From dahl.joachim at gmail.com  Thu Mar  1 02:38:29 2007
From: dahl.joachim at gmail.com (Joachim Dahl)
Date: Thu, 1 Mar 2007 08:38:29 +0100
Subject: [SciPy-user] scipy and cvxopt
In-Reply-To: <45E6080D.8090103@ee.byu.edu>
References: <32e43bb70702281340r39d69ea5t2e1287ed340eee9b@mail.gmail.com>
	<45E6080D.8090103@ee.byu.edu>
Message-ID: <47347f490702282338r54d5ce6cx1e9719f0e2f90c66@mail.gmail.com>

On 2/28/07, Travis Oliphant <oliphant at ee.byu.edu> wrote:
>
> Emin.shopper Martinian.shopper wrote:
>
> > Dear Experts,
> >
> > I need to solve some quadratic programs (and potentially other
> > nonlinear programs). While scipy.optimize.fmin_cobyla seems like it
> > can do this, it seems orders of magnitude slower than cvxopt. Are
> > there plans to merge/include cvxopt in scipy or otherwise improve
> > scipy's quadratic/nonlinear constrained optimization routines?
>
>
> Yes, eventually.  I have talked to the author of CVXOPT at NIPS 2006.
> The plan is to move NumPy's matrix object into C and move CVXOPTs
> implementation over to use it, possibly integrating the cvxopt
> algorithms into at least a scikits library (of GPL code).


That would be great!  We're considering different a license also.

Porting CVXOPT to Numpy/Scipy would require sparse matrices also -
preferably
implemented so that there's not much difference between dense and sparse
matrices from the user's perspective.  The sparse matrix class in SciPy
looks nice,  but it appears to
"behave" different from a dense matrix wrt. to indexing, assignments,
creation etc;
although it's not essential, it would make a CVXOPT port easier if some of
the differences between dense and sparse matrices were also ironed out when
Travis revamps the
Numpy matrix object.

Please let me know if there's anything I can do to help in process.

Joachim


But, I won't have time for that until April or May.
>
> -Travis
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070301/1c23517c/attachment.html>

From williams at astro.ox.ac.uk  Thu Mar  1 05:45:12 2007
From: williams at astro.ox.ac.uk (Michael Williams)
Date: Thu, 1 Mar 2007 10:45:12 +0000
Subject: [SciPy-user] [Numpy-discussion] NumPy in Teaching
In-Reply-To: <45E4FFA6.9010408@shrogers.com>
References: <45E4FFA6.9010408@shrogers.com>
Message-ID: <20070301104511.GA23386@astro.ox.ac.uk>

On Tue, Feb 27, 2007 at 09:05:58PM -0700, Steven H. Rogers wrote:
> I'm doing an informal survey on the use of Array Programming Languages 
> for teaching.  If you're using NumPy in this manner I'd like to hear 
> from you.  What subject was/is taught, academic level, results, lessons 
> learned, etc.

If Numeric counts, I used that back in 2002 as part of an introductory
programming course I wrote for the Department of Physics at Oxford. We
really only used to to provide an element-wise array method.

Brief introduction: http://pentangle.net/python/pyzine.php

The course (aka "Handbook") and report on the course's successes and
failures: http://pentangle.net/python/

-- Mike


From nwagner at iam.uni-stuttgart.de  Thu Mar  1 06:59:01 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Thu, 01 Mar 2007 12:59:01 +0100
Subject: [SciPy-user] fmin_cobyla
Message-ID: <45E6C005.5050507@iam.uni-stuttgart.de>

Hi all,

Is it possible to obtain more return values wrt to fmin_cobyla ?
The output is currently limited to the minimum (if any exists).

I am interested in

      fopt -- the value of f(xopt).
      func_calls -- the number of function_calls.
      allvecs  --  if retall then this vector of the iterates is returned


Nils



From gruben at bigpond.net.au  Thu Mar  1 08:07:22 2007
From: gruben at bigpond.net.au (Gary Ruben)
Date: Fri, 02 Mar 2007 00:07:22 +1100
Subject: [SciPy-user] NumPy in Teaching
In-Reply-To: <45E4FFA6.9010408@shrogers.com>
References: <45E4FFA6.9010408@shrogers.com>
Message-ID: <45E6D00A.5030706@bigpond.net.au>

Hi Steven,

Last year I helped out in teaching some basic programming with no 
prerequisites to 3rd year undergrad physics students at Monash 
University. It was really a 1st or 2nd year level course, but we had a 
wide spectrum of background experience levels - from no programming 
experience to proficient in C++. To deal with this variation in 
experience, we created some basic and some more advanced teaching labs. 
We divided the subject in half, giving a single C lecture first, 
followed by a few C labs, then a single Python lecture followed by a few 
Python labs; a deliberately chosen order and obviously very ambitious.

Our department is fairly IDL-centric, but Python's advantage of being 
free and its greater general applicability/flexibility was accepted by 
the course coordinator as sufficient reason to teach it. The hope is to 
get continuing 4th year honours students comfortable with a language for 
their 4th year physics projects.

We took the view - shared by colleagues in the Computer Science dept. - 
that getting the students to struggle with pointers and see the C-syntax 
would be good for their character :-) and that numpy would allow much 
higher level tasks to be attempted at an early stage and would get them 
used to using an array-processing mindset. I think that, since some of 
the students had prior C experience, they were able to help each other a 
bit more in the C labs. We found that we were busier answering questions 
in the Python labs as a result.

We had them create arrays in C, populating them with sinc functions to 
get them to deal with division by zero etc. and repeat the exercise in 
Python with numpy. We had them do some file i/o in both languages - I 
used scipy.io read_array and write_array to read data printf'ed by their 
C code. We did some fft-based filtering with Python and used pylab to 
view the results. We used Enthought Python with "ipython -pylab" shells 
and Idle as the editor. One lesson learned is that I tried to be a bit 
too ambitious with Python - they struggled with trying to figure out how 
to use functions. The labs were written as a mixture of "modify this 
example" code and "find the function which does this" - they found the 
latter too hard because the number of functions in numpy/scipy is a bit 
overwhelming and not easily navigable for the uninitiated.

We'll be re-running this course in a few weeks and perhaps introducing a 
physics modelling/numerical methods subject (perhaps language-neutral) 
later in the year. It may also be a prerequisite for some of the 3rd 
year astronomical data processing labs currently being written.

Gary R.

Steven H. Rogers wrote:
> I'm doing an informal survey on the use of Array Programming Languages 
> for teaching.  If you're using NumPy in this manner I'd like to hear 
> from you.  What subject was/is taught, academic level, results, lessons 
> learned, etc.
> 
> Regards,
> Steve



From lxander.m at gmail.com  Thu Mar  1 08:19:45 2007
From: lxander.m at gmail.com (Alexander Michael)
Date: Thu, 1 Mar 2007 08:19:45 -0500
Subject: [SciPy-user] Using SciPy/NumPy optimization THANKS!
In-Reply-To: <6.0.1.1.2.20070228135020.026f3008@pop.uky.edu>
References: <6.0.1.1.2.20070227131721.0287bfe0@pop.uky.edu>
	<45E493D7.2070703@gmail.com>
	<6.0.1.1.2.20070227170539.0283f608@pop.uky.edu>
	<6.0.1.1.2.20070228093647.02798d38@pop.uky.edu>
	<45E5BF5E.6020601@gmail.com>
	<6.0.1.1.2.20070228135020.026f3008@pop.uky.edu>
Message-ID: <525f23e80703010519y6cc4952bifd9b8a728324cfbf@mail.gmail.com>

On 2/28/07, Brandon Nuttall <bnuttall at uky.edu> wrote:
> Folks,
>
> Thanks to Alok Singhal and Robert Kern I have not only learned a great deal
> about SciPy and NumPy, but I have code that works. Thanks for the tip on
> not looping; it does make cleaner code. I have two issues: 1) there must be
> a better way to convert a list of data pairs to two arrays, 2) I'm not sure
> of a graceful way to transition from one plot to the next and then close.
>

To add to the cacophony of coding and style suggestions. The <>
operator is likely to be removed in the future, so you should use
'!='. My personal preference would be to move the plotting
functionality to a method so that initializing the data is separate
from acting on the data. I find this to be a helpful distinction as I
usually don't want to plot at the time of construction, but your
mileage may vary. Lastly, you can wrap the non-plotting portion of the
test function into a doctest
(<http://www.python.org/doc/lib/module-doctest.html>) which would then
both serve as an example in the code and as a correctness test which
is easy to check when other things change, like upgrades to numpy and
scipy. I find this to be immensely helpful.

Have fun!
Alex


From steve at shrogers.com  Thu Mar  1 09:09:01 2007
From: steve at shrogers.com (Steven H. Rogers)
Date: Thu, 01 Mar 2007 07:09:01 -0700
Subject: [SciPy-user] [Numpy-discussion] NumPy in Teaching
In-Reply-To: <20070301104511.GA23386@astro.ox.ac.uk>
References: <45E4FFA6.9010408@shrogers.com>
	<20070301104511.GA23386@astro.ox.ac.uk>
Message-ID: <45E6DE7D.9020307@shrogers.com>

Thanks Mike:

Michael Williams wrote:
> On Tue, Feb 27, 2007 at 09:05:58PM -0700, Steven H. Rogers wrote:
>   
>> I'm doing an informal survey on the use of Array Programming Languages 
>> for teaching.  If you're using NumPy in this manner I'd like to hear 
>> from you.  What subject was/is taught, academic level, results, lessons 
>> learned, etc.
>>     
>
> If Numeric counts, I used that back in 2002 as part of an introductory
> programming course I wrote for the Department of Physics at Oxford. We
> really only used to to provide an element-wise array method.
>   
Yes, Numeric and Numarray certainly count.    The comments I've received 
about Matlab and IDL are also welcome. 
> Brief introduction: http://pentangle.net/python/pyzine.php
>
> The course (aka "Handbook") and report on the course's successes and
> failures: http://pentangle.net/python/
>
> -- Mike
> ________
# Steve


From perry at stsci.edu  Thu Mar  1 11:03:34 2007
From: perry at stsci.edu (Perry Greenfield)
Date: Thu, 1 Mar 2007 11:03:34 -0500
Subject: [SciPy-user] [Numpy-discussion] NumPy in Teaching
In-Reply-To: <200703010032.l210WfhI005995@glup.physics.ucf.edu>
References: <200703010032.l210WfhI005995@glup.physics.ucf.edu>
Message-ID: <D54CE207-CAB6-464C-8A3B-38E1D2154FFB@stsci.edu>


On Feb 28, 2007, at 7:32 PM, Joe Harrington wrote:

> Hi Steve,
>
> I have taught Astronomical Data Analysis twice at Cornell using IDL,
> and I will be teaching it next Fall at UCF using NumPy.  Though I've
> been active here in the recent past, I'm actually not a regular NumPy
> user myself yet (I used Numeric experimentally for about 6 months in
> 1997), so I'm a bit nervous.  There isn't the kind of documentation
> and how-to support for Numpy that there is for IDL, though our web
> site is a start in that direction.  One thought I've had in making the
> transition easier is to put up a syntax and function concordance,
> similar to that available for MATLAB.  I thought this existed.  Maybe
> Perry can point me to it.  Just adding a column to the MATLAB one
> would be fine.

I made one for IDL, but I don't recall one for matlab. If anyone has  
done one, I sure would like to incorporate it into the tutorial I'm  
revising if possible.

Perry
y


From strawman at astraw.com  Thu Mar  1 11:57:46 2007
From: strawman at astraw.com (Andrew Straw)
Date: Thu, 01 Mar 2007 08:57:46 -0800
Subject: [SciPy-user] [Numpy-discussion] NumPy in Teaching
In-Reply-To: <D54CE207-CAB6-464C-8A3B-38E1D2154FFB@stsci.edu>
References: <200703010032.l210WfhI005995@glup.physics.ucf.edu>
	<D54CE207-CAB6-464C-8A3B-38E1D2154FFB@stsci.edu>
Message-ID: <45E7060A.9040105@astraw.com>

Perry Greenfield wrote:
> On Feb 28, 2007, at 7:32 PM, Joe Harrington wrote:
>
>   
>> Hi Steve,
>>
>> I have taught Astronomical Data Analysis twice at Cornell using IDL,
>> and I will be teaching it next Fall at UCF using NumPy.  Though I've
>> been active here in the recent past, I'm actually not a regular NumPy
>> user myself yet (I used Numeric experimentally for about 6 months in
>> 1997), so I'm a bit nervous.  There isn't the kind of documentation
>> and how-to support for Numpy that there is for IDL, though our web
>> site is a start in that direction.  One thought I've had in making the
>> transition easier is to put up a syntax and function concordance,
>> similar to that available for MATLAB.  I thought this existed.  Maybe
>> Perry can point me to it.  Just adding a column to the MATLAB one
>> would be fine.
>>     
>
> I made one for IDL, but I don't recall one for matlab. If anyone has  
> done one, I sure would like to incorporate it into the tutorial I'm  
> revising if possible.
>
>   
I believe the reference is to this: http://scipy.org/NumPy_for_Matlab_Users


From oliphant at ee.byu.edu  Thu Mar  1 12:34:01 2007
From: oliphant at ee.byu.edu (Travis Oliphant)
Date: Thu, 01 Mar 2007 10:34:01 -0700
Subject: [SciPy-user] scipy and cvxopt
In-Reply-To: <47347f490702282338r54d5ce6cx1e9719f0e2f90c66@mail.gmail.com>
References: <32e43bb70702281340r39d69ea5t2e1287ed340eee9b@mail.gmail.com>	<45E6080D.8090103@ee.byu.edu>
	<47347f490702282338r54d5ce6cx1e9719f0e2f90c66@mail.gmail.com>
Message-ID: <45E70E89.4060501@ee.byu.edu>

Joachim Dahl wrote:

>
>
> On 2/28/07, *Travis Oliphant* <oliphant at ee.byu.edu 
> <mailto:oliphant at ee.byu.edu>> wrote:
>
>     Emin.shopper Martinian.shopper wrote:
>
>     > Dear Experts,
>     >
>     > I need to solve some quadratic programs (and potentially other
>     > nonlinear programs). While scipy.optimize.fmin_cobyla seems like it
>     > can do this, it seems orders of magnitude slower than cvxopt. Are
>     > there plans to merge/include cvxopt in scipy or otherwise improve
>     > scipy's quadratic/nonlinear constrained optimization routines?
>
>
>     Yes, eventually.  I have talked to the author of CVXOPT at NIPS 2006.
>     The plan is to move NumPy's matrix object into C and move CVXOPTs
>     implementation over to use it, possibly integrating the cvxopt
>     algorithms into at least a scikits library (of GPL code). 
>
>
> That would be great!  We're considering different a license also.
>
> Porting CVXOPT to Numpy/Scipy would require sparse matrices also - 
> preferably
> implemented so that there's not much difference between dense and 
> sparse matrices from the user's perspective.  The sparse matrix class 
> in SciPy looks nice,  but it appears to
> "behave" different from a dense matrix wrt. to indexing, assignments, 
> creation etc;
> although it's not essential, it would make a CVXOPT port easier if 
> some of the differences between dense and sparse matrices were also 
> ironed out when Travis revamps the
> Numpy matrix object.

I'd like to move a the Sparse matrix representation into NumPy as well 
at that time.  We will need some discussion on what exactly to move over 
and how that is to be done.    There are some differences in indexing 
that CVXOPT uses for its matrix classes that we will need to discuss how 
to handle.  For example, we could construct a separate matrix subclass 
with the CVXOPT-style indexing.  I think it would be easier to just have 
one style of indexing, though and just make sure that sparse matrices 
also handle that style.

The most important thing to me, though, is to expose more users to the 
great work that CVXOPT has done and integrate that work into the larger 
NumPy/SciPy world.

-Travis





From dahl.joachim at gmail.com  Thu Mar  1 14:41:46 2007
From: dahl.joachim at gmail.com (Joachim Dahl)
Date: Thu, 1 Mar 2007 20:41:46 +0100
Subject: [SciPy-user] scipy and cvxopt
In-Reply-To: <45E70E89.4060501@ee.byu.edu>
References: <32e43bb70702281340r39d69ea5t2e1287ed340eee9b@mail.gmail.com>
	<45E6080D.8090103@ee.byu.edu>
	<47347f490702282338r54d5ce6cx1e9719f0e2f90c66@mail.gmail.com>
	<45E70E89.4060501@ee.byu.edu>
Message-ID: <47347f490703011141geea3bfdx4e7c3f3e155a0489@mail.gmail.com>

>
> I'd like to move a the Sparse matrix representation into NumPy as well
> at that time.  We will need some discussion on what exactly to move over
> and how that is to be done.    There are some differences in indexing
> that CVXOPT uses for its matrix classes that we will need to discuss how
> to handle.  For example, we could construct a separate matrix subclass
> with the CVXOPT-style indexing.  I think it would be easier to just have
> one style of indexing, though and just make sure that sparse matrices
> also handle that style.
>
> The most important thing to me, though, is to expose more users to the
> great work that CVXOPT has done and integrate that work into the larger
> NumPy/SciPy world.


that sounds great - I look forward to helping any way I can!

- Joachim
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070301/a622b43e/attachment.html>

From asreeve at maine.edu  Thu Mar  1 17:03:32 2007
From: asreeve at maine.edu (ASReeve)
Date: Thu, 1 Mar 2007 17:03:32 -0500 (EST)
Subject: [SciPy-user] PhD Opportunity
Message-ID: <Pine.LNX.4.64.0703011700260.11738@hydro1.geology.um.maine.edu>

My apologies if this if off topic.

I am seeking a Ph.D. Student interested in using python for hydrologic
simulation. Funding for a student is available for 4 years through an
NSF grant. The research focus of this project will be the creation of
finite-volume models, likely using FiPy, to simulate ground-water
flow, solute transport, and carbon accumulation in a wetland system.

Prospective Ph.D. students interested in working on this project should 
contact Andy Reeve (e-mail:asreeve at maine dot edu) for additional 
details. Please e-mail me directly and do not reply to this message 
through the mailing list.

Thanks,
Andy

----------------------
Andrew Reeve
Dept. of Earth Science
University of Maine
Orono, ME 04469



From fperez.net at gmail.com  Thu Mar  1 17:06:37 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Thu, 1 Mar 2007 15:06:37 -0700
Subject: [SciPy-user] PhD Opportunity
In-Reply-To: <Pine.LNX.4.64.0703011700260.11738@hydro1.geology.um.maine.edu>
References: <Pine.LNX.4.64.0703011700260.11738@hydro1.geology.um.maine.edu>
Message-ID: <db6b5ecc0703011406t4c1abc4dj6fefb38ba5ff176f@mail.gmail.com>

On 3/1/07, ASReeve <asreeve at maine.edu> wrote:
> My apologies if this if off topic.

Personally, I don't think this is off topic /at all/.  I hope one day
will come when python is so widely accepted as a research tool that
we'll have to ask people not to post such announcements here, but I
think that while we are in the 'early evangelism' phase, messages such
as yours are actually very valuable.  They show students that this
isn't necessarily a professional dead end.

Just my opinion...

Best,

f


From drfredkin at ucsd.edu  Thu Mar  1 16:06:42 2007
From: drfredkin at ucsd.edu (Donald Fredkin)
Date: Thu, 1 Mar 2007 21:06:42 +0000 (UTC)
Subject: [SciPy-user] Any Books on SciPy?
References: <D73DF6B6-D5A3-45E4-B7BA-F514307B3321@airmail.net>
	<db6b5ecc0702272219k3a5d67dft7d2f1980b86950d1@mail.gmail.com>
	<45E567A2.3050406@shrogers.com>
	<200702280821.35981.dd55@cornell.edu>
Message-ID: <es7f91$tbm$1@sea.gmane.org>

Darren Dale wrote:

> I haven't seen anyone mention "Numerical Methods in Engineering with
> Python" by Jaan Kiusalaas. It was published in 2005, and uses
> numarray, so it is somewhat dated considering all the impressive
> developments with NumPy and SciPy. I mention it for the sake of
> completeness.

This book has a first chapter on the basics of Python. The remainder is
a light weight introduction to numerical methods and does not teach
anything about Python.

-- 




From david at ar.media.kyoto-u.ac.jp  Thu Mar  1 22:56:33 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Fri, 02 Mar 2007 12:56:33 +0900
Subject: [SciPy-user] [Numpy-discussion] NumPy in Teaching
In-Reply-To: <45E7060A.9040105@astraw.com>
References: <200703010032.l210WfhI005995@glup.physics.ucf.edu>	<D54CE207-CAB6-464C-8A3B-38E1D2154FFB@stsci.edu>
	<45E7060A.9040105@astraw.com>
Message-ID: <45E7A071.5020303@ar.media.kyoto-u.ac.jp>

Andrew Straw wrote:
> Perry Greenfield wrote:
>> On Feb 28, 2007, at 7:32 PM, Joe Harrington wrote:
>>
>>   
>>> Hi Steve,
>>>
>>> I have taught Astronomical Data Analysis twice at Cornell using IDL,
>>> and I will be teaching it next Fall at UCF using NumPy.  Though I've
>>> been active here in the recent past, I'm actually not a regular NumPy
>>> user myself yet (I used Numeric experimentally for about 6 months in
>>> 1997), so I'm a bit nervous.  There isn't the kind of documentation
>>> and how-to support for Numpy that there is for IDL, though our web
>>> site is a start in that direction.  One thought I've had in making the
>>> transition easier is to put up a syntax and function concordance,
>>> similar to that available for MATLAB.  I thought this existed.  Maybe
>>> Perry can point me to it.  Just adding a column to the MATLAB one
>>> would be fine.
>>>     
>> I made one for IDL, but I don't recall one for matlab. If anyone has  
>> done one, I sure would like to incorporate it into the tutorial I'm  
>> revising if possible.
>>
>>   
> I believe the reference is to this: http://scipy.org/NumPy_for_Matlab_Users
Another reference which was really useful for me was this one

http://37mm.no/matlab-python-xref.html

I actually found it more useful than the wiki (at the time I made the 
step towards numpy for all my numerical needs, that is around one year 
ago). Note that this site also have reference wrt IDL and R, the latter 
having been useful for me when I thought about ditching matlab for R

David



From david.warde.farley at utoronto.ca  Thu Mar  1 23:13:25 2007
From: david.warde.farley at utoronto.ca (David Warde-Farley)
Date: Thu, 01 Mar 2007 23:13:25 -0500
Subject: [SciPy-user] Any Books on SciPy?
In-Reply-To: <es7f91$tbm$1@sea.gmane.org>
References: <D73DF6B6-D5A3-45E4-B7BA-F514307B3321@airmail.net>
	<db6b5ecc0702272219k3a5d67dft7d2f1980b86950d1@mail.gmail.com>
	<45E567A2.3050406@shrogers.com> <200702280821.35981.dd55@cornell.edu>
	<es7f91$tbm$1@sea.gmane.org>
Message-ID: <1172808805.14827.23.camel@rodimus>

On Thu, 2007-03-01 at 21:06 +0000, Donald Fredkin wrote:
> > I haven't seen anyone mention "Numerical Methods in Engineering with
> > Python" by Jaan Kiusalaas. It was published in 2005, and uses
> > numarray, so it is somewhat dated considering all the impressive
> > developments with NumPy and SciPy. I mention it for the sake of
> > completeness.
> 
> This book has a first chapter on the basics of Python. The remainder
> is
> a light weight introduction to numerical methods and does not teach
> anything about Python.

A lab mate of mine showed me this today. It does seem useful in that it
contains lots of little recipes. Unfortunately, I think that the recipe
I was interested in, namely the conjugate gradient solver linear
systems, was incorrect or at least incompatible with NumPy (the
residuals were *diverging*, somehow) and I ended up rewriting it from
scratch.

David





From robert.vergnes at yahoo.fr  Fri Mar  2 01:33:16 2007
From: robert.vergnes at yahoo.fr (Robert VERGNES)
Date: Fri, 2 Mar 2007 07:33:16 +0100 (CET)
Subject: [SciPy-user] QME-Dev Workbench (wxSciPy) Alpha 1.6 - RELEASED TODAY
	on sourceforge
Message-ID: <20070302063316.92779.qmail@web27411.mail.ukl.yahoo.com>


The QME-DEV Data analysis Workbench Alpha 1.6 has been released:
 
https://sourceforge.net/project/showfiles.php?group_id=181979

Looking for courageous testers..for the workbench.

Solved issues:
- window resizing, load/save, graph/plotting, lost of focus, change of the editor.

Lib-Needed : Nympy/Scipy/Pylab, Python 2.4 , wxpython 2.8

Best Regards,

Robert





 		
---------------------------------
 D?couvrez une nouvelle fa?on d'obtenir des r?ponses ? toutes vos questions ! Profitez des connaissances, des opinions et des exp?riences des internautes sur Yahoo! Questions/R?ponses.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070302/c155506d/attachment.html>

From franckm at aims.ac.za  Fri Mar  2 05:15:50 2007
From: franckm at aims.ac.za (Franck Kalala Mutombo)
Date: Fri, 02 Mar 2007 12:15:50 +0200
Subject: [SciPy-user] Fitting sphere to 3d data points
In-Reply-To: <BF49BEA8-57C4-4207-A919-EEA780576C9D@nih.gov>
References: <715CC8AB-4406-41A8-83EF-72F419738896@nih.gov>	<91cf711d0701250631r35a9602g8f8ed48979e7cde3@mail.gmail.com>
	<BF49BEA8-57C4-4207-A919-EEA780576C9D@nih.gov>
Message-ID: <45E7F956.9050400@aims.ac.za>

James Vincent wrote:
> David,
> 
> Thanks for the reply. I think I should have been clearer about the
> problem. I have a surface patch of data points, not an actual whole
> sphere. It will probably be a very small section of total sphere
> surface. I would like to fit the sphere that best fits just those points
> that I have. The center of the sphere will be highly dependent on the
> curvature of the points. 
> 
> I think the  leastsq routine is right, I just can't figure out how to
> pass the data in yet (it's my first work with scipy). 
> 
> Jim
> 
> 
> On Jan 25, 2007, at 9:31 AM, David Huard wrote:
> 
>> Hi James,
>>
>> As a first guess, I'd say the center of the sphere is simply the mean
>> of your data points, if they're all weighted equally. With only one
>> parameter left to fit, it should be easy enough. However, you may want
>> to look at the paper:
>>
>> Werman, Michael and Keren, Daniel
>> A Bayesian method for fitting parametric and nonparametric models to
>> noisy data
>> Ieee Transactions on Pattern Analysis and Machine Intelligence, 23, 2001.
>>
>> They write that the Mean Square Error approach overestimates the
>> radius in the case of circles. They don't talk about the 3D case, but
>> I'd guess similar problems arise. They provide a method to fit
>> parametric shapes with some robustness to data errors.
>>
>> Cheers,
>>
>> David
>>
>>
>>
>> 2007/1/25, James Vincent <jjv5 at nih.gov <mailto:jjv5 at nih.gov>>:
>>
>>     Hello,
>>
>>     Is it possible to fit a sphere to 3D data points
>>     using scipy.optimize.leastsq? I'd like to minimize the residual
>>     for the distance from the actual x,y,z point and the fitted sphere
>>     surface. I can see how to minimize for z, but that's not really
>>     what I'm looking for. Is there a better way to do this? Thanks for
>>     any help.
>>
>>     params = a,b,c and r
>>     a,b,c are the fitted center point of the sphere, r is the radius
>>
>>     err = distance-to-center - radius
>>     err = sqrt( x-a)**2 + (y-b)**2 + (z-c)**2) - r
>>
>>
>>
>>     ----
>>     James J. Vincent, Ph.D.
>>     National Cancer Institute
>>     National Institutes of Health
>>     Laboratory of Molecular Biology
>>     Building 37, Room 5120
>>     37 Convent Drive, MSC 4264
>>     Bethesda, MD 20892 USA
>>
>>     301-451-8755
>>     jjv5 at nih.gov <mailto:jjv5 at nih.gov>
>>
>>
>>
>>     _______________________________________________
>>     SciPy-user mailing list
>>     SciPy-user at scipy.org <mailto:SciPy-user at scipy.org>
>>     http://projects.scipy.org/mailman/listinfo/scipy-user
>>
>>
>>
>> _______________________________________________
>> SciPy-user mailing list
>> SciPy-user at scipy.org <mailto:SciPy-user at scipy.org>
>> http://projects.scipy.org/mailman/listinfo/scipy-user
> 
> ----
> James J. Vincent, Ph.D.
> National Cancer Institute
> National Institutes of Health
> Laboratory of Molecular Biology
> Building 37, Room 5120
> 37 Convent Drive, MSC 4264
> Bethesda, MD 20892 USA
> 
> 301-451-8755
> jjv5 at nih.gov <mailto:jjv5 at nih.gov>
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
Hi,

I have a sequence of decimal number (1,2,...16) for example, I want if
there is any function  which can convert each one in binary.

Thanks

-- 
Franck
African Institute for Mathematical Sciences -- www.aims.ac.za


From jelle.feringa at ezct.net  Fri Mar  2 07:26:21 2007
From: jelle.feringa at ezct.net (Jelle Feringa / EZCT Architecture & Design Research)
Date: Fri, 2 Mar 2007 13:26:21 +0100
Subject: [SciPy-user] writing string from array
Message-ID: <001301c75cc5$fcf0f130$c000a8c0@JELLE>

Hi,

 

I'm trying to send a grid of points to a subprocess.Popen object, via the
.communicate(input=some_array) method.

To do so I need to write a string that looks like this:

 

xA, yA, zA, xB, yB, zB

 

To do so I intend to use the mgrid function:

sample_grid = sp.mgrid[min_x:max_x:spacing, min_y:max_y:spacing,
min_z:max_z:spacing ]

 

To obtain the xA, yA, zA values, and to add an array((0,0,0.1)) to these
values to obtain xB, yB, zB

 

Unfortuantely I have no idea on how I can do so (using the .tostring
method?)

Any pointers would be appreciated for sure!

 

Thanks,

 

-jelle

 

 

Jelle Feringa
EZCT ARCHITECTURE & DESIGN RESEARCH


Office:  +33 (0) 1 42 40 19 81
Fax:     +33 (0) 1 43 14 94 61

Cell Nl:  +31 (0) 6 44 02 10 15

Cell Fr:  +33 (0) 6 63 51 27 46
 <http://www.ezct.net/> www.ezct.net
 <mailto:jelle.feringal at ezct.net> jelle.feringa at ezct.net

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070302/4372558c/attachment.html>

From lorenzo.isella at gmail.com  Fri Mar  2 08:25:29 2007
From: lorenzo.isella at gmail.com (Lorenzo Isella)
Date: Fri, 2 Mar 2007 14:25:29 +0100
Subject: [SciPy-user] Optimization Working only with a Specific Expression
	of the input Parameters
Message-ID: <a2b3004b0703020525k3f01a49dke1fcdcb5b6c365a1@mail.gmail.com>

Dear All,
I was trying to fit some data using the leastsq package in
scipy.optimize. The function I would like to use to fit my data is:

log(10.0)*A1/sqrt(2.0*pi)/log(myvar1)*exp(-((log(x/mu1))**2.0)/2.0/log(myvar1)/log(myvar1)))

 where A1, mu1 and myvar1 are fitting parameters.
For some reason, I used to get an error message from scipy.optimize
telling me that I was not working with an array of floats.
I suppose that this is due to the fact that the optimizer also tries
solving for negative values of mu1 and myvar1, for which the log
function (x is always positive) does not exist.
In fact, if I use the fitting function:

log(10.0)*A1/sqrt(2.0*pi)/log(myvar1**2.0)*exp(-((log(x/mu1**2.0))**2.0)/2.0/log(myvar1**2.0)/log(myvar1**2.0)))

Where mu1 and myvar1 appear squared, then the problem does not exist
any longer and the results are absolutely ok.
Can anyone enlighten me here and confirm this is what is really going on?
Kind Regards

Lorenzo


From nmarais at sun.ac.za  Fri Mar  2 08:37:49 2007
From: nmarais at sun.ac.za (Neilen Marais)
Date: Fri, 02 Mar 2007 15:37:49 +0200
Subject: [SciPy-user] Preconditioned iterative matrix solution
Message-ID: <pan.2007.03.02.10.13.22.213171@sun.ac.za>

Hi,

Is there any practical examples/documentation of how to do preconditioned
iterative matrix solution with scipy? If it turns out there isn't any I
could put my experiences up on the wiki.

Thanks
Neilen

-- 
you know its kind of tragic 
we live in the new world
but we've lost the magic
-- Battery 9 (www.battery9.co.za)



From nwagner at iam.uni-stuttgart.de  Fri Mar  2 08:50:54 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Fri, 02 Mar 2007 14:50:54 +0100
Subject: [SciPy-user] Preconditioned iterative matrix solution
In-Reply-To: <pan.2007.03.02.10.13.22.213171@sun.ac.za>
References: <pan.2007.03.02.10.13.22.213171@sun.ac.za>
Message-ID: <45E82BBE.8040901@iam.uni-stuttgart.de>

Neilen Marais wrote:
> Hi,
>
> Is there any practical examples/documentation of how to do preconditioned
> iterative matrix solution with scipy? If it turns out there isn't any I
> could put my experiences up on the wiki.
>
> Thanks
> Neilen
>
>   
Hi Neilen,

I am not aware of any scipy example. Thus it would be great if you could
add it.

Thanks in advance !

Cheers
                      Nils
 


From bnuttall at uky.edu  Fri Mar  2 08:50:53 2007
From: bnuttall at uky.edu (Brandon Nuttall)
Date: Fri, 02 Mar 2007 08:50:53 -0500
Subject: [SciPy-user] Optimization Working only with a Specific
 Expression of the input Parameters
In-Reply-To: <a2b3004b0703020525k3f01a49dke1fcdcb5b6c365a1@mail.gmail.co
 m>
References: <a2b3004b0703020525k3f01a49dke1fcdcb5b6c365a1@mail.gmail.com>
Message-ID: <6.0.1.1.2.20070302083425.027a3548@pop.uky.edu>

Hello,

I'm just an amateur, but it seems to me like the array data in myvar1 are 
likely integers. When you raise the data to a power of type float (i.e. 
2.0) all the members of the array are automatically converted to real 
(float) types. Easiest and fastest thing I know to do would be:

      myvar1 = myvar1*1.0

Or, and probably preferred (assuming you are using the numpy array type and 
have imported it):

      myvar1 = numpy.array(myvar1,dtype=float)

Brandon

At 08:25 AM 3/2/2007, you wrote:
>Dear All,
>I was trying to fit some data using the leastsq package in
>scipy.optimize. The function I would like to use to fit my data is:
>
>log(10.0)*A1/sqrt(2.0*pi)/log(myvar1)*exp(-((log(x/mu1))**2.0)/2.0/log(myvar1)/log(myvar1)))
>
>  where A1, mu1 and myvar1 are fitting parameters.
>For some reason, I used to get an error message from scipy.optimize
>telling me that I was not working with an array of floats.
>I suppose that this is due to the fact that the optimizer also tries
>solving for negative values of mu1 and myvar1, for which the log
>function (x is always positive) does not exist.
>In fact, if I use the fitting function:
>
>log(10.0)*A1/sqrt(2.0*pi)/log(myvar1**2.0)*exp(-((log(x/mu1**2.0))**2.0)/2.0/log(myvar1**2.0)/log(myvar1**2.0)))
>
>Where mu1 and myvar1 appear squared, then the problem does not exist
>any longer and the results are absolutely ok.
>Can anyone enlighten me here and confirm this is what is really going on?
>Kind Regards
>
>Lorenzo
>_______________________________________________
>SciPy-user mailing list
>SciPy-user at scipy.org
>http://projects.scipy.org/mailman/listinfo/scipy-user


Brandon C. Nuttall

BNUTTALL at UKY.EDU                         Kentucky Geological Survey
(859) 257-5500                                     University of Kentucky
(859) 257-1147 (fax)                              228 Mining & Mineral 
Resources Bldg
http://www.uky.edu/KGS/home.htm       Lexington, Kentucky 40506-0107  



From robert.vergnes at yahoo.fr  Fri Mar  2 08:57:30 2007
From: robert.vergnes at yahoo.fr (Robert VERGNES)
Date: Fri, 2 Mar 2007 14:57:30 +0100 (CET)
Subject: [SciPy-user] odeint-lsoda
Message-ID: <20070302135733.14339.qmail@web27403.mail.ukl.yahoo.com>

Hello,

is there a manual for odeint and lsoda ?

On the following small code (see below) oedint gives a warning and I would like to understand what is it... exactly and how to adjust it.

Thanx

Robert

Warning messgae:

Commandline: C:\Python25\python.exe F:\RFV-WO~1\DT_DAN~1\SOFTWA~1\MQE_PE~1\PYTHON~1\FEB200~2\SmallODE.py
##Workingdirectory: F:\RFV-WorkFiles\DT_Danil_EQM\SoftwarePendulumEQM\MQE_Pendulum_2006\Python_rfv\Feb20007_QPendulum
##Timeout: 0 ms
##
## lsoda--  at current t (=r1), mxstep (=i1) steps
##       taken on this call before reaching tout
##      In above message,  I1 =       500
##      In above message,  R1 =  0.3691558106967E+02
##Excess work done on this call (perhaps wrong Dfun type).
##Run with full_output = 1 to get quantitative information.
##[[ 0.05        0.785     ]
## [-0.08825151  0.78480528]
## [-0.22531056  0.78323728]
## ...,
## [ 0.          0.        ]
## [ 0.          0.        ]
## [ 0.          0.        ]]
##
##Process "Pyhton Interpreter" terminated, ExitCode: 00000000

small ode code:


#Import of librairies used:
from scipy import *
from scipy.integrate import *
from math import *
#from pylab import *
#from string import *
#import time
import types
#import logging
------------------------------------------------------
Small ode test
-------------------------------------------------------

def func_r3(y,t):
    return [ - fo*sign(y[0])- w0**2*sin(y[1])+ eval(RHS), y[0]]
    # rhs as alpha*sin(gamma*t+beta)

def func_f3(y,t):
    return [ - fo*sign(y[0])- w0**2*sin(y[1]), y[0]]

#----Param Settings:----
#Set the time:
t=arange(1. ,50. ,0.01)   # from 1 to aaa in steps of bbb
#Define wo
w0 = sqrt (9.81/0.5)
#Define fo
fo = 0.15   #example value fo=0.2
#we define parameters of the electrical or mechanical circuit
alpha = 1.5 # try between 0.3 and 5
gamma =  4.43 #for 50Hz - gamma = 314.159 / and for 60Hz gamma = 376.99 - or w0 for forced oscillation
beta = 0  #(used only if we now the exact phase change
RHS=rhs = 'alpha*sin(gamma*t+beta)'
#Define Initial Condtion
rfv_y0_0 =  0.05  # = theta_dot(t=0) in rd/s
rfv_y1_0 =  0.785  # = theta (t=0) in rd
y0 = [rfv_y0_0, rfv_y1_0]  # definition of the initial condition vector
# 0.08727 rd = 5 deg
rfv_qm=-0.08727
rfv_qp=0.08727
#---------------------------------------
#QPendulum(Pendultype , y0, t, w0, fo =0, RHS='', qp=0, qm=0, args='', Dfun='')

y_final = array([[0,0]]) # we define y_final with 2 vectors - the same
y=odeint(func_f3,y0,t)

print y

 		
---------------------------------
 D?couvrez une nouvelle fa?on d'obtenir des r?ponses ? toutes vos questions ! Profitez des connaissances, des opinions et des exp?riences des internautes sur Yahoo! Questions/R?ponses.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070302/53394059/attachment.html>

From david.warde.farley at utoronto.ca  Fri Mar  2 11:41:54 2007
From: david.warde.farley at utoronto.ca (David Warde-Farley)
Date: Fri, 2 Mar 2007 11:41:54 -0500
Subject: [SciPy-user] dot product of vectors stored as "matrix"?
Message-ID: <31AF6D91-60CE-42C6-B09A-224EBF10C43F@utoronto.ca>

  This is more a question of elegance than anything else, but, given  
2 column vectors (or row vectors) stored as type "matrix", why  
doesn't dot(x,y) give you their dot product? I get complaints about  
"matrices not aligned".

Of course dot(x.T, y) works (with the annoying caveat that you get a  
matrix and have to dereference its single element), and so does dot 
(x.A.squeeze(), y.A.squeeze()), and so too does sum(multiply(x,y)).

Is there a preferred notation for linear-algebra dot products, one  
that's faster than the others?

David


From drfredkin at ucsd.edu  Fri Mar  2 11:46:13 2007
From: drfredkin at ucsd.edu (Donald Fredkin)
Date: Fri, 2 Mar 2007 16:46:13 +0000 (UTC)
Subject: [SciPy-user] odeint-lsoda
References: <20070302135733.14339.qmail@web27403.mail.ukl.yahoo.com>
Message-ID: <es9kck$hjp$1@sea.gmane.org>

Robert VERGNES wrote:

> is there a manual for odeint and lsoda ?

You can view the source code for lsoda on netlib. The extensive
comments at the head of the source file are clear and complete. Odeint
is just a wrapper for lsoda, so when you understand what lsoda is doing
you will be all set.

In your case, you might simply be computing for to coarse a mesh in t,
so "too much work" has to be done for each step. The error message
tells you what value of t was reached before lsoda gave up.

-- 




From jan at aims.ac.za  Fri Mar  2 11:53:22 2007
From: jan at aims.ac.za (Jan Groenewald)
Date: Fri, 2 Mar 2007 18:53:22 +0200
Subject: [SciPy-user] Integer to binary [Was: Fitting sphere to 3d data
	points]
In-Reply-To: <45E7F956.9050400@aims.ac.za>
References: <715CC8AB-4406-41A8-83EF-72F419738896@nih.gov>
	<91cf711d0701250631r35a9602g8f8ed48979e7cde3@mail.gmail.com>
	<BF49BEA8-57C4-4207-A919-EEA780576C9D@nih.gov>
	<45E7F956.9050400@aims.ac.za>
Message-ID: <20070302165322.GA25987@aims.ac.za>

Hi Franck

When you post, please don't reply to another thread of messages with
a good descriptive subject. Your message not only gets lost by any
email client that sorts by threads (mails with the same subject matter)
it also annoys some readers who use this feature.

Please start a blank message if you have a new question, and give it
a descriptive subject.

On Fri, Mar 02, 2007 at 12:15:50PM +0200, Franck Kalala Mutombo wrote:
> I have a sequence of decimal number (1,2,...16) for example, I want if
> there is any function  which can convert each one in binary.

A quick google of int2bin+python found this:

def int2bin(num, width=32):
    return ''.join(['%c'%(ord('0')+bool((1<<k)&num)) for k in range((width-1),-1,-1)])

cheers,
Jan
-- 
   .~.
   /V\     Jan Groenewald
  /( )\    www.aims.ac.za
  ^^-^^


From lorenzo.isella at gmail.com  Fri Mar  2 14:30:35 2007
From: lorenzo.isella at gmail.com (Lorenzo Isella)
Date: Fri, 02 Mar 2007 20:30:35 +0100
Subject: [SciPy-user] Optimization Working only with a Specific
 Expression of the input Parameters
Message-ID: <45E87B5B.4000703@gmail.com>

Hi Brandon,
Thanks for your advice, but I am a bit confused: myvar1 is simply a 
fitting parameter (i.e. it is used to return an output), nothing is 
stored in it to start from.
I do not define it anywhere.
It is not an array. Furthermore, I have to say that if I define the 
error function as the absolute difference between my data and the 
function I want to use for the fitting, then the code executes even 
without raising any parameter to the second power, but returns nonsense 
(negative variance and so on).
Instead, without the abs(), I still get the same problem mentioned in my 
previous email.
There is something I must be misunderstanding...it is not a tough 
optimization at all the one I am carrying out...
Cheers

Lorenzo









Hello,

I'm just an amateur, but it seems to me like the array data in myvar1 are
likely integers. When you raise the data to a power of type float (i.e.
2.0) all the members of the array are automatically converted to real
(float) types. Easiest and fastest thing I know to do would be:

      myvar1 = myvar1*1.0

Or, and probably preferred (assuming you are using the numpy array type and
have imported it):

      myvar1 = numpy.array(myvar1,dtype=float)

Brandon

At 08:25 AM 3/2/2007, you wrote:

> >Dear All,
> >I was trying to fit some data using the leastsq package in
> >scipy.optimize. The function I would like to use to fit my data is:
> >
> >log(10.0)*A1/sqrt(2.0*pi)/log(myvar1)*exp(-((log(x/mu1))**2.0)/2.0/log(myvar1)/log(myvar1)))
> >
> >  where A1, mu1 and myvar1 are fitting parameters.
> >For some reason, I used to get an error message from scipy.optimize
> >telling me that I was not working with an array of floats.
> >I suppose that this is due to the fact that the optimizer also tries
> >solving for negative values of mu1 and myvar1, for which the log
> >function (x is always positive) does not exist.
> >In fact, if I use the fitting function:
> >
> >log(10.0)*A1/sqrt(2.0*pi)/log(myvar1**2.0)*exp(-((log(x/mu1**2.0))**2.0)/2.0/log(myvar1**2.0)/log(myvar1**2.0)))
> >
> >Where mu1 and myvar1 appear squared, then the problem does not exist
> >any longer and the results are absolutely ok.
> >Can anyone enlighten me here and confirm this is what is really going on?
> >Kind Regards
> >
> >Lorenzo
> >_______________________________________________
> >SciPy-user mailing list
> >SciPy-user at scipy.org
> >http://projects.scipy.org/mailman/listinfo/scipy-user
>   


Brandon C. Nuttall

> >Dear All,
> >I was trying to fit some data using the leastsq package in
> >scipy.optimize. The function I would like to use to fit my data is:
> >
> >log(10.0)*A1/sqrt(2.0*pi)/log(myvar1)*exp(-((log(x/mu1))**2.0)/2.0/log(myvar1)/log(myvar1)))
> >
> >  where A1, mu1 and myvar1 are fitting parameters.
> >For some reason, I used to get an error message from scipy.optimize
> >telling me that I was not working with an array of floats.
> >I suppose that this is due to the fact that the optimizer also tries
> >solving for negative values of mu1 and myvar1, for which the log
> >function (x is always positive) does not exist.
> >In fact, if I use the fitting function:
> >
> >log(10.0)*A1/sqrt(2.0*pi)/log(myvar1**2.0)*exp(-((log(x/mu1**2.0))**2.0)/2.0/log(myvar1**2.0)/log(myvar1**2.0)))
> >
> >Where mu1 and myvar1 appear squared, then the problem does not exist
> >any longer and the results are absolutely ok.
> >Can anyone enlighten me here and confirm this is what is really going on?
> >Kind Regards
> >
> >Lorenzo
> >_______________________________________________
> >SciPy-user mailing list
> >SciPy-user at scipy.org
> >http://projects.scipy.org/mailman/listinfo/scipy-user
>   


Brandon C. Nuttall




From franckm at aims.ac.za  Fri Mar  2 16:31:27 2007
From: franckm at aims.ac.za (Franck Kalala Mutombo)
Date: Fri, 02 Mar 2007 23:31:27 +0200
Subject: [SciPy-user] Integer to binary [Was: Fitting sphere to 3d data
 points]
In-Reply-To: <20070302165322.GA25987@aims.ac.za>
References: <715CC8AB-4406-41A8-83EF-72F419738896@nih.gov>	<91cf711d0701250631r35a9602g8f8ed48979e7cde3@mail.gmail.com>	<BF49BEA8-57C4-4207-A919-EEA780576C9D@nih.gov>	<45E7F956.9050400@aims.ac.za>
	<20070302165322.GA25987@aims.ac.za>
Message-ID: <45E897AF.5010505@aims.ac.za>

Jan Groenewald wrote:
> Hi Franck
> 
> When you post, please don't reply to another thread of messages with
> a good descriptive subject. Your message not only gets lost by any
> email client that sorts by threads (mails with the same subject matter)
> it also annoys some readers who use this feature.
> 
> Please start a blank message if you have a new question, and give it
> a descriptive subject.
> 
> On Fri, Mar 02, 2007 at 12:15:50PM +0200, Franck Kalala Mutombo wrote:
>> I have a sequence of decimal number (1,2,...16) for example, I want if
>> there is any function  which can convert each one in binary.
> 
> A quick google of int2bin+python found this:
> 
> def int2bin(num, width=32):
>     return ''.join(['%c'%(ord('0')+bool((1<<k)&num)) for k in range((width-1),-1,-1)])
> 
> cheers,
> Jan
Dear Jan

Thank you for your advices and for your help.

cheers

Franck

-- 
Franck
African Institute for Mathematical Sciences -- www.aims.ac.za


From wbaxter at gmail.com  Fri Mar  2 16:43:38 2007
From: wbaxter at gmail.com (Bill Baxter)
Date: Sat, 3 Mar 2007 06:43:38 +0900
Subject: [SciPy-user] [Numpy-discussion]  NumPy in Teaching
In-Reply-To: <45E7060A.9040105@astraw.com>
References: <200703010032.l210WfhI005995@glup.physics.ucf.edu>
	<D54CE207-CAB6-464C-8A3B-38E1D2154FFB@stsci.edu>
	<45E7060A.9040105@astraw.com>
Message-ID: <e86a5fd00703021343u3d6ffb92la4031cb012a9c7ca@mail.gmail.com>

On 3/2/07, Andrew Straw <strawman at astraw.com> wrote:
> Perry Greenfield wrote:
> > On Feb 28, 2007, at 7:32 PM, Joe Harrington wrote:
> >
> >
> >
> >
> I believe the reference is to this: http://scipy.org/NumPy_for_Matlab_Users

That page also has this link at the bottom:
http://37mm.no/matlab-python-xref.html

That page contains a Numpy command cross reference for Matlab IDL and R.

--bb


From philfarm at fastmail.fm  Fri Mar  2 23:27:56 2007
From: philfarm at fastmail.fm (philfarm at fastmail.fm)
Date: Fri, 02 Mar 2007 22:27:56 -0600
Subject: [SciPy-user] Best way to structure dynamic system-of-systems
	simulations?
Message-ID: <1172896076.28906.1177490111@webmail.messagingengine.com>

I'm looking for any pointers on the best way to to structure the
simulation of a system composed of subsystems, i.e., I want to build my
system model as an assemblage of component models, where some components
may use as U values the Y values from other components. Each of the
components can be modeled in state-space form, i.e.:

X_dot = A(X,U,t)
    Y = B(X,U,t)

where A and B are matrix functions (not necessarily linear).

Is anyone using any sort of framework to simplify this task? What's the
best approach?

If it makes any difference, I've been using odeint.

Thanks,
Phil

-- 
http://www.fastmail.fm - And now for something completely different




From ckkart at hoc.net  Sat Mar  3 00:53:40 2007
From: ckkart at hoc.net (Christian Kristukat)
Date: Sat, 03 Mar 2007 14:53:40 +0900
Subject: [SciPy-user] Optimization Working only with a Specific
 Expression of the input Parameters
In-Reply-To: <45E87B5B.4000703@gmail.com>
References: <45E87B5B.4000703@gmail.com>
Message-ID: <45E90D64.7040304@hoc.net>

Lorenzo Isella wrote:
> Hi Brandon,
> Thanks for your advice, but I am a bit confused: myvar1 is simply a 
> fitting parameter (i.e. it is used to return an output), nothing is 
> stored in it to start from.
 >
 > It is not an array. Furthermore, I have to say that if I define the
> error function as the absolute difference between my data and the 
> function I want to use for the fitting, then the code executes even 
> without raising any parameter to the second power, but returns nonsense 
> (negative variance and so on).
> Instead, without the abs(), I still get the same problem mentioned in my 
> previous email.

Can you post your code, including the data?

Christian


From robert.vergnes at yahoo.fr  Sat Mar  3 04:05:40 2007
From: robert.vergnes at yahoo.fr (Robert VERGNES)
Date: Sat, 3 Mar 2007 10:05:40 +0100 (CET)
Subject: [SciPy-user] RE : Re:  odeint-lsoda
In-Reply-To: <es9kck$hjp$1@sea.gmane.org>
Message-ID: <20070303090540.29592.qmail@web27408.mail.ukl.yahoo.com>

I looked in the site package before but can't see anything ..
(C:\Python25\Lib\site-packages\scipy\integrate)
All tests from the tests folder work, but 
I don't understand how the odepack links to the odeint (albeit i see the def...). And can't find any file lsoda.py (or .f)

I probably missed something...?


Donald Fredkin <drfredkin at ucsd.edu> a ?crit : Robert VERGNES wrote:

> is there a manual for odeint and lsoda ?

You can view the source code for lsoda on netlib. The extensive
comments at the head of the source file are clear and complete. Odeint
is just a wrapper for lsoda, so when you understand what lsoda is doing
you will be all set.

In your case, you might simply be computing for to coarse a mesh in t,
so "too much work" has to be done for each step. The error message
tells you what value of t was reached before lsoda gave up.

-- 


_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user


 		
---------------------------------
 D?couvrez une nouvelle fa?on d'obtenir des r?ponses ? toutes vos questions ! Profitez des connaissances, des opinions et des exp?riences des internautes sur Yahoo! Questions/R?ponses.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070303/e55a1eab/attachment.html>

From lorenzo.isella at gmail.com  Sat Mar  3 15:23:16 2007
From: lorenzo.isella at gmail.com (Lorenzo Isella)
Date: Sat, 03 Mar 2007 21:23:16 +0100
Subject: [SciPy-user] Optimization Working only with a Specific
 Expression of the input Parameters
Message-ID: <45E9D934.20202@gmail.com>

Can you post your code, including the data?

Christian



Hi Cristian,
Actually I already posted this on the mailing list, but nevertheless I 
will give it another try.
Now, this is the code I fixed after posting on the mailing list:



#! /usr/bin/env python
from scipy import *
import pylab  # used to read the .csv file

# now I want to try reading some .csv file
data = pylab.load("120km-TPM.csv",delimiter=',')

vecdim=shape(data) # now I introduce a vector with the dimensions of 
data, the file I read
print "the dimensions of data are"
print vecdim  # now very careful! in Python the arrays start with index 
zero.
x=data[0:vecdim[0],0] # it means: slice the rows from the 1st one (0) to 
the last one (
# (vecdim[0]) for the first column (0)

#plot(x,data[:,1])
#show() # uncomment them to plot a distribution

# 1st problem: if uncomment the previous two lines, I get a warning and 
until I close
# the window, the script does not progress.


y_meas=data[0:vecdim[0],1]  # measured data, for example the 2nd column 
of the .csv file

# Here I define my own error function i.e. the function I want to minimize

def residuals(p, y, x):
    A1,mu1,myvar1 = p
    err = abs( 
y-log(10.0)*A1/sqrt(2.0*pi)/log(abs(myvar1))*exp(-((log(x/abs(mu1)))**2.0)/2.0/log(abs(myvar1))/log(abs(myvar1))))
    return err

#def peval(x, p):
#    return  
log(10.0)*p[0]/sqrt(2.0*pi)/log(abs(p[2]))*exp(-((log(x/abs(p[1])))**2.0)/2.0/log(abs(p[2]))/log(abs(p[2])))

# NB: I am using the mean as a mu1**2. and the std as myvar1**2.
# otherwise I run into problems (probably the optimizers tries some 
negative values of the
# mean and stops as it sees a complex number and it stops).


print "the initial guesses are"
p0 = [50000.0,90.0, 1.59]
print array(p0)

# now I try performing a least-square fitting
from scipy.optimize import leastsq

# now I try actually solving the problem
#print "ok up to here"
plsq = leastsq(residuals, p0, args=(y_meas, x))

#print "ok up to here2"

print "the optimized values of the parameters are:"

print plsq[0]

coeff=plsq[0]
#coeff[1]=coeff[1]**2.
#coeff[2]=coeff[2]**2.

print "so the amplitude is", coeff[0]

print "and the mean", coeff[1]

print "and the std", coeff[2]

print"and these are the same results as those provided by R!"

print "So far so good"




This way the code works, but if you try running it without the abs() 
around mu1 and myvar1, then you get again the error message concerning 
the floating point numbers.
I also (re)attach the data, which you will be able to access via a link.
Kind Regards

Lorenzo
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 120km-TPM.csv
Type: text/csv
Size: 9696 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070303/aa1139b1/attachment.csv>

From ckkart at hoc.net  Sat Mar  3 19:17:51 2007
From: ckkart at hoc.net (Christian Kristukat)
Date: Sun, 04 Mar 2007 09:17:51 +0900
Subject: [SciPy-user] Optimization Working only with a Specific
 Expression of the input Parameters
In-Reply-To: <45E9D934.20202@gmail.com>
References: <45E9D934.20202@gmail.com>
Message-ID: <45EA102F.9070505@hoc.net>

Hi Lorenzo,

Lorenzo Isella wrote:
> Actually I already posted this on the mailing list, but nevertheless I
> will give it another try.

Sorry I have not seen it.
The problem is as you guessed that leastsq is trying negative parameters which
will turn the residual complex. But the reason is simple and easy to solve: your
initital guess for the first parameters is bad:

> print "the initial guesses are"
> p0 = [50000.0,90.0, 1.59]

If you set it to 1e8 it works very well.

Christian


From gael.varoquaux at normalesup.org  Sun Mar  4 14:27:39 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Sun, 4 Mar 2007 20:27:39 +0100
Subject: [SciPy-user] TraitsUI tutorial
Message-ID: <20070304192738.GK1070@clipper.ens.fr>

TraitsUI is a package to create user interfaces by building dialogs from
objects. It actually does much more and provides tools for dynamical
notifications and is thus a good framework for data-flow programming and
inversion of control. As it is based on WxPython it is portable and
integrates in other GUIs (it can be used in an application with
matplotlib, for instance).

I have used it very successfully to write a software to control a lab
experiment. I find that it is the easiest way to build GUI under python.

I had a few difficulties in the beginning, due my ignorance of basics
facts in GUI programming. I have writing a tutorial that targets the
casual programmer, hopping to leverage traitsUI to just about any
scientist that uses python. The tutorial has been mentioned a few times
on the ML, but this time I have made the last modifications I wanted to,
so I announce the first version that I am not ashamed of:
http://www.gael-varoquaux.info/computers/traits_tutorial/index.html.
I have also made a cookbook entry: http://scipy.org/TraitsUI

If you agree with me that this is the easiest way to build graphical
interfaces under python, please help me spread the word.

Cheers,

Ga?l


From davbrow at gmail.com  Mon Mar  5 03:43:12 2007
From: davbrow at gmail.com (Dave)
Date: Mon, 5 Mar 2007 00:43:12 -0800
Subject: [SciPy-user] OSX _fitpack import problem
Message-ID: <588f52020703050043g9cfd772kc4dbb30ef27cfbe1@mail.gmail.com>

I have not been able to build a completely working scipy for python
2.5 on OS X (PPC).  Much of it seems to work but I get an error when
trying to import scipy.interpolate:

/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/scipy/interpolate/fitpack.py
in <module>()
     32     'bisplrep', 'bisplev']
     33 __version__ = "$Revision: 2252 $"[10:-1]
---> 34 import _fitpack
     35 from numpy import atleast_1d, array, ones, zeros, sqrt, ravel,
transpose, \
     36      dot, sin, cos, pi, arange, empty, int32

<type 'exceptions.ImportError'>:
dlopen(/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/scipy/interpolate/_fitpack.so,
2): Symbol not found: _e_wsle
  Referenced from:
/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/scipy/interpolate/_fitpack.so
  Expected in: dynamic lookup


I installed using instructions from the scipy web site
(http://www.scipy.org/Installing_SciPy/Mac_OS_X).  Here's additional
info.

gfortran --version
GNU Fortran 95 (GCC) 4.3.0 20061223 (experimental)

gcc --version
powerpc-apple-darwin8-gcc-4.0.1 (GCC) 4.0.1 (Apple Computer, Inc. build 5367)

I installed scipy from the tar archive (scipy-0.5.2), not svn.

In searching I found postings where others have had this problem over
the last few months but I didn't find a solution.  Does anyone know
what's going on and how this can be resolved?

-- David


From giorgio.luciano at chimica.unige.it  Mon Mar  5 05:50:42 2007
From: giorgio.luciano at chimica.unige.it (Giorgio Luciano)
Date: Mon, 05 Mar 2007 11:50:42 +0100
Subject: [SciPy-user] new repository for free software and forum community
	started
Message-ID: <45EBF602.6010902@chimica.unige.it>

Dear All
I'm happy and proud to announce that a repository for free chemometric 
software (in particular Python oriented) has started.
at www.chemometrics.it

Some time ago I've  sent an email to the ICS  asking if anybody knows 
the esistence of a python repository for chemometrics,
I had a positive feedback from the community and with the help of other 
chemometricians I've decided to  gather  free
software available for doing common chemometric tasks.

In the site you will also find a forum where to discuss

-> theoretical aspects of chemometrics
-> job opportunities and cooperation with other chemometricians
-> software request/news


I've tried to link to existing sites that already give a huge contribute 
to chemometrics (KVL,Chemometrics.se just to cite two of them)
but I will be very glad if you would like to contribute with your own links

I hope this initiave could help to spread chemometrics and the use of 
free routine/software for chemometrics but that
it also will become a place of discussion for anybody interested.
Obviously we need your help and enthusiasm.

If you would like to upload software links, routines etc, just register 
for free and do it, it's just so easy and you are very encouraged to do it.

The initiave is just at its beginning so feel free to report any 
technical problem
Any kind of feedback is appreciated


Since it's a no profit personal initiave (I bought the domain and the 
rent the hosting)  finally let me say that any mirroring or hosting  
will be greatly appreciated.
I hope that this initiative will be useful ;)

Best Regards

Giorgio Luciano

-- 
-======================-
Dr Giorgio Luciano Ph.D.
Di.C.T.F.A. Dipartimento di Chimica e Tecnologie Farmaceutiche e Alimentari
Via Brigata Salerno (ponte) - 16147 Genova (GE) - Italy
email luciano at dictfa.unige.it
-======================-


From gonzalezmancera+scipy at gmail.com  Mon Mar  5 13:30:32 2007
From: gonzalezmancera+scipy at gmail.com (Andres Gonzalez-Mancera)
Date: Mon, 5 Mar 2007 13:30:32 -0500
Subject: [SciPy-user]  OSX _fitpack import problem
Message-ID: <c990a6a50703051030qd346ba4maa2df83d9e3e1e0a@mail.gmail.com>

David, I'm not an expert on the subject but I have Scipy installed on
Mac os 10.4.8. I wasn't successful using python 2.5 and so I
downgraded to 2.4 and everything builds out of the box with no
problems following the instructions. I don't know if the SVN version
plays nicer with 2.5 but the release version which I also used didn't.
Also remember to use:

sudo python setup.py config_fc --fcompiler=gnu95 build

if you have both g77 and gfortran installed.

I hope this helps,

Andres

On 3/5/07, scipy-user-request at scipy.org <scipy-user-request at scipy.org> wrote:
>
> Message: 2
> Date: Mon, 5 Mar 2007 00:43:12 -0800
> From: Dave <davbrow at gmail.com>
> Subject: [SciPy-user] OSX _fitpack import problem
> To: scipy-user at scipy.org
> Message-ID:
>         <588f52020703050043g9cfd772kc4dbb30ef27cfbe1 at mail.gmail.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
> I have not been able to build a completely working scipy for python
> 2.5 on OS X (PPC).  Much of it seems to work but I get an error when
> trying to import scipy.interpolate:
>
> /Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/scipy/interpolate/fitpack.py
> in <module>()
>      32     'bisplrep', 'bisplev']
>      33 __version__ = "$Revision: 2252 $"[10:-1]
> ---> 34 import _fitpack
>      35 from numpy import atleast_1d, array, ones, zeros, sqrt, ravel,
> transpose, \
>      36      dot, sin, cos, pi, arange, empty, int32
>
> <type 'exceptions.ImportError'>:
> dlopen(/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/scipy/interpolate/_fitpack.so,
> 2): Symbol not found: _e_wsle
>   Referenced from:
> /Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/scipy/interpolate/_fitpack.so
>   Expected in: dynamic lookup
>
>
> I installed using instructions from the scipy web site
> (http://www.scipy.org/Installing_SciPy/Mac_OS_X).  Here's additional
> info.
>
> gfortran --version
> GNU Fortran 95 (GCC) 4.3.0 20061223 (experimental)
>
> gcc --version
> powerpc-apple-darwin8-gcc-4.0.1 (GCC) 4.0.1 (Apple Computer, Inc. build 5367)
>
> I installed scipy from the tar archive (scipy-0.5.2), not svn.
>
> In searching I found postings where others have had this problem over
> the last few months but I didn't find a solution.  Does anyone know
> what's going on and how this can be resolved?
>
> -- David
>
>
> ------------------------------

-- 
Andres Gonzalez-Mancera
Biofluid Mechanics Lab
Department of Mechanical Engineering
University of Maryland, Baltimore County
andres.gonzalez at umbc.edu
410-455-3347


From steveire at gmail.com  Mon Mar  5 15:24:15 2007
From: steveire at gmail.com (Stephen Kelly)
Date: Mon, 5 Mar 2007 20:24:15 +0000
Subject: [SciPy-user] Arrayfns in numpy?
Message-ID: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>

Hi,

I attempted to post this to the numpy-discussion list, but it seems that it
has not appeared.

I'm working on a project that requires interpolation, and I found this
post (http://mail.python.org/pipermail/python-list/2000-August/050462.html
) which works fine, but depends on Numeric and does not seem to be available
on the latest numpy.

Is there some other way I should be doing interpolation using numpy,
or is the omission of arrayfns an oversight?

Kind regards,

Stephen.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070305/9bcf77d4/attachment.html>

From as8ca at virginia.edu  Mon Mar  5 15:56:44 2007
From: as8ca at virginia.edu (Alok Singhal)
Date: Mon, 5 Mar 2007 15:56:44 -0500
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
Message-ID: <20070305205644.GA18701@virginia.edu>

On 05/03/07: 20:24, Stephen Kelly wrote:
>    Is there some other way I should be doing interpolation using numpy,
>    or is the omission of arrayfns an oversight?

You could use the functions in scipy.interpolate.  For linear
interpolation, use something like:

from scipy.interpolate import interp1d

# x, y are the original data values
intp = interp1d(x, y)

# xx is an array of new x values for which to interpolate (or just a
# scalar)
yy = intp(xx)

If you can't use scipy, then it shouldn't be too hard to write your
own linear interpolation function.  You might find searchsorted()
function useful in that case (in case your original x data is not
regularly spaced).

-Alok

-- 
Alok Singhal                               *   *          
Graduate Student, dept. of Astronomy   *           *     *
University of Virginia                                    
http://www.astro.virginia.edu/~as8ca/              *    * 


From drfredkin at ucsd.edu  Mon Mar  5 19:49:11 2007
From: drfredkin at ucsd.edu (Donald Fredkin)
Date: Tue, 6 Mar 2007 00:49:11 +0000 (UTC)
Subject: [SciPy-user] RE : Re: odeint-lsoda
References: <es9kck$hjp$1@sea.gmane.org>
	<20070303090540.29592.qmail@web27408.mail.ukl.yahoo.com>
Message-ID: <esidq6$h22$1@sea.gmane.org>

You can find the source for lsoda at http://netlib.org/. It's good to
become accustomed to searching in netlib, but you can "cheat" and go
directly to the source at http://netlib.org/alliant/ode/prog/lsoda.f.

Robert VERGNES wrote:

> I probably missed something...?
> 
> 
> Donald Fredkin <drfredkin at ucsd.edu> a icrit : Robert VERGNES wrote:
> 
> > is there a manual for odeint and lsoda ?
> 
> You can view the source code for lsoda on netlib. The extensive
> comments at the head of the source file are clear and complete. Odeint
> is just a wrapper for lsoda, so when you understand what lsoda is
> doing you will be all set.
> 
> In your case, you might simply be computing for to coarse a mesh in t,
> so "too much work" has to be done for each step. The error message
> tells you what value of t was reached before lsoda gave up.



-- 




From steveire at gmail.com  Tue Mar  6 05:54:19 2007
From: steveire at gmail.com (Stephen Kelly)
Date: Tue, 6 Mar 2007 10:54:19 +0000
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <20070305205644.GA18701@virginia.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
Message-ID: <18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>

Hi,

Thanks for the reply.

I don't want to depend on scipy for a small task like this. I want to keep
dependencies small.

I looked at the searchsorted function, but don't see how it would be useful.
My data is not regularly spaced. It is X-ray diffraction data in which each
intensity has been shifted slightly. I am currently using arrayfns from
Numeric to get data with regular spacing. Could you give more information on
how to interpolate the data? I don't know where to start.

As an aside, here's some things in Numeric that aren't in numpy. Is this
oversight, or are there no plans to implement them?
1. arrayfns module
2. UserArray module with UserArray class (comparible to UserDict, UserList).

Kind regards,

Stephen.

On 3/5/07, Alok Singhal <as8ca at virginia.edu> wrote:
>
> On 05/03/07: 20:24, Stephen Kelly wrote:
> >    Is there some other way I should be doing interpolation using numpy,
> >    or is the omission of arrayfns an oversight?
>
> You could use the functions in scipy.interpolate.  For linear
> interpolation, use something like:
>
> from scipy.interpolate import interp1d
>
> # x, y are the original data values
> intp = interp1d(x, y)
>
> # xx is an array of new x values for which to interpolate (or just a
> # scalar)
> yy = intp(xx)
>
> If you can't use scipy, then it shouldn't be too hard to write your
> own linear interpolation function.  You might find searchsorted()
> function useful in that case (in case your original x data is not
> regularly spaced).
>
> -Alok
>
> --
> Alok Singhal                               *   *
> Graduate Student, dept. of Astronomy   *           *     *
> University of Virginia
> http://www.astro.virginia.edu/~as8ca/              *    *
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070306/7407dd8c/attachment.html>

From steveire at gmail.com  Tue Mar  6 05:54:19 2007
From: steveire at gmail.com (Stephen Kelly)
Date: Tue, 6 Mar 2007 10:54:19 +0000
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <20070305205644.GA18701@virginia.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
Message-ID: <18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>

Hi,

Thanks for the reply.

I don't want to depend on scipy for a small task like this. I want to keep
dependencies small.

I looked at the searchsorted function, but don't see how it would be useful.
My data is not regularly spaced. It is X-ray diffraction data in which each
intensity has been shifted slightly. I am currently using arrayfns from
Numeric to get data with regular spacing. Could you give more information on
how to interpolate the data? I don't know where to start.

As an aside, here's some things in Numeric that aren't in numpy. Is this
oversight, or are there no plans to implement them?
1. arrayfns module
2. UserArray module with UserArray class (comparible to UserDict, UserList).

Kind regards,

Stephen.

On 3/5/07, Alok Singhal <as8ca at virginia.edu> wrote:
>
> On 05/03/07: 20:24, Stephen Kelly wrote:
> >    Is there some other way I should be doing interpolation using numpy,
> >    or is the omission of arrayfns an oversight?
>
> You could use the functions in scipy.interpolate.  For linear
> interpolation, use something like:
>
> from scipy.interpolate import interp1d
>
> # x, y are the original data values
> intp = interp1d(x, y)
>
> # xx is an array of new x values for which to interpolate (or just a
> # scalar)
> yy = intp(xx)
>
> If you can't use scipy, then it shouldn't be too hard to write your
> own linear interpolation function.  You might find searchsorted()
> function useful in that case (in case your original x data is not
> regularly spaced).
>
> -Alok
>
> --
> Alok Singhal                               *   *
> Graduate Student, dept. of Astronomy   *           *     *
> University of Virginia
> http://www.astro.virginia.edu/~as8ca/              *    *
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070306/7407dd8c/attachment-0001.html>

From jelle.feringa at ezct.net  Tue Mar  6 07:18:40 2007
From: jelle.feringa at ezct.net (Jelle Feringa / EZCT Architecture & Design Research)
Date: Tue, 6 Mar 2007 13:18:40 +0100
Subject: [SciPy-user] tvtk GL2PS
Message-ID: <001501c75fe9$942463e0$c000a8c0@JELLE>

Hi,

 

Currently I'm using the enthon (win32, 2.4) 1.0.0 release, which is
terrific!

Though there is a minor, but to me important glitch in (t)vtk; 

 

>>>window

<enthought.tvtk.tools.ivtk.IVTKWithCrustAndBrowser object at 0x02840D80>

>>>window.scene.save_gl2ps('test.pdf')

Saving as a vector PS/EPS/PDF/TeX file using GL2PS is either not supported
by your version of VTK or you have not configured VTK to work with GL2PS --
read the documentation for the vtkGL2PSExporter class.

 

Which is really too bad. 

 

 

-jelle

 

 

 

Ps:

 

If you have ran into the same issues and have been able to recompile vtk
with the VTK_USE_GL2PS flag enabled, it'd be great if you would be willing
to share the binaries with me! I know
http://mayavi.sourceforge.net/cgi-bin/moin.cgi/BuildingVTKOnWin32 has
precise building instructions, but I'm afraid I'd need to confess I'm
compiler-challenged ;')

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070306/0171b5f7/attachment.html>

From as8ca at virginia.edu  Tue Mar  6 08:19:20 2007
From: as8ca at virginia.edu (Alok Singhal)
Date: Tue, 6 Mar 2007 08:19:20 -0500
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
Message-ID: <20070306131920.GA11989@virginia.edu>

On 06/03/07: 10:54, Stephen Kelly wrote:
>    I don't want to depend on scipy for a small task like this. I want to keep
>    dependencies small.
> 
>    I looked at the searchsorted function, but don't see how it would be
>    useful.

Given a sorted arry a, and some values v,

ind = searchsorted(a, v)

returns ind such that a[ind[i]] > v[i] > a[ind[i]-1] (I hope I got
that right :-) ).


>    My data is not regularly spaced. It is X-ray diffraction data in
>    which each intensity has been shifted slightly. I am currently using
>    arrayfns from Numeric to get data with regular spacing. Could you give
>    more information on how to interpolate the data? I don't know where to
>    start.

Here is how I would go about it:

import numpy

# Test data (non-uniform spacing in x)
x = numpy.array([0.0, 3.5, 6.0, 10.0, 15.0, 25.0])
y = numpy.array([13.0, 6.0, -6.0, 0.0, 4.0, 18.0])

# The data values for which interpolation is required.  xx[0] should
# be > x[0], and xx[-1] should be < x[-1].  Otherwise, undefined
# behavior.
xx = numpy.mgrid[0.5:25:0.5]

# High indices
hi = numpy.searchsorted(x, xx)

# Low indices
lo = hi - 1

slopes = (y[hi] - y[lo])/(x[hi] - x[lo])

# Interpolated data
yy = y[lo] + slopes*(xx - x[lo])

# To plot using matplotlib:
import pylab
pylab.plot(x, y, 'ro', xx, yy, 'b')
pylab.show()


>    As an aside, here's some things in Numeric that aren't in numpy. Is this
>    oversight, or are there no plans to implement them?
>    1. arrayfns module
>    2. UserArray module with UserArray class (comparible to UserDict,
>    UserList).

I don't know the answer to that - maybe interpolation etc., are more
suited to be in a 'scientific library' than an 'array library', so the
creators of scipy/numpy moved the functionality of those modules to
scipy and removed the modules from numpy?  I don't know what else
(except interpolation) did the UserArray class/module have, so I can't
say for sure.

-Alok

-- 
Alok Singhal                               *   *          
Graduate Student, dept. of Astronomy   *           *     *
University of Virginia                                    
http://www.astro.virginia.edu/~as8ca/              *    * 


From mathias.wagner at physik.tu-darmstadt.de  Tue Mar  6 09:20:12 2007
From: mathias.wagner at physik.tu-darmstadt.de (Mathias Wagner)
Date: Tue, 6 Mar 2007 15:20:12 +0100
Subject: [SciPy-user] optimize.fmin_cg
Message-ID: <200703061520.14440.mathias.wagner@physik.tu-darmstadt.de>

Hi,

I use optimize.fmin_cg to find the minimum of 1-dimensional and n-dimensional 
functions.

For 1-dimensional functions everything is fine, but for n-dimensional 
functions I get a strange result.

The function returns an array-of-array-of-arry depending on the number of 
iterations, for example I get 

In [6]:scipy.optimize.fmin_ncg(a.potstmu, [60.0] , 
a.grad_potstmu,args=(10,0,0),disp=1)
Optimization terminated successfully.
         Current function value: -3464534650.284432
         Iterations: 3
         Function evaluations: 51
         Gradient evaluations: 9
         Hessian evaluations: 0
Out[6]:array([[[[[[[[[[[ 121.31834513]]]]]]]]]]])


The function potstmu is quite complicated, I am still searching for an easier 
example which needs more than 1 iteration.
Can anyone confirm this behavior?
Or is it a problem with my function?


Mathias


From nwagner at iam.uni-stuttgart.de  Tue Mar  6 09:32:51 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Tue, 06 Mar 2007 15:32:51 +0100
Subject: [SciPy-user] optimize.fmin_cg
In-Reply-To: <200703061520.14440.mathias.wagner@physik.tu-darmstadt.de>
References: <200703061520.14440.mathias.wagner@physik.tu-darmstadt.de>
Message-ID: <45ED7B93.6020803@iam.uni-stuttgart.de>

Mathias Wagner wrote:
> Hi,
>
> I use optimize.fmin_cg to find the minimum of 1-dimensional and n-dimensional 
> functions.
>
> For 1-dimensional functions everything is fine, but for n-dimensional 
> functions I get a strange result.
>
> The function returns an array-of-array-of-arry depending on the number of 
> iterations, for example I get 
>
> In [6]:scipy.optimize.fmin_ncg(a.potstmu, [60.0] , 
> a.grad_potstmu,args=(10,0,0),disp=1)
> Optimization terminated successfully.
>          Current function value: -3464534650.284432
>          Iterations: 3
>          Function evaluations: 51
>          Gradient evaluations: 9
>          Hessian evaluations: 0
> Out[6]:array([[[[[[[[[[[ 121.31834513]]]]]]]]]]])
>
>
> The function potstmu is quite complicated, I am still searching for an easier 
> example which needs more than 1 iteration.
> Can anyone confirm this behavior?
> Or is it a problem with my function?
>
>
> Mathias
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   
Hi Mathias,

Maybe I am missing something, but you can't optimize vector valued
functions.

For example you can optimize the Rayleigh quotient

R(v) = dot(v.T,A*v)/dot(v.T,B*v)  \quad    R \in \mathds{R},\quad v \in
\mathds{R}^n

where v is a vector. Please can you expand on your function potstmu.
 
Nils



From mathias.wagner at physik.tu-darmstadt.de  Tue Mar  6 09:45:18 2007
From: mathias.wagner at physik.tu-darmstadt.de (Mathias Wagner)
Date: Tue, 6 Mar 2007 15:45:18 +0100
Subject: [SciPy-user] optimize.fmin_cg
In-Reply-To: <45ED7B93.6020803@iam.uni-stuttgart.de>
References: <200703061520.14440.mathias.wagner@physik.tu-darmstadt.de>
	<45ED7B93.6020803@iam.uni-stuttgart.de>
Message-ID: <200703061545.19971.mathias.wagner@physik.tu-darmstadt.de>

Hi,

potstmu is a scalar function. The optimization works, it is just that the 
output is not

array([xmin]) 

as expect and also works for a scalar function that is optimized with respect 
to a n-dimensional (n>1) argument. In this case the result is

array([xmin0,xmin1,...])

I will try to find a simple example for which I can also post the function I 
want to minimize.

Mathias


On Tuesday 06 March 2007 15:32, you wrote:
> Mathias Wagner wrote:
> > Hi,
> >
> > I use optimize.fmin_cg to find the minimum of 1-dimensional and
> > n-dimensional functions.
> >
> > For 1-dimensional functions everything is fine, but for n-dimensional
> > functions I get a strange result.
> >
> > The function returns an array-of-array-of-arry depending on the number of
> > iterations, for example I get
> >
> > In [6]:scipy.optimize.fmin_ncg(a.potstmu, [60.0] ,
> > a.grad_potstmu,args=(10,0,0),disp=1)
> > Optimization terminated successfully.
> >          Current function value: -3464534650.284432
> >          Iterations: 3
> >          Function evaluations: 51
> >          Gradient evaluations: 9
> >          Hessian evaluations: 0
> > Out[6]:array([[[[[[[[[[[ 121.31834513]]]]]]]]]]])
> >
> >
> > The function potstmu is quite complicated, I am still searching for an
> > easier example which needs more than 1 iteration.
> > Can anyone confirm this behavior?
> > Or is it a problem with my function?
> >
> >
> > Mathias
> > _______________________________________________
> > SciPy-user mailing list
> > SciPy-user at scipy.org
> > http://projects.scipy.org/mailman/listinfo/scipy-user
>
> Hi Mathias,
>
> Maybe I am missing something, but you can't optimize vector valued
> functions.
>
> For example you can optimize the Rayleigh quotient
>
> R(v) = dot(v.T,A*v)/dot(v.T,B*v)  \quad    R \in \mathds{R},\quad v \in
> \mathds{R}^n
>
> where v is a vector. Please can you expand on your function potstmu.
>
> Nils

-- 
// ***************************************************************
// ** Mathias Wagner                                            **
// ** Institut fuer Kernphysik, TU Darmstadt                    **
// ** Schlossgartenstr. 9, 64289 Darmstadt, Germany             **
// **                                                           **
// ** email: mathias.wagner at physik.tu-darmstadt.de              **
// ** www  : http://crunch.ikp.physik.tu-darmstadt.de/~wagner   **
// ***************************************************************


From oliphant at ee.byu.edu  Tue Mar  6 11:19:22 2007
From: oliphant at ee.byu.edu (Travis Oliphant)
Date: Tue, 06 Mar 2007 09:19:22 -0700
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
Message-ID: <45ED948A.2040708@ee.byu.edu>

Stephen Kelly wrote:
> Hi,
>
> Thanks for the reply.
>
> I don't want to depend on scipy for a small task like this. I want to 
> keep dependencies small.
>
> I looked at the searchsorted function, but don't see how it would be 
> useful. My data is not regularly spaced. It is X-ray diffraction data 
> in which each intensity has been shifted slightly. I am currently 
> using arrayfns from Numeric to get data with regular spacing. Could 
> you give more information on how to interpolate the data? I don't know 
> where to start.
>
> As an aside, here's some things in Numeric that aren't in numpy. Is 
> this oversight, or are there no plans to implement them?
> 1. arrayfns module
> 2. UserArray module with UserArray class (comparible to UserDict, 
> UserList).
>

The UserArray module is there (it is imported when you import numpy as 
numpy.lib.user_array)

The UserArray class is called numpy.lib.user_array.container

The arrayfns module was not an oversight.  Some of the arrayfns module 
has been implemented except for functions that belong in SciPy.

The trend has been to move more things out of NumPy and make SciPy more 
modular so that pieces can be installed separately.  

There will always be people who want something moved into NumPy.  It is 
hard to know where to stop.

Which function in arrayfns are you missing?

-Travis



From steveire at gmail.com  Tue Mar  6 11:25:13 2007
From: steveire at gmail.com (Stephen Kelly)
Date: Tue, 6 Mar 2007 17:25:13 +0100
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <45ED948A.2040708@ee.byu.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
	<45ED948A.2040708@ee.byu.edu>
Message-ID: <18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>

I'm missing arrayfns.interp. It would be a shame to have to depend on either
Numeric or Scipy for such a trivial but useful function. I'll see about
implementing something like it myself. I have the module source code, but
installing C modules in python is a little beyond what I know how to do with
it.

Thanks for the UserArray info.

Kind regards,

Stephen.


On 3/6/07, Travis Oliphant <oliphant at ee.byu.edu> wrote:
>
> Stephen Kelly wrote:
> > Hi,
> >
> > Thanks for the reply.
> >
> > I don't want to depend on scipy for a small task like this. I want to
> > keep dependencies small.
> >
> > I looked at the searchsorted function, but don't see how it would be
> > useful. My data is not regularly spaced. It is X-ray diffraction data
> > in which each intensity has been shifted slightly. I am currently
> > using arrayfns from Numeric to get data with regular spacing. Could
> > you give more information on how to interpolate the data? I don't know
> > where to start.
> >
> > As an aside, here's some things in Numeric that aren't in numpy. Is
> > this oversight, or are there no plans to implement them?
> > 1. arrayfns module
> > 2. UserArray module with UserArray class (comparible to UserDict,
> > UserList).
> >
>
> The UserArray module is there (it is imported when you import numpy as
> numpy.lib.user_array)
>
> The UserArray class is called numpy.lib.user_array.container
>
> The arrayfns module was not an oversight.  Some of the arrayfns module
> has been implemented except for functions that belong in SciPy.
>
> The trend has been to move more things out of NumPy and make SciPy more
> modular so that pieces can be installed separately.
>
> There will always be people who want something moved into NumPy.  It is
> hard to know where to stop.
>
> Which function in arrayfns are you missing?
>
> -Travis
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070306/3b646f87/attachment.html>

From oliphant at ee.byu.edu  Tue Mar  6 11:43:27 2007
From: oliphant at ee.byu.edu (Travis Oliphant)
Date: Tue, 06 Mar 2007 09:43:27 -0700
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>	<20070305205644.GA18701@virginia.edu>	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>	<45ED948A.2040708@ee.byu.edu>
	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
Message-ID: <45ED9A2F.3050803@ee.byu.edu>

Stephen Kelly wrote:
> I'm missing arrayfns.interp. It would be a shame to have to depend on 
> either Numeric or Scipy for such a trivial but useful function. I'll 
> see about implementing something like it myself. I have the module 
> source code, but installing C modules in python is a little beyond 
> what I know how to do with it.
It is easy enough to just pull interp out.  Some of the functions in 
arrayfns are in numpy.lib already (in the _compiled_base.c module. 

Perhaps we could put interp in there as well (it doesn't look too big). 
  A simple 1-d interpolation would probably be a useful thing to have in 
NumPy.   What do others think?

-Travis



From beckers at orn.mpg.de  Tue Mar  6 11:56:05 2007
From: beckers at orn.mpg.de (Gabriel J.L. Beckers)
Date: Tue, 06 Mar 2007 17:56:05 +0100
Subject: [SciPy-user] mean of recarray
Message-ID: <1173200165.19821.8.camel@beckerspc>

Is there an easy way to get the mean of a record array?

There is a mean() method, but it doesn't seem to work for me.

In [1]: import numpy

In [2]: desc = numpy.dtype({'names':['a','b'],'formats':['d','d']})

In [3]: ar = numpy.array([(1.,2.),(3.,4.)],dtype=desc)

In [4]: ar.mean()
---------------------------------------------------------------------------
exceptions.TypeError                                 Traceback (most
recent call last)

/home/user/<ipython console> 

TypeError: cannot perform reduce with flexible type

I would actually be able to get the mean along the second axis, like
ar.mean(1). Can this be done in a simple way?

Cheers, Gabriel








From paul.ray at nrl.navy.mil  Tue Mar  6 12:02:19 2007
From: paul.ray at nrl.navy.mil (Paul Ray)
Date: Tue, 6 Mar 2007 12:02:19 -0500
Subject: [SciPy-user] SciPy-user Digest, Vol 43, Issue 9
In-Reply-To: <mailman.28360.1173199886.8339.scipy-user@scipy.org>
References: <mailman.28360.1173199886.8339.scipy-user@scipy.org>
Message-ID: <5F553E0C-06F7-4FBF-8117-2CEBF9AD705D@nrl.navy.mil>


On Mar 6, 2007, at 11:51 AM, scipy-user-request at scipy.org wrote:

> It is easy enough to just pull interp out.  Some of the functions  
> in arrayfns are in numpy.lib already (in the _compiled_base.c module.
> Perhaps we could put interp in there as well (it doesn't look too  
> big).  A simple 1-d interpolation would probably be a useful thing  
> to have in NumPy.   What do others think?

It sounds good to me.  I often have to regrid irregularly spaced data  
onto a uniform grid, which I use interpolate.splrep and  
interpolate.splev from scipy to do.  This is often the only  
dependency on scipy in my codes that otherwise use only numpy.  I'd  
love having a 1-d interp in numpy.

Cheers,

-- Paul





From peridot.faceted at gmail.com  Tue Mar  6 12:10:58 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Tue, 6 Mar 2007 12:10:58 -0500
Subject: [SciPy-user] SciPy-user Digest, Vol 43, Issue 9
In-Reply-To: <5F553E0C-06F7-4FBF-8117-2CEBF9AD705D@nrl.navy.mil>
References: <mailman.28360.1173199886.8339.scipy-user@scipy.org>
	<5F553E0C-06F7-4FBF-8117-2CEBF9AD705D@nrl.navy.mil>
Message-ID: <ce557a360703060910m65e27e7crd6d805547bc982ae@mail.gmail.com>

On 06/03/07, Paul Ray <paul.ray at nrl.navy.mil> wrote:
>
> On Mar 6, 2007, at 11:51 AM, scipy-user-request at scipy.org wrote:
>
> > It is easy enough to just pull interp out.  Some of the functions
> > in arrayfns are in numpy.lib already (in the _compiled_base.c module.
> > Perhaps we could put interp in there as well (it doesn't look too
> > big).  A simple 1-d interpolation would probably be a useful thing
> > to have in NumPy.   What do others think?
>
> It sounds good to me.  I often have to regrid irregularly spaced data
> onto a uniform grid, which I use interpolate.splrep and
> interpolate.splev from scipy to do.  This is often the only
> dependency on scipy in my codes that otherwise use only numpy.  I'd
> love having a 1-d interp in numpy.

Why not just use scipy? splev works fine, and numpy doesn't need to
grow any more... it seems to me that numpy should include just the
basics of working with arrays (indexing functions, ufuncs, matrix
multiplication, things like searchsorted) and all the mathematics
(ffts, linear algebra, interpolation, integration) should go in scipy.

Anne M. Archibald


From oliphant at ee.byu.edu  Tue Mar  6 12:24:43 2007
From: oliphant at ee.byu.edu (Travis Oliphant)
Date: Tue, 06 Mar 2007 10:24:43 -0700
Subject: [SciPy-user] SciPy-user Digest, Vol 43, Issue 9
In-Reply-To: <ce557a360703060910m65e27e7crd6d805547bc982ae@mail.gmail.com>
References: <mailman.28360.1173199886.8339.scipy-user@scipy.org>	<5F553E0C-06F7-4FBF-8117-2CEBF9AD705D@nrl.navy.mil>
	<ce557a360703060910m65e27e7crd6d805547bc982ae@mail.gmail.com>
Message-ID: <45EDA3DB.7030907@ee.byu.edu>

Anne Archibald wrote:
> On 06/03/07, Paul Ray <paul.ray at nrl.navy.mil> wrote:
>   
>> On Mar 6, 2007, at 11:51 AM, scipy-user-request at scipy.org wrote:
>>
>>     
>>> It is easy enough to just pull interp out.  Some of the functions
>>> in arrayfns are in numpy.lib already (in the _compiled_base.c module.
>>> Perhaps we could put interp in there as well (it doesn't look too
>>> big).  A simple 1-d interpolation would probably be a useful thing
>>> to have in NumPy.   What do others think?
>>>       
>> It sounds good to me.  I often have to regrid irregularly spaced data
>> onto a uniform grid, which I use interpolate.splrep and
>> interpolate.splev from scipy to do.  This is often the only
>> dependency on scipy in my codes that otherwise use only numpy.  I'd
>> love having a 1-d interp in numpy.
>>     
>
> Why not just use scipy? splev works fine, and numpy doesn't need to
> grow any more... it seems to me that numpy should include just the
> basics of working with arrays (indexing functions, ufuncs, matrix
> multiplication, things like searchsorted) and all the mathematics
> (ffts, linear algebra, interpolation, integration) should go in scipy.
>   
That is generally what we believe as well.  The problem is that Numeric 
already included several additional features.   Trying to maintain some 
semblance of backward compatibility is why NumPy has not shrunk even more.

Because the interp function was already in Numeric and is not that big, 
perhaps it should be added to NumPy.

-Travis



From Glen.Mabey at swri.org  Tue Mar  6 13:02:52 2007
From: Glen.Mabey at swri.org (Glen W. Mabey)
Date: Tue, 6 Mar 2007 12:02:52 -0600
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <45ED9A2F.3050803@ee.byu.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
	<45ED948A.2040708@ee.byu.edu>
	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
	<45ED9A2F.3050803@ee.byu.edu>
Message-ID: <20070306180252.GA25007@swri16wm.electro.swri.edu>

On Tue, Mar 06, 2007 at 09:43:27AM -0700, Travis Oliphant wrote:
> Perhaps we could put interp in there as well (it doesn't look too big). 
>   A simple 1-d interpolation would probably be a useful thing to have in 
> NumPy.   What do others think?

+1



From souheil.inati at nyu.edu  Tue Mar  6 14:15:15 2007
From: souheil.inati at nyu.edu (Souheil Inati)
Date: Tue, 6 Mar 2007 14:15:15 -0500
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <20070306180252.GA25007@swri16wm.electro.swri.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
	<45ED948A.2040708@ee.byu.edu>
	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
	<45ED9A2F.3050803@ee.byu.edu>
	<20070306180252.GA25007@swri16wm.electro.swri.edu>
Message-ID: <B323ED65-1903-415D-B155-D00AAD800E11@nyu.edu>

-100

I am strongly opposed to mixing computations into NumPY.  Why is 1-d  
interpolation good to have inside of NumPy? why not nd-interpolation?  
why not fft? or SVD? or some linear algebra thing?

Python is good because of it lets you have a lot of granularity.

NumPy should stick to what it's good at.  And it should be very good  
at a small number of things, namely it should be a library for  
efficient storage and access to data that is well represented by N- 
dimensional arrays.

---------------------------------

Souheil Inati, PhD
Assistant Professor
Center for Neural Science
New York University
4 Washington Place, Room 809
New York, N.Y., 10003-6621
Office: (212) 998-3741
Email: souheil.inati at nyu.edu



-Souheil

On Mar 6, 2007, at 1:02 PM, Glen W. Mabey wrote:

> On Tue, Mar 06, 2007 at 09:43:27AM -0700, Travis Oliphant wrote:
>> Perhaps we could put interp in there as well (it doesn't look too  
>> big).
>>   A simple 1-d interpolation would probably be a useful thing to  
>> have in
>> NumPy.   What do others think?
>
> +1
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user



From nmarais at sun.ac.za  Tue Mar  6 14:34:14 2007
From: nmarais at sun.ac.za (Neilen Marais)
Date: Tue, 06 Mar 2007 21:34:14 +0200
Subject: [SciPy-user] Using umfpack to calculate a incomplete LU
	factorisation (ILU)
Message-ID: <pan.2007.03.06.19.34.13.563384@sun.ac.za>

I'm trying to get an incomplete LU factorisation of a sparse matrix using
scipy.linsolve.splu. It's using umfpack as far as I can tell. I'm telling
it to drop elements using the drop_tol keyword argument, though it seems
to be having no effect:

In [135]: splu(A, drop_tol=1.0).nnz
Use minimum degree ordering on A'+A.
Out[135]: 6894814

In [136]: splu(A, drop_tol=0.00001).nnz
Use minimum degree ordering on A'+A.
Out[136]: 6894814

In [137]: splu(A, drop_tol=1000).nnz
Use minimum degree ordering on A'+A.
Out[137]: 6894814

In [138]: splu(A, drop_tol=10000000000.).nnz
Use minimum degree ordering on A'+A.
Out[138]: 6894814

In [139]: splu(A, drop_tol=100000000000000000000000.).nnz
Use minimum degree ordering on A'+A.
Out[139]: 6894814

In [140]: splu(A, drop_tol=0.000000000000000001).nnz
Use minimum degree ordering on A'+A.
Out[140]: 6894814

Am I doing or understanding something worng? I'm using a recent SVN build:

In [19]: scipy.version.version
Out[20]: '0.5.3.dev2827'

Any help, or pointers to better routines better suited to ILU generation
welcomed!

Thanks
Neilen

-- 
you know its kind of tragic 
we live in the new world
but we've lost the magic
-- Battery 9 (www.battery9.co.za)



From vallis.35530172 at bloglines.com  Tue Mar  6 14:37:31 2007
From: vallis.35530172 at bloglines.com (vallis.35530172 at bloglines.com)
Date: 6 Mar 2007 19:37:31 -0000
Subject: [SciPy-user] Arrayfns in numpy?
Message-ID: <1173209851.1509303837.21637.sendItem@bloglines.com>

+10

This is not about adding new functionality to numpy. It's about keeping
something that was in Numeric (arrayfns.interp, in particular), and that was
taken out for numpy.

Whereas porting code to numpy for Numeric is trivial
in most cases, if the code included arrayfns.interp then a dependency on scipy
must be added (which is a big deal!), or a new interpolation function must
be written from scratch.

To make it obvious that the new numpy should not
include computations, perhaps this functionality should sit in the oldnumeric
compatibility layer. Would that be satisfactory to Souheil?

Michele


--- SciPy Users List <scipy-user at scipy.org wrote:
-100
> 
> I am strongly
opposed to mixing computations into NumPY.  Why is 1-d  
> interpolation
good to have inside of NumPy? why not nd-interpolation?  
> why not fft?
or SVD? or some linear algebra thing?
> 
> Python is good because of it
lets you have a lot of granularity.
> 
> NumPy should stick to what it's
good at.  And it should be very good  
> at a small number of things, namely
it should be a library for  
> efficient storage and access to data that
is well represented by N- 
> dimensional arrays.
> 
> ---------------------------------

> 
> Souheil Inati, PhD
> Assistant Professor
> Center for Neural Science

> New York University
> 4 Washington Place, Room 809
> New York, N.Y.,
10003-6621
> Office: (212) 998-3741
> Email: souheil.inati at nyu.edu
> 
> 
> 
> -Souheil
> 
> On Mar 6, 2007, at 1:02 PM, Glen W. Mabey wrote:

> 
> > On Tue, Mar 06, 2007 at 09:43:27AM -0700, Travis Oliphant wrote:

> >> Perhaps we could put interp in there as well (it doesn't look too  

> >> big).
> >>   A simple 1-d interpolation would probably be a useful
thing to  
> >> have in
> >> NumPy.   What do others think?
> >
> > +1

> >
> > _______________________________________________
> > SciPy-user
mailing list
> > SciPy-user at scipy.org
> > http://projects.scipy.org/mailman/listinfo/scipy-user

> 
> _______________________________________________
> SciPy-user mailing
list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user

> 


From nwagner at iam.uni-stuttgart.de  Tue Mar  6 15:11:34 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Tue, 06 Mar 2007 21:11:34 +0100
Subject: [SciPy-user] Using umfpack to calculate a incomplete
	LU	factorisation (ILU)
In-Reply-To: <pan.2007.03.06.19.34.13.563384@sun.ac.za>
References: <pan.2007.03.06.19.34.13.563384@sun.ac.za>
Message-ID: <web-80047720@uni-stuttgart.de>

On Tue, 06 Mar 2007 21:34:14 +0200
  Neilen Marais <nmarais at sun.ac.za> wrote:
> I'm trying to get an incomplete LU factorisation of a 
>sparse matrix using
> scipy.linsolve.splu. It's using umfpack as far as I can 
>tell. I'm telling
> it to drop elements using the drop_tol keyword argument, 
>though it seems
> to be having no effect:
> 
> In [135]: splu(A, drop_tol=1.0).nnz
> Use minimum degree ordering on A'+A.
> Out[135]: 6894814
> 
> In [136]: splu(A, drop_tol=0.00001).nnz
> Use minimum degree ordering on A'+A.
> Out[136]: 6894814
> 
> In [137]: splu(A, drop_tol=1000).nnz
> Use minimum degree ordering on A'+A.
> Out[137]: 6894814
> 
> In [138]: splu(A, drop_tol=10000000000.).nnz
> Use minimum degree ordering on A'+A.
> Out[138]: 6894814
> 
> In [139]: splu(A, 
>drop_tol=100000000000000000000000.).nnz
> Use minimum degree ordering on A'+A.
> Out[139]: 6894814
> 
> In [140]: splu(A, drop_tol=0.000000000000000001).nnz
> Use minimum degree ordering on A'+A.
> Out[140]: 6894814
> 
> Am I doing or understanding something worng? I'm using a 
>recent SVN build:
> 
> In [19]: scipy.version.version
> Out[20]: '0.5.3.dev2827'
> 
> Any help, or pointers to better routines better suited 
>to ILU generation
> welcomed!
> 
> Thanks
> Neilen
> 
> -- 
> you know its kind of tragic 
> we live in the new world
> but we've lost the magic
> -- Battery 9 (www.battery9.co.za)
> 
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user

Neilen,

You might be interested in ILUPACK

http://www.math.tu-berlin.de/ilupack/
  
Nils


From thn at cs.utexas.edu  Tue Mar  6 16:27:42 2007
From: thn at cs.utexas.edu (Thomas Nelson)
Date: Tue, 6 Mar 2007 15:27:42 -0600 (CST)
Subject: [SciPy-user] probability tools in scipy
Message-ID: <Pine.LNX.4.64.0703061522300.20560@cofferdam.cs.utexas.edu>

I'm looking for a function something like:

normalcdf(lo,hi,mu,sigma)
that returns the probability that
lo < X < hi
where X is a normal random variable with mean = mu and standard deviation 
= sigma.  Is there something like this, or a tool to make something like 
this, in scipy?  I looked at the scipy.stats.distributions module, but as 
far as I can tell that seems to be more of a random number generator.  Am 
I correct about this?  does scipy have something like the functionality 
I'm looking for?

Thanks for your time,
Thomas N.


From jks at iki.fi  Tue Mar  6 16:37:07 2007
From: jks at iki.fi (=?iso-8859-1?Q?Jouni_K=2E_Sepp=E4nen?=)
Date: Tue, 06 Mar 2007 23:37:07 +0200
Subject: [SciPy-user] probability tools in scipy
References: <Pine.LNX.4.64.0703061522300.20560@cofferdam.cs.utexas.edu>
Message-ID: <m2r6s2awak.fsf@iki.fi>

Thomas Nelson <thn at cs.utexas.edu> writes:

> normalcdf(lo,hi,mu,sigma)
> that returns the probability that
> lo < X < hi
> where X is a normal random variable with mean = mu and standard deviation 
> = sigma. 

Look at scipy.stats.norm.cdf. I think this is one way to define your
function:

def normalcdf(lo, hi, mu, sigma):
    return numpy.dot([-1, 1], scipy.stats.norm.cdf([lo, hi], mu, sigma))

-- 
Jouni K. Sepp?nen
http://www.iki.fi/jks



From robert.kern at gmail.com  Tue Mar  6 16:40:05 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 06 Mar 2007 15:40:05 -0600
Subject: [SciPy-user] probability tools in scipy
In-Reply-To: <Pine.LNX.4.64.0703061522300.20560@cofferdam.cs.utexas.edu>
References: <Pine.LNX.4.64.0703061522300.20560@cofferdam.cs.utexas.edu>
Message-ID: <45EDDFB5.2060701@gmail.com>

Thomas Nelson wrote:
> I'm looking for a function something like:
> 
> normalcdf(lo,hi,mu,sigma)
> that returns the probability that
> lo < X < hi
> where X is a normal random variable with mean = mu and standard deviation 
> = sigma.  Is there something like this, or a tool to make something like 
> this, in scipy?  I looked at the scipy.stats.distributions module, but as 
> far as I can tell that seems to be more of a random number generator.  Am 
> I correct about this?  does scipy have something like the functionality 
> I'm looking for?

from scipy import stats

stats.norm.cdf(hi, mu, sigma) - stats.norm.cdf(lo, mu, sigma)

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From thn at cs.utexas.edu  Tue Mar  6 17:06:33 2007
From: thn at cs.utexas.edu (Thomas Nelson)
Date: Tue, 6 Mar 2007 16:06:33 -0600 (CST)
Subject: [SciPy-user] probability tools in scipy
In-Reply-To: <45EDDFB5.2060701@gmail.com>
References: <Pine.LNX.4.64.0703061522300.20560@cofferdam.cs.utexas.edu>
	<45EDDFB5.2060701@gmail.com>
Message-ID: <Pine.LNX.4.64.0703061605400.24294@capstan.cs.utexas.edu>

Thanks, this is exactly what I needed.

On Tue, 6 Mar 2007, Robert Kern wrote:

> Thomas Nelson wrote:
>> I'm looking for a function something like:
>>
>> normalcdf(lo,hi,mu,sigma)
>> that returns the probability that
>> lo < X < hi
>> where X is a normal random variable with mean = mu and standard deviation
>> = sigma.  Is there something like this, or a tool to make something like
>> this, in scipy?  I looked at the scipy.stats.distributions module, but as
>> far as I can tell that seems to be more of a random number generator.  Am
>> I correct about this?  does scipy have something like the functionality
>> I'm looking for?
>
> from scipy import stats
>
> stats.norm.cdf(hi, mu, sigma) - stats.norm.cdf(lo, mu, sigma)
>
> -- 
> Robert Kern
>
> "I have come to believe that the whole world is an enigma, a harmless enigma
> that is made terrible by our own mad attempt to interpret it as though it had
> an underlying truth."
>  -- Umberto Eco
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From chanley at stsci.edu  Tue Mar  6 21:07:58 2007
From: chanley at stsci.edu (Christopher Hanley)
Date: Tue, 06 Mar 2007 21:07:58 -0500
Subject: [SciPy-user] PyFITS 1.1 "candidate" RELEASE
Message-ID: <45EE1E7E.4050900@stsci.edu>

------------------
| PYFITS Release |
------------------

Space Telescope Science Institute is pleased to announce the
candidate release of PyFITS 1.1.  This release includes support
for both the NUMPY and NUMARRAY array packages. This software
can be downloaded at:

http://www.stsci.edu/resources/software_hardware/pyfits/Download

If you encounter bugs, please send bug reports to "help at stsci.edu".

We intend to support NUMARRAY and NUMPY simultaneously for a
transition period of no less than 6 months. Eventually, however,
support for NUMARRAY will disappear. During this period, it is
likely that new features will appear only for NUMPY. The
support for NUMARRAY will primarily be to fix serious bugs and
handle platform updates.


-----------
| Version |
-----------

Version 1.1rc1; March 6, 2007

-------------------------------
| Major Changes since v1.1b3  |
-------------------------------

    * The NUMPY version of PyFITS now supports FITS group data.

    * The NUMPY version of PyFITS now support variable length
      column tables.

    * Keyboard interrputs are now blocked during a file flush.

    * Many minor bug fixes.

-------------------------
| Software Requirements |
-------------------------

PyFITS Version 1.1rc1 REQUIRES:

       * Python 2.3 or later
       * NUMPY 1.0.1(or later) or NUMARRAY

---------------------
| Installing PyFITS |
---------------------
PyFITS 1.1rc1 is distributed as a Python distutils module.
Installation simply involves unpacking the package
and executing

% python setup.py install

to install it in Python's site-packages directory.

Alternatively the command

%python setup.py install --local="/destination/directory/"

will install PyFITS in an arbitrary directory which should
be placed on PYTHONPATH. Once numarray or numpy has been
installed, then PyFITS should be available for use under
Python.

-----------------
| Download Site |
-----------------

http://www.stsci.edu/resources/software_hardware/pyfits/Download

----------
| Usage  |
----------

Users will issue an "import pyfits" command as in the past.
However, the use of the NUMPY or NUMARRAY version of PyFITS
will be controlled by an environment variable called NUMERIX.

Set NUMERIX to 'numarray' for the NUMARRAY version of PyFITS.
Set NUMERIX to 'numpy' for the NUMPY version of pyfits.

If only one array package is installed, that package's version
of PyFITS will be imported.  If both packages are installed
the NUMERIX value is used to decide which version to import.
If no NUMERIX value is set then the NUMARRAY version of PyFITS
will be imported.

Anything else will raise an exception upon import.

---------------
| Bug Reports |
---------------

Please send all PyFITS bug reports to help at stsci.edu

------------------
| Advanced Users |
------------------

Users who would like the "bleeding" edge of PyFITS can retrieve
the software from our SUBVERSION repository hosted at:

http://astropy.scipy.org/svn/pyfits/trunk

We also provide a Trac site at:

http://projects.scipy.org/astropy/pyfits/wiki



From oliphant at ee.byu.edu  Wed Mar  7 02:13:26 2007
From: oliphant at ee.byu.edu (Travis Oliphant)
Date: Wed, 07 Mar 2007 00:13:26 -0700
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <B323ED65-1903-415D-B155-D00AAD800E11@nyu.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>	<20070305205644.GA18701@virginia.edu>	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>	<45ED948A.2040708@ee.byu.edu>	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>	<45ED9A2F.3050803@ee.byu.edu>	<20070306180252.GA25007@swri16wm.electro.swri.edu>
	<B323ED65-1903-415D-B155-D00AAD800E11@nyu.edu>
Message-ID: <45EE6616.4080306@ee.byu.edu>


> -100
>
> I am strongly opposed to mixing computations into NumPY.  Why is 1-d  
> interpolation good to have inside of NumPy? why not nd-interpolation?  
> why not fft? or SVD? or some linear algebra thing?
>   
I agree with the idea of the statement.  However, practicality may beat 
purity here because of the backward-compatibility issue.  The only 
proposal on the table is pulling in more functions from arrayfns.   I'm 
fine if they live in the oldnumeric name-space (but it will actually be 
easier to put them in the lib namespace because the arrayfnsmodule.c 
file became the _compiled_base.c file in numpy.lib. 

Perhaps they can still live in _compiled_base.c  but not be pulled in to 
numpy.lib (only a  oldnumeric.arrayfns module).

-Travis



From peridot.faceted at gmail.com  Wed Mar  7 02:10:28 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Wed, 7 Mar 2007 02:10:28 -0500
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <45EE6616.4080306@ee.byu.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
	<45ED948A.2040708@ee.byu.edu>
	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
	<45ED9A2F.3050803@ee.byu.edu>
	<20070306180252.GA25007@swri16wm.electro.swri.edu>
	<B323ED65-1903-415D-B155-D00AAD800E11@nyu.edu>
	<45EE6616.4080306@ee.byu.edu>
Message-ID: <ce557a360703062310o669e0799j5406b1a8ae0f6b23@mail.gmail.com>

On 07/03/07, Travis Oliphant <oliphant at ee.byu.edu> wrote:

> I agree with the idea of the statement.  However, practicality may beat
> purity here because of the backward-compatibility issue.  The only
> proposal on the table is pulling in more functions from arrayfns.   I'm
> fine if they live in the oldnumeric name-space (but it will actually be
> easier to put them in the lib namespace because the arrayfnsmodule.c
> file became the _compiled_base.c file in numpy.lib.
>
> Perhaps they can still live in _compiled_base.c  but not be pulled in to
> numpy.lib (only a  oldnumeric.arrayfns module).

I like this idea - can it be a general policy for things that
shouldn't be in numpy but are included for backward compatibility?

Anne


From fperez.net at gmail.com  Wed Mar  7 02:46:36 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Wed, 7 Mar 2007 00:46:36 -0700
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <45EE6616.4080306@ee.byu.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
	<45ED948A.2040708@ee.byu.edu>
	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
	<45ED9A2F.3050803@ee.byu.edu>
	<20070306180252.GA25007@swri16wm.electro.swri.edu>
	<B323ED65-1903-415D-B155-D00AAD800E11@nyu.edu>
	<45EE6616.4080306@ee.byu.edu>
Message-ID: <db6b5ecc0703062346r6eba1927o250a08278c5441d0@mail.gmail.com>

On 3/7/07, Travis Oliphant <oliphant at ee.byu.edu> wrote:
>
> > -100
> >
> > I am strongly opposed to mixing computations into NumPY.  Why is 1-d
> > interpolation good to have inside of NumPy? why not nd-interpolation?
> > why not fft? or SVD? or some linear algebra thing?
> >
> I agree with the idea of the statement.  However, practicality may beat
> purity here because of the backward-compatibility issue.  The only
> proposal on the table is pulling in more functions from arrayfns.   I'm
> fine if they live in the oldnumeric name-space (but it will actually be
> easier to put them in the lib namespace because the arrayfnsmodule.c
> file became the _compiled_base.c file in numpy.lib.
>
> Perhaps they can still live in _compiled_base.c  but not be pulled in to
> numpy.lib (only a  oldnumeric.arrayfns module).

-1.  If they are going to go in, make them first-class citizens.

For better or worse, there are functional units in numpy that go
beyond pure array basics: linalg, fft are the leading examples.   Not
putting these in would break so much old Numeric code that it really
would have been unacceptable.  So if the argument for the
interpolation stuff is the same, then it should go also as part of
numpy.

IMO, oldnumeric is a way to keep the old *APIs* for things that can be
done differently (often better) in numpy, as a way of easing the
transition of old codes.

But if you put interpolation in oldnumeric, then /everyone/ who wants
interpolation and doesn't have scipy around is going to pull
oldnumeric in.  And that will have the unintended side effect of
basically promoting oldnumeric as a first-class API rather than a
backwards-compatibility layer for codes that haven't been properly
'numpyfied'.

Cheers,

f


From fperez.net at gmail.com  Wed Mar  7 02:53:25 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Wed, 7 Mar 2007 00:53:25 -0700
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <B323ED65-1903-415D-B155-D00AAD800E11@nyu.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
	<45ED948A.2040708@ee.byu.edu>
	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
	<45ED9A2F.3050803@ee.byu.edu>
	<20070306180252.GA25007@swri16wm.electro.swri.edu>
	<B323ED65-1903-415D-B155-D00AAD800E11@nyu.edu>
Message-ID: <db6b5ecc0703062353q4a550e7bu25843ee0fa157d03@mail.gmail.com>

On 3/6/07, Souheil Inati <souheil.inati at nyu.edu> wrote:
> -100
>
> I am strongly opposed to mixing computations into NumPY.  Why is 1-d
> interpolation good to have inside of NumPy? why not nd-interpolation?
> why not fft? or SVD? or some linear algebra thing?

We're already well past that point; ffts, svds and linear algebra are
all in there already, and won't be removed anytime soon:

In [1]: import numpy as N

In [2]: N.linalg?
Type:           module
Base Class:     <type 'module'>
Namespace:      Interactive
File:
/home/fperez/tmp/local/lib/python2.4/site-packages/numpy/linalg/__init__.py
Docstring:
    Core Linear Algebra Tools
    ===========

     Linear Algebra Basics:

       norm       --- Vector or matrix norm
       inv        --- Inverse of a square matrix
       solve      --- Solve a linear system of equations
       det        --- Determinant of a square matrix
       lstsq      --- Solve linear least-squares problem
       pinv       --- Pseudo-inverse (Moore-Penrose) using lstsq

     Eigenvalues and Decompositions:

       eig        --- Eigenvalues and vectors of a square matrix
       eigh       --- Eigenvalues and eigenvectors of a Hermitian matrix
       eigvals    --- Eigenvalues of a square matrix
       eigvalsh   --- Eigenvalues of a Hermitian matrix.
       svd        --- Singular value decomposition of a matrix
       cholesky   --- Cholesky decomposition of a matrix


In [3]: N.fft?
Type:           module
Base Class:     <type 'module'>
Namespace:      Interactive
File:
/home/fperez/tmp/local/lib/python2.4/site-packages/numpy/fft/__init__.py
Docstring:
    Core FFT routines
    ==================

     Standard FFTs

       fft
       ifft
       fft2
       ifft2
       fftn
       ifftn


etc...

Numpy is intended to be a reasonable replacement of all the
*functionality* of Numeric.  I think Travis tried very hard to make
sure that if an old code could run only on Numeric (without SciPy),
then it could also run on numpy (possibly after some manual work).

As much as I pushed for breaking all backwards compatibility for the
sake of *clean APIs* in numpy, I think it's a good decision to keep
b.compat. in terms of functionality.  Otherwise a number of projects
could be stuck on Numeric if for whatever reason they consider SciPy
too large/complex a dependency.

Cheers,


f


From robert.kern at gmail.com  Wed Mar  7 02:53:48 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 07 Mar 2007 01:53:48 -0600
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <db6b5ecc0703062346r6eba1927o250a08278c5441d0@mail.gmail.com>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>	<20070305205644.GA18701@virginia.edu>	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>	<45ED948A.2040708@ee.byu.edu>	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>	<45ED9A2F.3050803@ee.byu.edu>	<20070306180252.GA25007@swri16wm.electro.swri.edu>	<B323ED65-1903-415D-B155-D00AAD800E11@nyu.edu>	<45EE6616.4080306@ee.byu.edu>
	<db6b5ecc0703062346r6eba1927o250a08278c5441d0@mail.gmail.com>
Message-ID: <45EE6F8C.8070408@gmail.com>

Fernando Perez wrote:
> On 3/7/07, Travis Oliphant <oliphant at ee.byu.edu> wrote:
>>> -100
>>>
>>> I am strongly opposed to mixing computations into NumPY.  Why is 1-d
>>> interpolation good to have inside of NumPy? why not nd-interpolation?
>>> why not fft? or SVD? or some linear algebra thing?
>>>
>> I agree with the idea of the statement.  However, practicality may beat
>> purity here because of the backward-compatibility issue.  The only
>> proposal on the table is pulling in more functions from arrayfns.   I'm
>> fine if they live in the oldnumeric name-space (but it will actually be
>> easier to put them in the lib namespace because the arrayfnsmodule.c
>> file became the _compiled_base.c file in numpy.lib.
>>
>> Perhaps they can still live in _compiled_base.c  but not be pulled in to
>> numpy.lib (only a  oldnumeric.arrayfns module).
> 
> -1.  If they are going to go in, make them first-class citizens.

Except for interp, those functions really don't deserve to be first class. I
can't even understand what most of them do on a first (and second and third,
frankly) pass through their comments.

http://fresh.t-systems-sfr.com/unix/src/privat2/gmath-0.3.tar.gz:a/GMatH/lib/NumPySession/doc/NumPy/module-arrayfns.html

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From fperez.net at gmail.com  Wed Mar  7 03:12:37 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Wed, 7 Mar 2007 01:12:37 -0700
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <45EE6F8C.8070408@gmail.com>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
	<45ED948A.2040708@ee.byu.edu>
	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
	<45ED9A2F.3050803@ee.byu.edu>
	<20070306180252.GA25007@swri16wm.electro.swri.edu>
	<B323ED65-1903-415D-B155-D00AAD800E11@nyu.edu>
	<45EE6616.4080306@ee.byu.edu>
	<db6b5ecc0703062346r6eba1927o250a08278c5441d0@mail.gmail.com>
	<45EE6F8C.8070408@gmail.com>
Message-ID: <db6b5ecc0703070012r4da62b46m4927def69b194073@mail.gmail.com>

On 3/7/07, Robert Kern <robert.kern at gmail.com> wrote:

> > -1.  If they are going to go in, make them first-class citizens.
>
> Except for interp, those functions really don't deserve to be first class. I
> can't even understand what most of them do on a first (and second and third,
> frankly) pass through their comments.

Well, then maybe they just don't belong in the code at all if they are so bad?

My take on it is just a simple one: if it's functionally unique, it
shouldn't be in oldnumeric because then people will pull it out of
there forever, because no amount of 'numpyfying' their code will give
them that particular feature, since it's not an API issue.

Hopefully oldnumeric is a module whose usage should decrease with
time, and I'm trying to avoid something that would prevent that from
being possible.

But that's why I qualified my statement with '*if* they are going in'.
 Perhaps the right solution is for them NOT to go in, if they are
really bad enough.

I'll shut up now, ultimately I won't be using this anyway :)

Cheers,

f


From h.schulz at fzd.de  Wed Mar  7 04:26:58 2007
From: h.schulz at fzd.de (Schulz, Henrik)
Date: Wed, 07 Mar 2007 10:26:58 +0100
Subject: [SciPy-user] Problem with scipy.integrate
Message-ID: <03199d44edb6d343943cd00587129f77@fz-rossendorf.de>

Hi all,

I have a problem with the installation of Scipy, especially using the
fblas-libraries:

After I fixed the problem described here:
http://cens.ioc.ee/~pearu/scipy/INSTALL.html#blas-sources-shipped-with-l
apack-are-incomplete

by downloading blas and compiling with:
ifc -shared -O3 -unroll -w90 -w95 -cm -fPIC -FI *.f -o fblas.so

I get the following messages when trying to import scipy.integrate:

python
Python 2.3.4 (#1, Oct  9 2006, 18:28:26)
[GCC 3.4.4 20050721 (Red Hat 3.4.4-2)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import scipy.integrate
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "/usr/lib64/python2.3/site-packages/scipy/integrate/__init__.py",
line 9, in ?
    from quadrature import *
  File
"/usr/lib64/python2.3/site-packages/scipy/integrate/quadrature.py", line
8, in ?
    from scipy.special.orthogonal import p_roots
  File "/usr/lib64/python2.3/site-packages/scipy/special/__init__.py",
line 10, in ?
    import orthogonal
  File "/usr/lib64/python2.3/site-packages/scipy/special/orthogonal.py",
line 65, in ?
    from scipy.linalg import eig
  File "/usr/lib64/python2.3/site-packages/scipy/linalg/__init__.py",
line 8, in ?
    from basic import *
  File "/usr/lib64/python2.3/site-packages/scipy/linalg/basic.py", line
227, in ?
    import decomp
  File "/usr/lib64/python2.3/site-packages/scipy/linalg/decomp.py", line
21, in ?
    from blas import get_blas_funcs
  File "/usr/lib64/python2.3/site-packages/scipy/linalg/blas.py", line
14, in ?
    from scipy.linalg import fblas
ImportError: dynamic module does not define init function (initfblas)


What can I do to solve this problem?

Thanks!
Henrik






From cimrman3 at ntc.zcu.cz  Wed Mar  7 04:45:01 2007
From: cimrman3 at ntc.zcu.cz (Robert Cimrman)
Date: Wed, 07 Mar 2007 10:45:01 +0100
Subject: [SciPy-user] Using umfpack to calculate a incomplete
 LU	factorisation (ILU)
In-Reply-To: <pan.2007.03.06.19.34.13.563384@sun.ac.za>
References: <pan.2007.03.06.19.34.13.563384@sun.ac.za>
Message-ID: <45EE899D.9040304@ntc.zcu.cz>

Neilen Marais wrote:
> I'm trying to get an incomplete LU factorisation of a sparse matrix using
> scipy.linsolve.splu. It's using umfpack as far as I can tell. I'm telling
> it to drop elements using the drop_tol keyword argument, though it seems
> to be having no effect:

Just one correction: linsolve.splu uses SuperLU function *gstrf, so if
the threshold does not work, the problem is there...

Otherwise you can get LU factors with UmfpackContext.lu, but those are
always complete.

cheers,
r.


From davbrow at gmail.com  Wed Mar  7 07:00:27 2007
From: davbrow at gmail.com (Dave)
Date: Wed, 7 Mar 2007 04:00:27 -0800
Subject: [SciPy-user] Arrayfns in numpy?
Message-ID: <588f52020703070400x316a2677x1c07a6ce441b1dcf@mail.gmail.com>

On Tue, 06 Mar 2007 Travis Oliphant wrote:
>That is generally what we believe as well.  The problem is that Numeric
>already included several additional features.   Trying to maintain some
>semblance of backward compatibility is why NumPy has not shrunk even more.

>Because the interp function was already in Numeric and is not that big,
>perhaps it should be added to NumPy.

I just realized I have been struggling to install scipy, as it turns
out, just to use an interpolation function.  I usually try to minimize
dependencies for anything I plan to distribute to others.  My current
experience seems to support arguments for both sides on whether to add
interpolation to numpy.

As I see it there are three useful levels of array functionality.
First, python itself should ideally allow for handling array objects
with ease and efficiency, and promote "...one way to do it..." as per
python dogma.  Hopefully ndarray added to the standard library will
make this more of a reality some day.

Second there are common math functions and manipulations over arrays
that are potentially useful for pedestrian as well as esoteric
applications.   Sin, cos, basic integration and differentiation,
matrix and vector math, simple statistics, etc., are all reasonably
within this category.  Hobbyist game programmers as well as scientists
can use them.  Basic 1d and perhaps 2d interpolation are also within
this level of functionality in my opinion.  The capabilities and the
API need to be stable and extremely reliable.  Numpy today provides a
relatively rich and efficient package with many of these features and
properties.

There is no hard cutoff but obscure (to me) probability distributions,
simulation frameworks, image processing functions named after a living
person, field-specific packages, etc, are all wonderful to have when
you need them but are best suited for the third level of
functionality, scipy.  Sophisticated packages will have dependencies
and might be updated often or have API changes as capabilities evolve.
 It makes sense to isolate this from the more generic and stable
features in numpy.

Included in numpy for the important practical goal of backward
compatibility are objects like, I would guess, numpy.kaiser and others
that don't seem to fit well the stated divisions.  If interpolation is
in Numeric then numpy is at least deserving of a compatibility
function.  But as a user without legacy requirements I would prefer
one designed for numpy and for future applications.

The other aspect of the argument concerns the dependency issue.  I
have not been able to use the scipy.interpolate module because of what
appears to be some dependency that I can't resolve.  I certainly don't
want the problem moved to numpy and I sympathize very much with "less
is more" arguments in this respect.  So I would like to have available
basic interpolation features in numpy if the functions can be added
without adding new dependencies.  In general I don't see a problem
with adding select numerical features if it is done carefully and with
costs/benefits of the whole package in mind.

In summary:
+1 interpolation for numpy,
-1 new dependencies for numpy,
+1 balanced practical approach to adding numpy features


-- David


From cimrman3 at ntc.zcu.cz  Wed Mar  7 07:09:19 2007
From: cimrman3 at ntc.zcu.cz (Robert Cimrman)
Date: Wed, 07 Mar 2007 13:09:19 +0100
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <588f52020703070400x316a2677x1c07a6ce441b1dcf@mail.gmail.com>
References: <588f52020703070400x316a2677x1c07a6ce441b1dcf@mail.gmail.com>
Message-ID: <45EEAB6F.5070501@ntc.zcu.cz>

Dave wrote:
> On Tue, 06 Mar 2007 Travis Oliphant wrote:
>> That is generally what we believe as well.  The problem is that Numeric
>> already included several additional features.   Trying to maintain some
>> semblance of backward compatibility is why NumPy has not shrunk even more.
> 
>> Because the interp function was already in Numeric and is not that big,
>> perhaps it should be added to NumPy.
> 
> I just realized I have been struggling to install scipy, as it turns
> out, just to use an interpolation function.  I usually try to minimize
> dependencies for anything I plan to distribute to others.  My current
> experience seems to support arguments for both sides on whether to add
> interpolation to numpy.
> 
> As I see it there are three useful levels of array functionality.
> First, python itself should ideally allow for handling array objects
> with ease and efficiency, and promote "...one way to do it..." as per
> python dogma.  Hopefully ndarray added to the standard library will
> make this more of a reality some day.
> 
> Second there are common math functions and manipulations over arrays
> that are potentially useful for pedestrian as well as esoteric
> applications.   Sin, cos, basic integration and differentiation,
> matrix and vector math, simple statistics, etc., are all reasonably
> within this category.  Hobbyist game programmers as well as scientists
> can use them.  Basic 1d and perhaps 2d interpolation are also within
> this level of functionality in my opinion.  The capabilities and the
> API need to be stable and extremely reliable.  Numpy today provides a
> relatively rich and efficient package with many of these features and
> properties.
> 
> There is no hard cutoff but obscure (to me) probability distributions,
> simulation frameworks, image processing functions named after a living
> person, field-specific packages, etc, are all wonderful to have when
> you need them but are best suited for the third level of
> functionality, scipy.  Sophisticated packages will have dependencies
> and might be updated often or have API changes as capabilities evolve.
>  It makes sense to isolate this from the more generic and stable
> features in numpy.
> 
> Included in numpy for the important practical goal of backward
> compatibility are objects like, I would guess, numpy.kaiser and others
> that don't seem to fit well the stated divisions.  If interpolation is
> in Numeric then numpy is at least deserving of a compatibility
> function.  But as a user without legacy requirements I would prefer
> one designed for numpy and for future applications.
> 
> The other aspect of the argument concerns the dependency issue.  I
> have not been able to use the scipy.interpolate module because of what
> appears to be some dependency that I can't resolve.  I certainly don't
> want the problem moved to numpy and I sympathize very much with "less
> is more" arguments in this respect.  So I would like to have available
> basic interpolation features in numpy if the functions can be added
> without adding new dependencies.  In general I don't see a problem
> with adding select numerical features if it is done carefully and with
> costs/benefits of the whole package in mind.
> 
> In summary:
> +1 interpolation for numpy,
> -1 new dependencies for numpy,
> +1 balanced practical approach to adding numpy features

As Dave correctly mentioned, there are three levels of functionality:
1. core of numpy (ndarray) - basic array handling
2. 'general purpose' functions (like interp, linalg.*, ...) with no
external dependencies, no fortran, easy to install
3. full scipy

Currently, 1. and 2. are in one package (numpy) - why not make two?

r.



From souheil.inati at nyu.edu  Wed Mar  7 08:58:16 2007
From: souheil.inati at nyu.edu (Souheil Inati)
Date: Wed, 7 Mar 2007 08:58:16 -0500
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <45EEAB6F.5070501@ntc.zcu.cz>
References: <588f52020703070400x316a2677x1c07a6ce441b1dcf@mail.gmail.com>
	<45EEAB6F.5070501@ntc.zcu.cz>
Message-ID: <798643F3-E8AA-42EB-99BB-00116F221BC2@nyu.edu>

I agree with Dave and Robert's view of the three layers of  
functionality.  I am agnostic as to whether layers 1 and 2 should be  
one package or two as long as the API for both is extremely stable.

In this scheme, the old numeric compatibility layer is part of level 2.

-Souheil

On Mar 7, 2007, at 7:09 AM, Robert Cimrman wrote:

> Dave wrote:
>> On Tue, 06 Mar 2007 Travis Oliphant wrote:
>>> That is generally what we believe as well.  The problem is that  
>>> Numeric
>>> already included several additional features.   Trying to  
>>> maintain some
>>> semblance of backward compatibility is why NumPy has not shrunk  
>>> even more.
>>
>>> Because the interp function was already in Numeric and is not  
>>> that big,
>>> perhaps it should be added to NumPy.
>>
>> >snip
>>
>> In summary:
>> +1 interpolation for numpy,
>> -1 new dependencies for numpy,
>> +1 balanced practical approach to adding numpy features
>
> As Dave correctly mentioned, there are three levels of functionality:
> 1. core of numpy (ndarray) - basic array handling
> 2. 'general purpose' functions (like interp, linalg.*, ...) with no
> external dependencies, no fortran, easy to install
> 3. full scipy
>
> Currently, 1. and 2. are in one package (numpy) - why not make two?
>
> r.



From matthieu.brucher at gmail.com  Wed Mar  7 09:17:09 2007
From: matthieu.brucher at gmail.com (Matthieu Brucher)
Date: Wed, 7 Mar 2007 15:17:09 +0100
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <45ED9A2F.3050803@ee.byu.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
	<45ED948A.2040708@ee.byu.edu>
	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
	<45ED9A2F.3050803@ee.byu.edu>
Message-ID: <e76aa17f0703070617j6aeeccf7g19f81896eba2e9d9@mail.gmail.com>

2007/3/6, Travis Oliphant <oliphant at ee.byu.edu>:
>
> Perhaps we could put interp in there as well (it doesn't look too big).
>   A simple 1-d interpolation would probably be a useful thing to have in
> NumPy.   What do others think?
>
> -Travis


I'm opposed to this as well, interpolation is not linear algebra, it is
signal processing, and as Souheil said, why not nd-interpolation with
B-Splines then ?

Matthieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070307/7fc5ae84/attachment.html>

From aisaac at american.edu  Wed Mar  7 10:14:39 2007
From: aisaac at american.edu (Alan Isaac)
Date: Wed, 07 Mar 2007 10:14:39 -0500
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <e76aa17f0703070617j6aeeccf7g19f81896eba2e9d9@mail.gmail.com>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>	<20070305205644.GA18701@virginia.edu>	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>	<45ED948A.2040708@ee.byu.edu>	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>	<45ED9A2F.3050803@ee.byu.edu>
	<e76aa17f0703070617j6aeeccf7g19f81896eba2e9d9@mail.gmail.com>
Message-ID: <45EED6DF.9020502@american.edu>

Matthieu Brucher wrote:
> I'm opposed to this as well, interpolation is not linear algebra, it is
> signal processing, and as Souheil said, why not nd-interpolation with
> B-Splines then ?

I am not taking sides on this issue, but I do
take issue with trying to settle it with this
kind of "argument".  A rhetorical question does
not set out the issues.  For example, treating
a rhetorical question as a real question for a
moment, one might respond that a simple 1-d
interpolation would
- find wide use among people who need nothing more
- enhance backward compatibility
- impose minimum maintenance requirements

Again, I am not taking sides on the issue, so I
am in no way suggesting that such points are
decisive.  However opposition that fails to
address such points is hardly decisive either.

My point of reference would be: if it is likely
to drain even modest developer effort away from
core NumPy issues, then it is harmful. Otherwise
it is harmless as long as it introduces no new
dependencies.

Alan Isaac


From matthieu.brucher at gmail.com  Wed Mar  7 11:03:51 2007
From: matthieu.brucher at gmail.com (Matthieu Brucher)
Date: Wed, 7 Mar 2007 17:03:51 +0100
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <45EED6DF.9020502@american.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
	<45ED948A.2040708@ee.byu.edu>
	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
	<45ED9A2F.3050803@ee.byu.edu>
	<e76aa17f0703070617j6aeeccf7g19f81896eba2e9d9@mail.gmail.com>
	<45EED6DF.9020502@american.edu>
Message-ID: <e76aa17f0703070803l2b68b846h8023876ac8273bbb@mail.gmail.com>

>
> Again, I am not taking sides on the issue, so I
> am in no way suggesting that such points are
> decisive.  However opposition that fails to
> address such points is hardly decisive either.



I read the end of the discussion and I see your point ;)


My point of reference would be: if it is likely
> to drain even modest developer effort away from
> core NumPy issues, then it is harmful. Otherwise
> it is harmless as long as it introduces no new
> dependencies.


I think that separating interpolation in two parts can be harmful, but as it
is needed for a backward compatiblity reason, it can be understood.
The problem could be that people will want to have more complex
interpolation in NumPy as the simple one exist, and then this question will
be raised again :(
Not quite sure there is an answer that will satisfy everybody - in fact I'm
sure of the opposite :) -

Matthieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070307/0f354b1e/attachment.html>

From oliphant at ee.byu.edu  Wed Mar  7 12:25:57 2007
From: oliphant at ee.byu.edu (Travis Oliphant)
Date: Wed, 07 Mar 2007 10:25:57 -0700
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <45EED6DF.9020502@american.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>	<20070305205644.GA18701@virginia.edu>	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>	<45ED948A.2040708@ee.byu.edu>	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>	<45ED9A2F.3050803@ee.byu.edu>	<e76aa17f0703070617j6aeeccf7g19f81896eba2e9d9@mail.gmail.com>
	<45EED6DF.9020502@american.edu>
Message-ID: <45EEF5A5.6080202@ee.byu.edu>

Alan Isaac wrote:

>Matthieu Brucher wrote:
>  
>
>>I'm opposed to this as well, interpolation is not linear algebra, it is
>>signal processing, and as Souheil said, why not nd-interpolation with
>>B-Splines then ?
>>    
>>
>
>I am not taking sides on this issue, but I do
>take issue with trying to settle it with this
>kind of "argument".  A rhetorical question does
>not set out the issues.  For example, treating
>a rhetorical question as a real question for a
>moment, one might respond that a simple 1-d
>interpolation would
>- find wide use among people who need nothing more
>- enhance backward compatibility
>- impose minimum maintenance requirements
>
>Again, I am not taking sides on the issue, so I
>am in no way suggesting that such points are
>decisive.  However opposition that fails to
>address such points is hardly decisive either.
>
>My point of reference would be: if it is likely
>to drain even modest developer effort away from
>core NumPy issues, then it is harmful. Otherwise
>it is harmless as long as it introduces no new
>dependencies.
>  
>
This is an impoirtant point for me.  I'm only considering adding interp 
because it's already done and just a matter of cut-and-paste.   

My take on this, is that I would rather use developer time to make scipy 
more modular and easier to install (this usually translates to binary 
builds for multiple platforms).   That way you could grab just the 
modules of interest.  Ideally, grabbing these modules is a matter of 
using setuptools (or some other package management concept) to do it 
automatically as needed.

-Travis




From prabhu at aero.iitb.ac.in  Wed Mar  7 23:34:31 2007
From: prabhu at aero.iitb.ac.in (Prabhu Ramachandran)
Date: Thu, 8 Mar 2007 10:04:31 +0530
Subject: [SciPy-user] tvtk GL2PS
In-Reply-To: <001501c75fe9$942463e0$c000a8c0@JELLE>
References: <001501c75fe9$942463e0$c000a8c0@JELLE>
Message-ID: <17903.37463.446065.89579@gargle.gargle.HOWL>

>>>>> "Jelle" == Jelle Feringa / EZCT Architecture & Design Research <jelle.feringa at ezct.net> writes:

    Jelle> Hi, Currently I'm using the enthon (win32, 2.4) 1.0.0
    Jelle> release, which is terrific!

Thanks to Bryce and the enthought build team (or is that a one man
army?)  for all their work!

    Jelle> Though there is a minor, but to me important glitch in
    Jelle> (t)vtk; 
[...]
    Jelle> Saving as a vector PS/EPS/PDF/TeX file using GL2PS is
    Jelle> either not supported by your version of VTK or you have not
    Jelle> configured VTK to work with GL2PS -- read the documentation
    Jelle> for the vtkGL2PSExporter class.
 
    Jelle> Which is really too bad.

This would require Bryce to build VTK with the VTK_USE_GL2PS flag set
to on for the next release.  There are no other dependencies that this
brings up.
 
cheers,
prabhu


From nwagner at iam.uni-stuttgart.de  Thu Mar  8 08:19:54 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Thu, 08 Mar 2007 14:19:54 +0100
Subject: [SciPy-user] Using umfpack to calculate a incomplete
 LU	factorisation (ILU)
In-Reply-To: <45EE899D.9040304@ntc.zcu.cz>
References: <pan.2007.03.06.19.34.13.563384@sun.ac.za>
	<45EE899D.9040304@ntc.zcu.cz>
Message-ID: <45F00D7A.4060603@iam.uni-stuttgart.de>

Robert Cimrman wrote:
> Neilen Marais wrote:
>   
>> I'm trying to get an incomplete LU factorisation of a sparse matrix using
>> scipy.linsolve.splu. It's using umfpack as far as I can tell. I'm telling
>> it to drop elements using the drop_tol keyword argument, though it seems
>> to be having no effect:
>>     
>
> Just one correction: linsolve.splu uses SuperLU function *gstrf, so if
> the threshold does not work, the problem is there...
>
> Otherwise you can get LU factors with UmfpackContext.lu, but those are
> always complete.
>
> cheers,
> r.
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   
Just out of interest

Is there something comparable with

sigma_solve = linsolve.splu(A - sigma*B).solve

available in UMFPACK ?

Nils
 


From jelle.feringa at ezct.net  Thu Mar  8 10:41:27 2007
From: jelle.feringa at ezct.net (jelle)
Date: Thu, 8 Mar 2007 15:41:27 +0000 (UTC)
Subject: [SciPy-user] tvtk GL2PS
References: <001501c75fe9$942463e0$c000a8c0@JELLE>
	<17903.37463.446065.89579@gargle.gargle.HOWL>
Message-ID: <loom.20070308T164014-938@post.gmane.org>


> This would require Bryce to build VTK with the VTK_USE_GL2PS flag set
> to on for the next release.  There are no other dependencies that this
> brings up.
> 
> cheers,
> prabhu


Thanks Prabu,

That really would be wonderful!
Being able to export to a vector format really is a huge feauture!

Cheers,

-jelle




From nwagner at iam.uni-stuttgart.de  Thu Mar  8 11:06:38 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Thu, 08 Mar 2007 17:06:38 +0100
Subject: [SciPy-user] Usage of fmin_ncg
Message-ID: <45F0348E.6070306@iam.uni-stuttgart.de>

Hi all,

I have a question wrt optimize.fmin_ncg(F,v_0,Fp, fhess_p = Fhess_p).
especially the last argument is of interest. I am not sure if I applied
it correctly.

The help function yields

      fhess_p -- a function to compute the Hessian of f times an
                 arbitrary vector: fhess_p (x, p, *args)

Is p the arbitrary vector ?
Any hint will be appreciated.


Nils

P.S.: The Hessian is not directly available.


from scipy import *
from pylab import plot, show, semilogy, xlabel, ylabel, legend, savefig,
title, figure, ylim
#
#
def R(v):
    """ Rayleigh quotient """
    return dot(v.T,A*v)/dot(v.T,B*v)

def F(v):
    """ Objective function """
    Bv = bsolve(v)
    res = linalg.norm(A*Bv-R(Bv)*B*Bv)/linalg.norm(B*Bv)
    h =    0.25*dot(bsolve(v).T,v)**2+0.5*dot(sigma_solve(v).T,v)
    data.append(res)
    return h

def Fp(v):
    """ Gradient of the objective function """
    return dot(bsolve(v).T,v)*bsolve(v)+sigma_solve(v)

def Fhess_p(v,p):
    """ Hessian applied to an arbitrary vector p """
    return
dot(bsolve(v).T,v)*bsolve(p)+sigma_solve(p)+2*dot(outer(bsolve(v),bsolve(v)),p)

#
# Test matrices from MatrixMarket
#
A = io.mmread('bcsstk02.mtx.gz')
B = io.mmread('bcsstm02.mtx.gz')
n = A.shape[0]

sigma = 47.5                                       # Preassigned shift
bsolve = linsolve.splu(B).solve                    # inv(B)*( )
sigma_solve = linsolve.splu(A - sigma*B).solve     # inv(A-sigma*B)*( )

random.seed(10)
v_0 = random.rand(n)                               # Initial vector

data=[]
v = optimize.fmin_ncg(F,v_0,Fp, fhess_p = Fhess_p)
semilogy(arange(0,len(data)),data)
print  'Newton CG',-1./(2*sqrt(-F(v)))+sigma
print R(bsolve(v))
print



From peridot.faceted at gmail.com  Thu Mar  8 11:16:48 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Thu, 8 Mar 2007 11:16:48 -0500
Subject: [SciPy-user] Random number generation and scipy.stats
Message-ID: <ce557a360703080816s34ea97b3k7716ef2ade879b62@mail.gmail.com>

Hi,

scipy.stats defines a number of distribution objects (scipy.stats.chi2
represents the chi-squared distribution, for example). How useful
would it be to add a .generate(size=(1,)) method which generated
random numbers according to the distribution? There's practically no
extra effort involved for those functions which have a .ppf() method,
since that allows one to use the transformation method:

In [1]: import numpy, scipy.stats

In [2]: scipy.stats.chi2.ppf(numpy.random.rand(5),2)
Out[2]: array([ 3.19017391,  3.27772839,  3.44953492,  8.09066854,  0.29628122])

Maybe the simplicity makes it not worth implementing, but on the one
hand there may be distributions that allow much more efficient
generation, and on the other this may not occur to users...

Anne M. Archibald


From cimrman3 at ntc.zcu.cz  Thu Mar  8 11:23:32 2007
From: cimrman3 at ntc.zcu.cz (Robert Cimrman)
Date: Thu, 08 Mar 2007 17:23:32 +0100
Subject: [SciPy-user] Using umfpack to calculate a incomplete
 LU	factorisation (ILU)
In-Reply-To: <45F00D7A.4060603@iam.uni-stuttgart.de>
References: <pan.2007.03.06.19.34.13.563384@sun.ac.za>	<45EE899D.9040304@ntc.zcu.cz>
	<45F00D7A.4060603@iam.uni-stuttgart.de>
Message-ID: <45F03884.3000101@ntc.zcu.cz>

Nils Wagner wrote:
> Just out of interest
> 
> Is there something comparable with
> 
> sigma_solve = linsolve.splu(A - sigma*B).solve
> 
> available in UMFPACK ?

Sure, although it is not a oneliner. This is the relevant part of the
docstring:

---
Sparse matrix in CSR or CSC format: mtx
Righthand-side: rhs
Solution: sol

import scipy.linsolve.umfpack as um
# Contruct the solver.
umfpack = um.UmfpackContext()
# Make LU decomposition.
umfpack.numeric( mtx )
...
# Use already LU-decomposed matrix.
sol1 = umfpack( um.UMFPACK_A, mtx, rhs1, autoTranspose = True )
sol2 = umfpack( um.UMFPACK_A, mtx, rhs2, autoTranspose = True )
# same as:
sol1 = umfpack.solve( um.UMFPACK_A, mtx, rhs1, autoTranspose = True )
sol2 = umfpack.solve( um.UMFPACK_A, mtx, rhs2, autoTranspose = True )
---

So basically, after calling umfpack.numeric( mtx ), all subsequent calls
to umfpack or umfpack.solve() will reuse the LU decomposition (if mtx
stays the same...)

Would you be interested in exposing this in linsolve, as, say,

def factorized( A ):
    if isUmfpack and useUmfpack:
	# This must be written...
    else:
        return splu( A ).solve


r.


From jdh2358 at gmail.com  Thu Mar  8 11:43:25 2007
From: jdh2358 at gmail.com (John Hunter)
Date: Thu, 8 Mar 2007 10:43:25 -0600
Subject: [SciPy-user] Random number generation and scipy.stats
In-Reply-To: <ce557a360703080816s34ea97b3k7716ef2ade879b62@mail.gmail.com>
References: <ce557a360703080816s34ea97b3k7716ef2ade879b62@mail.gmail.com>
Message-ID: <88e473830703080843o73f8649ei5a91840a46091816@mail.gmail.com>

On 3/8/07, Anne Archibald <peridot.faceted at gmail.com> wrote:
> Hi,
>
> scipy.stats defines a number of distribution objects (scipy.stats.chi2
> represents the chi-squared distribution, for example). How useful
> would it be to add a .generate(size=(1,)) method which generated
> random numbers according to the distribution? There's practically no
> extra effort involved for those functions which have a .ppf() method,
> since that allows one to use the transformation method:

Isn't this what rvs already does, or am I misunderstanding you


In [184]: import scipy.stats as stats

In [185]: stats.expon(0,5).rvs(10)
Out[185]:
array([  6.49235468,   3.66423025,  11.22019501,   4.50247735,
         0.95703071,   4.67328849,   1.52512936,  22.77158456,
         7.32379865,   4.45632927])


JDH


From cimrman3 at ntc.zcu.cz  Thu Mar  8 12:10:30 2007
From: cimrman3 at ntc.zcu.cz (Robert Cimrman)
Date: Thu, 08 Mar 2007 18:10:30 +0100
Subject: [SciPy-user] linsolve.factorized was: Re: Using umfpack to
 calculate a incomplete LU factorisation (ILU)
In-Reply-To: <45F03884.3000101@ntc.zcu.cz>
References: <pan.2007.03.06.19.34.13.563384@sun.ac.za>	<45EE899D.9040304@ntc.zcu.cz>	<45F00D7A.4060603@iam.uni-stuttgart.de>
	<45F03884.3000101@ntc.zcu.cz>
Message-ID: <45F04386.6010105@ntc.zcu.cz>

Robert Cimrman wrote:
> Nils Wagner wrote:
>> Just out of interest
>>
>> Is there something comparable with
>>
>> sigma_solve = linsolve.splu(A - sigma*B).solve
>>
>> available in UMFPACK ?
> 
> Sure, although it is not a oneliner. This is the relevant part of the
> docstring:

> Would you be interested in exposing this in linsolve, as, say,
> 
> def factorized( A ):
>     if isUmfpack and useUmfpack:
> 	# This must be written...
>     else:
>         return splu( A ).solve


Well, I did it since I am going to need this, too :-)

In [3]:scipy.linsolve.factorized?
...
Definition:     scipy.linsolve.factorized(A)
Docstring:
    Return a fuction for solving a linear system, with A pre-factorized.

    Example:
      solve = factorized( A ) # Makes LU decomposition.
      x1 = solve( rhs1 ) # Uses the LU factors.
      x2 = solve( rhs2 ) # Uses again the LU factors.

This uses UMFPACK if available.

cheers,
r.


From peridot.faceted at gmail.com  Thu Mar  8 12:32:37 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Thu, 8 Mar 2007 12:32:37 -0500
Subject: [SciPy-user] Random number generation and scipy.stats
In-Reply-To: <88e473830703080843o73f8649ei5a91840a46091816@mail.gmail.com>
References: <ce557a360703080816s34ea97b3k7716ef2ade879b62@mail.gmail.com>
	<88e473830703080843o73f8649ei5a91840a46091816@mail.gmail.com>
Message-ID: <ce557a360703080932i5986991v669bebf5438ca69f@mail.gmail.com>

On 08/03/07, John Hunter <jdh2358 at gmail.com> wrote:

> Isn't this what rvs already does, or am I misunderstanding you

Indeed. I feel silly now.

Sorry for the noise!

Anne


From jpeacock at mesoscopic.mines.edu  Thu Mar  8 12:34:36 2007
From: jpeacock at mesoscopic.mines.edu (Jared Peacock)
Date: Thu, 8 Mar 2007 10:34:36 -0700
Subject: [SciPy-user] Phase Unwrapping Algorithm
Message-ID: <17904.18732.283675.88678@hipparchus.mines.edu>


Does anybody know of a good phase unwrapping algorithm in Python?

J. Peacock


From jdh2358 at gmail.com  Thu Mar  8 12:55:26 2007
From: jdh2358 at gmail.com (John Hunter)
Date: Thu, 8 Mar 2007 11:55:26 -0600
Subject: [SciPy-user] Phase Unwrapping Algorithm
In-Reply-To: <17904.18732.283675.88678@hipparchus.mines.edu>
References: <17904.18732.283675.88678@hipparchus.mines.edu>
Message-ID: <88e473830703080955u7d677dd2he524aee7380d704a@mail.gmail.com>

On 3/8/07, Jared Peacock <jpeacock at mesoscopic.mines.edu> wrote:
>
> Does anybody know of a good phase unwrapping algorithm in Python?

Does scipy.signal.hilbert do what you want?

"""
Analytic signal s(t) = a(t) * sin( phi(t) )
"""
from pylab import figure, nx, show
from scipy.signal import hilbert

t = nx.arange(0.0, 0.5, 0.001)
s = nx.sin(2*nx.pi*10*t) + nx.mlab.randn(len(t))
h = hilbert(s)

a = nx.absolute(h)
phase = nx.angle(h)

fig = figure()
ax1 = fig.add_subplot(211)
ax1.plot(t, s, 'g-', t, a*nx.cos(phase), 'b-')

ax2 = fig.add_subplot(212)
ax2.plot(t, s-a*nx.cos(phase))

show()


From lorenzo.isella at gmail.com  Thu Mar  8 13:11:59 2007
From: lorenzo.isella at gmail.com (Lorenzo Isella)
Date: Thu, 8 Mar 2007 19:11:59 +0100
Subject: [SciPy-user] Reading Binary Files
Message-ID: <a2b3004b0703081011y457d5c81m754ae03dee8e65aa@mail.gmail.com>

Dear All,
I hope this is not too off-topic. I have been given an old but
reliable fortran code for fluid dynamic simulations. It saves a a lot
of data using 3D arrays (q1,q2,q3,pr) as a binary file.
I cut and paste the part of the fortran code saving the data into a binary file:
(in the following iav=1 and iwrq2=1).


if(iav.eq.1) then
      namfil='field'//ipfi//'.dat'
      pnamfil='field'//ipfi//'.dat'
      open(13,file=namfil,form='unformatted')
                   else

      pnamfil=filcnw
      print*, "filcnw is", filcnw
      open(13,file=filcnw,form='unformatted')
                   endif
      print*, "iav is", iav

      write(6,*) pnamfil,'written at t=',
     1          time, ' prma mi=',prma,prmi
      nfil=13
      rewind(nfil)
      write(nfil) n1,n2,n3
      write(nfil) ros,alx3d,ren,time
      if(iwrq2.eq.1) then
c
c    large memory occupancy
c
       print*,"I am saving the extended results"
      write(nfil) (((q1(i,j,k),i=1,n1),j=1,n2),k=1,n3),
     1            (((q2(i,j,k),i=1,n1),j=1,n2),k=1,n3),
     1            (((q3(i,j,k),i=1,n1),j=1,n2),k=1,n3),
     1            (((pr(i,j,k),i=1,n1),j=1,n2),k=1,n3)
                     else
c
c    reduced memory occupancy
c    the pressure is not necessary for restarting files
c    even for post processing can be saved but then
c    the advancement of a time step should be performed
c
      write(nfil) (((q1(i,j,k),i=1,n1),j=1,n2),k=1,n3),
     1            (((q3(i,j,k),i=1,n1),j=1,n2),k=1,n3),
     1            (((pr(i,j,k),i=1,n1),j=1,n2),k=1,n3)
                     endif
      write(nfil) ntime,ntt,nav
      close(nfil)



The results is for instance file field0010.dat, which I try reading in
Python by using pylab and the statement:

s = file( './field0010.dat','rb' ).read( )

newarr = fromstring(s ,Float)

but the content of newarr seems absolutely wrong (number of the order
of 1e+309 which are not produced or saved in the simulations...).
Am I doing something wrong?

Kind Regards

Lorenzo


From nwagner at iam.uni-stuttgart.de  Thu Mar  8 13:17:32 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Thu, 08 Mar 2007 19:17:32 +0100
Subject: [SciPy-user] linsolve.factorized was: Re: Using umfpack to
 calculate a incomplete LU factorisation (ILU)
In-Reply-To: <45F04386.6010105@ntc.zcu.cz>
References: <pan.2007.03.06.19.34.13.563384@sun.ac.za>
	<45EE899D.9040304@ntc.zcu.cz>
	<45F00D7A.4060603@iam.uni-stuttgart.de>
	<45F03884.3000101@ntc.zcu.cz> <45F04386.6010105@ntc.zcu.cz>
Message-ID: <web-80147878@uni-stuttgart.de>

On Thu, 08 Mar 2007 18:10:30 +0100
  Robert Cimrman <cimrman3 at ntc.zcu.cz> wrote:
> Robert Cimrman wrote:
>> Nils Wagner wrote:
>>> Just out of interest
>>>
>>> Is there something comparable with
>>>
>>> sigma_solve = linsolve.splu(A - sigma*B).solve
>>>
>>> available in UMFPACK ?
>> 
>> Sure, although it is not a oneliner. This is the 
>>relevant part of the
>> docstring:
> 
>> Would you be interested in exposing this in linsolve, 
>>as, say,
>> 
>> def factorized( A ):
>>     if isUmfpack and useUmfpack:
>> 	# This must be written...
>>     else:
>>         return splu( A ).solve
> 
> 
> Well, I did it since I am going to need this, too :-)
> 
> In [3]:scipy.linsolve.factorized?
> ...
> Definition:     scipy.linsolve.factorized(A)
> Docstring:
>    Return a fuction for solving a linear system, with A 
>pre-factorized.
> 
>    Example:
>      solve = factorized( A ) # Makes LU decomposition.
>      x1 = solve( rhs1 ) # Uses the LU factors.
>      x2 = solve( rhs2 ) # Uses again the LU factors.
> 
> This uses UMFPACK if available.
> 
> cheers,
> r.
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
  
Great ! Thank you very much !

Nils


From wollez at gmx.net  Thu Mar  8 13:32:29 2007
From: wollez at gmx.net (WolfgangZ)
Date: Thu, 08 Mar 2007 19:32:29 +0100
Subject: [SciPy-user] Reading Binary Files
In-Reply-To: <a2b3004b0703081011y457d5c81m754ae03dee8e65aa@mail.gmail.com>
References: <a2b3004b0703081011y457d5c81m754ae03dee8e65aa@mail.gmail.com>
Message-ID: <espkrt$q1d$1@sea.gmane.org>

Lorenzo Isella schrieb:
> Dear All,
> I hope this is not too off-topic. I have been given an old but
> reliable fortran code for fluid dynamic simulations. It saves a a lot
> of data using 3D arrays (q1,q2,q3,pr) as a binary file.
> I cut and paste the part of the fortran code saving the data into a binary file:
> (in the following iav=1 and iwrq2=1).
> 
> 
> if(iav.eq.1) then
>       namfil='field'//ipfi//'.dat'
>       pnamfil='field'//ipfi//'.dat'
>       open(13,file=namfil,form='unformatted')
>                    else
> 
>       pnamfil=filcnw
>       print*, "filcnw is", filcnw
>       open(13,file=filcnw,form='unformatted')
>                    endif
>       print*, "iav is", iav
> 
>       write(6,*) pnamfil,'written at t=',
>      1          time, ' prma mi=',prma,prmi
>       nfil=13
>       rewind(nfil)
>       write(nfil) n1,n2,n3
>       write(nfil) ros,alx3d,ren,time
>       if(iwrq2.eq.1) then
> c
> c    large memory occupancy
> c
>        print*,"I am saving the extended results"
>       write(nfil) (((q1(i,j,k),i=1,n1),j=1,n2),k=1,n3),
>      1            (((q2(i,j,k),i=1,n1),j=1,n2),k=1,n3),
>      1            (((q3(i,j,k),i=1,n1),j=1,n2),k=1,n3),
>      1            (((pr(i,j,k),i=1,n1),j=1,n2),k=1,n3)
>                      else
> c
> c    reduced memory occupancy
> c    the pressure is not necessary for restarting files
> c    even for post processing can be saved but then
> c    the advancement of a time step should be performed
> c
>       write(nfil) (((q1(i,j,k),i=1,n1),j=1,n2),k=1,n3),
>      1            (((q3(i,j,k),i=1,n1),j=1,n2),k=1,n3),
>      1            (((pr(i,j,k),i=1,n1),j=1,n2),k=1,n3)
>                      endif
>       write(nfil) ntime,ntt,nav
>       close(nfil)
> 
> 
> 
> The results is for instance file field0010.dat, which I try reading in
> Python by using pylab and the statement:
> 
> s = file( './field0010.dat','rb' ).read( )
> 
> newarr = fromstring(s ,Float)
> 
> but the content of newarr seems absolutely wrong (number of the order
> of 1e+309 which are not produced or saved in the simulations...).
> Am I doing something wrong?
> 
> Kind Regards
> 
> Lorenzo


are you sure that fortran saves binary files? Have you  checked the 
files with a text editor?



From hasslerjc at comcast.net  Thu Mar  8 14:41:35 2007
From: hasslerjc at comcast.net (John Hassler)
Date: Thu, 08 Mar 2007 14:41:35 -0500
Subject: [SciPy-user] Reading Binary Files
In-Reply-To: <espkrt$q1d$1@sea.gmane.org>
References: <a2b3004b0703081011y457d5c81m754ae03dee8e65aa@mail.gmail.com>
	<espkrt$q1d$1@sea.gmane.org>
Message-ID: <45F066EF.5040709@comcast.net>

An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070308/d985ee26/attachment.html>

From gnurser at googlemail.com  Thu Mar  8 14:58:25 2007
From: gnurser at googlemail.com (George Nurser)
Date: Thu, 8 Mar 2007 19:58:25 +0000
Subject: [SciPy-user] Reading Binary Files
In-Reply-To: <45F066EF.5040709@comcast.net>
References: <a2b3004b0703081011y457d5c81m754ae03dee8e65aa@mail.gmail.com>
	<espkrt$q1d$1@sea.gmane.org> <45F066EF.5040709@comcast.net>
Message-ID: <1d1e6ea70703081158o5dc1d3c8x775406815f9a7939@mail.gmail.com>

This short routine might be useful it gets rid of the extra stuff at
the beginning and end that fortran adds. But I can't guarantee it will
work for you.

use swap=True if you are reading little-endian on big
endian/big-endian on little endian.
f8=false if you are using real*4 instead of real*8 ( double precision)

You also need to put somewhere:
from numpy import *

def readbin(file_in,swap=False,f8=True):
    if f8:
        htot =fromfile(file = file_in ,dtype=Float)
        c = htot.view(single)
        hc = c[1:-1].view(double)
    else:
        htot =fromfile(file = file_in ,dtype=Float32)
        hc = c[1:-1]

    if swap:
        hc = hc.byteswap()
    return hc


From hasslerjc at comcast.net  Thu Mar  8 15:28:11 2007
From: hasslerjc at comcast.net (John Hassler)
Date: Thu, 08 Mar 2007 15:28:11 -0500
Subject: [SciPy-user] Reading Binary Files
In-Reply-To: <a2b3004b0703081011y457d5c81m754ae03dee8e65aa@mail.gmail.com>
References: <a2b3004b0703081011y457d5c81m754ae03dee8e65aa@mail.gmail.com>
Message-ID: <45F071DB.9050209@comcast.net>

I realize that I was so eager to answer your question that I neglected 
to try to solve your problem.  If you have the source (and if this isn't 
legacy data files), just recompile with more usable I/O formats.  We 
used to use unformatted I/O back in the day when a BIG computer might 
have several tens of K (as in K) of memory, and the high speed I/O was 
mag. tape.  We'd dump to tape to get some free memory, backspace, and 
read it back in when we needed it.  "Unformatted" I/O was intended to be 
used by programs made by the same compiler on the same computer, and not 
for any external use.
john

Lorenzo Isella wrote:
> Dear All,
> I hope this is not too off-topic. I have been given an old but
> reliable fortran code for fluid dynamic simulations. It saves a a lot
> of data using 3D arrays (q1,q2,q3,pr) as a binary file.
> I cut and paste the part of the fortran code saving the data into a binary file:
> (in the following iav=1 and iwrq2=1).
>
>
> if(iav.eq.1) then
>       namfil='field'//ipfi//'.dat'
>       pnamfil='field'//ipfi//'.dat'
>       open(13,file=namfil,form='unformatted')
>                    else
>
>       pnamfil=filcnw
>       print*, "filcnw is", filcnw
>       open(13,file=filcnw,form='unformatted')
>                    endif
>       print*, "iav is", iav
>
>       write(6,*) pnamfil,'written at t=',
>      1          time, ' prma mi=',prma,prmi
>       nfil=13
>       rewind(nfil)
>       write(nfil) n1,n2,n3
>       write(nfil) ros,alx3d,ren,time
>       if(iwrq2.eq.1) then
> c
> c    large memory occupancy
> c
>        print*,"I am saving the extended results"
>       write(nfil) (((q1(i,j,k),i=1,n1),j=1,n2),k=1,n3),
>      1            (((q2(i,j,k),i=1,n1),j=1,n2),k=1,n3),
>      1            (((q3(i,j,k),i=1,n1),j=1,n2),k=1,n3),
>      1            (((pr(i,j,k),i=1,n1),j=1,n2),k=1,n3)
>                      else
> c
> c    reduced memory occupancy
> c    the pressure is not necessary for restarting files
> c    even for post processing can be saved but then
> c    the advancement of a time step should be performed
> c
>       write(nfil) (((q1(i,j,k),i=1,n1),j=1,n2),k=1,n3),
>      1            (((q3(i,j,k),i=1,n1),j=1,n2),k=1,n3),
>      1            (((pr(i,j,k),i=1,n1),j=1,n2),k=1,n3)
>                      endif
>       write(nfil) ntime,ntt,nav
>       close(nfil)
>
>
>
> The results is for instance file field0010.dat, which I try reading in
> Python by using pylab and the statement:
>
> s = file( './field0010.dat','rb' ).read( )
>
> newarr = fromstring(s ,Float)
>
> but the content of newarr seems absolutely wrong (number of the order
> of 1e+309 which are not produced or saved in the simulations...).
> Am I doing something wrong?
>
> Kind Regards
>
> Lorenzo
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>   


From bnuttall at uky.edu  Thu Mar  8 15:32:03 2007
From: bnuttall at uky.edu (Brandon Nuttall)
Date: Thu, 08 Mar 2007 15:32:03 -0500
Subject: [SciPy-user] Reading Binary Files
In-Reply-To: <45F066EF.5040709@comcast.net>
References: <a2b3004b0703081011y457d5c81m754ae03dee8e65aa@mail.gmail.com>
	<espkrt$q1d$1@sea.gmane.org> <45F066EF.5040709@comcast.net>
Message-ID: <6.0.1.1.2.20070308152737.0280ed60@pop.uky.edu>

An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070308/34f98cb8/attachment.html>

From peridot.faceted at gmail.com  Thu Mar  8 17:28:57 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Thu, 8 Mar 2007 17:28:57 -0500
Subject: [SciPy-user] Phase Unwrapping Algorithm
In-Reply-To: <17904.18732.283675.88678@hipparchus.mines.edu>
References: <17904.18732.283675.88678@hipparchus.mines.edu>
Message-ID: <ce557a360703081428u657f89dbk1e215de7faa0c24@mail.gmail.com>

On 08/03/07, Jared Peacock <jpeacock at mesoscopic.mines.edu> wrote:
>
> Does anybody know of a good phase unwrapping algorithm in Python?

If by phase unwrapping you mean taking values that have been reduced
modulo 1 (say) and trying to reconstitute the originals, the problem
is in principle underdetermined - we really can't distinguish big
jumps in the data from the results of phase wrapping. Still, for
reasonable data it's often enough to consider every jump bigger than
0.5 a phase wrap:

def unwrap(a):
     return a-numpy.concatenate(([0],numpy.cumsum(numpy.round(a[1:]-a[:-1]))))

That is, a[1:]-a[:-1] extracts the first difference, round converts
anything less than -0.5 to -1 and anything more than 0,5 to 1, we
concatenate a zero on the front, and then we subtract (because a jump
down of 0.6 is rounded to -1 and we want it to be a jump up) the
result from our original data.

I suppose this could be made more efficient by rewriting more of the
operations to be in-place.

Anne


From drfredkin at ucsd.edu  Thu Mar  8 19:12:17 2007
From: drfredkin at ucsd.edu (Donald Fredkin)
Date: Fri, 9 Mar 2007 00:12:17 +0000 (UTC)
Subject: [SciPy-user] Phase Unwrapping Algorithm
References: <17904.18732.283675.88678@hipparchus.mines.edu>
Message-ID: <esq8p0$so$1@sea.gmane.org>

Jared Peacock wrote:

> 
> Does anybody know of a good phase unwrapping algorithm in Python?
> 

Whatis wrong with scipy.unwrap, which is actually
numpy.lib.function_base.unwrap?

-- 




From maschke at molgen.mpg.de  Fri Mar  9 09:54:02 2007
From: maschke at molgen.mpg.de (Elisabeth Maschke-Dutz)
Date: Fri, 09 Mar 2007 15:54:02 +0100
Subject: [SciPy-user] Python embedded in C : import scipy fails
Message-ID: <45F1750A.6070002@molgen.mpg.de>

Hallo,

I want to use some programs that uses scipy
in an in C embedded python program.
import sys
import os
from string import join
..........
all those comands work fine.
The sys.path is found, it looks as if the whole environment
is the same as it is when I start python in a shell.
The directory .../python2.4/site-packages/scipy is also set
in sys.path.

But it is impossible to import scipy,
the program terminates with a segmentation fault or the function is not 
found.
The same problem occurs if I try to import numpy.
It seems that packages, which do not belong to
python can not be importet. Single programs anywhere
do not make any problems if the path is set in sys.path.


The import of scipy in all other python programs do
not make any problems:

 >>> import scipy
 >>>


the testmodel.c code:


#include <Python.h>



int main(int argc, char **argv)
{
  char *modelp = "modelp";
  char *testmodel = "testmodel";
  PyObject *pName, *pModule, *pFunc;
  PyObject *pArgs1, *pArgs2, *pValue, *hallop;
  int i;
  double l;
  l=100.0;
  Py_SetProgramName(argv[0]);
  Py_Initialize();
  PySys_SetArgv(argc, argv);
  pName = PyString_FromString(modelp);
  if (pName != NULL)
    {
      printf("pName != NULL");
    }
  else
    {
      printf("pName == NULL");
    }
  pModule = PyImport_Import(pName);
  Py_DECREF(pName);
  if (Py_IsInitialized())
    printf("Initialized!!!");
  if (pModule != NULL)
    {
     pFunc = PyObject_GetAttrString(pModule,testmodel);
     printf("Function found!!!!!!!!!");
    }
  else
    {
      printf("pModule == NULL!!!!!!!!!");
    }
  if (PyCallable_Check(pFunc))
    printf("function callable");
  pArgs1 = PyFloat_FromDouble(l);
  pArgs2 = PyTuple_New(1);
  i=0;
  PyTuple_SetItem(pArgs2, i,pArgs1);
  if (pArgs2 != NULL)
    printf("pArgs2!=Null");
  pValue = PyObject_CallObject(pFunc, pArgs2);
  Py_DECREF(pArgs1);
  Py_DECREF(pArgs2);
  Py_DECREF(pFunc);
  Py_DECREF(pModule);
  printf("3Result of call...............");
  printf("Result of call: %ld\n", PyInt_AsLong(pValue));
  return 1;
}

the python code:

import sys
import os
from string import join
#from math import *
from string import *
sys.path.append(".................../dir2")
sys.path.append(".......................python2.4/site-packages/scipy")

#import scipy

import testimport  # a single file (program)

def testmodel(a):
    print 'in testmodel................ a= ',a
    print sys.path
    print os.environ
    import distutils.sysconfig
    t1=distutils.sysconfig.get_config_var('LINKFORSHARED')
    print t1
    t2=distutils.sysconfig.get_config_vars()
    print t2
    t3=distutils.sysconfig.get_python_lib()
    print t3
    print 'nach import'
    #import numpy
    return 17


I use gcc with  CFLAGS  = -I/............../linux/include/python2.4 -c

and link  with:


 -pthread -lm -ldl -lutil   
................lib/python2.4/config/libpython2.4.a

.../getpath.o  ..../config.o

Has somebody an idee what I can do?

                  Elisabeth M.



From lorenzo.isella at gmail.com  Fri Mar  9 10:57:36 2007
From: lorenzo.isella at gmail.com (Lorenzo Isella)
Date: Fri, 9 Mar 2007 16:57:36 +0100
Subject: [SciPy-user] Array Bounds
Message-ID: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>

Dear All,
Again a newbie question: I noticed that Python (or at least
Numpy/Scipy) follows the C convention of labeling as zero the first
element in an array.
I can live with this, but I do not find very intuitive the loop behavior:

for i in range (1,5 )

means that i actually takes values 1,2,3,4 and not 5!
Why this choice? Is it a consequence of the array labeling? In case
one should not like it, can it be changed by the user?
Kind Regards

Lorenzo


From perry at stsci.edu  Fri Mar  9 12:12:09 2007
From: perry at stsci.edu (Perry Greenfield)
Date: Fri, 9 Mar 2007 12:12:09 -0500
Subject: [SciPy-user] job opportunity at Space Telescope Science Institute
Message-ID: <C3DCC608-5E9A-4EAF-A9A2-3341A0B9430C@stsci.edu>

We are looking for someone to fill a position at the Space Telescope  
Science Institute (in Baltimore, MD) to work on Python tools for  
astronomical data processing and analysis. Details can be found at:

http://www.stsci.edu/institute/brc/hr/co/external/Req559.html


From lbolla at gmail.com  Fri Mar  9 12:19:49 2007
From: lbolla at gmail.com (lorenzo bolla)
Date: Fri, 9 Mar 2007 18:19:49 +0100
Subject: [SciPy-user] job opportunity at Space Telescope Science
	Institute
In-Reply-To: <C3DCC608-5E9A-4EAF-A9A2-3341A0B9430C@stsci.edu>
References: <C3DCC608-5E9A-4EAF-A9A2-3341A0B9430C@stsci.edu>
Message-ID: <80c99e790703090919l5cb4dd0mb06216c8eaccdb77@mail.gmail.com>

I cannot access the link...
-lorenzo

On 3/9/07, Perry Greenfield <perry at stsci.edu> wrote:
>
> We are looking for someone to fill a position at the Space Telescope
> Science Institute (in Baltimore, MD) to work on Python tools for
> astronomical data processing and analysis. Details can be found at:
>
> http://www.stsci.edu/institute/brc/hr/co/external/Req559.html
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070309/b75a717e/attachment.html>

From robert.kern at gmail.com  Fri Mar  9 13:07:27 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 09 Mar 2007 12:07:27 -0600
Subject: [SciPy-user] Array Bounds
In-Reply-To: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>
References: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>
Message-ID: <45F1A25F.7010302@gmail.com>

Lorenzo Isella wrote:
> Dear All,
> Again a newbie question: I noticed that Python (or at least
> Numpy/Scipy) follows the C convention of labeling as zero the first
> element in an array.
> I can live with this, but I do not find very intuitive the loop behavior:
> 
> for i in range (1,5 )
> 
> means that i actually takes values 1,2,3,4 and not 5!
> Why this choice? Is it a consequence of the array labeling?

It is a consequence of *Python* using zero-indexing everywhere. numpy simply
follows suit. It means that range(len(sequence)) gives you all of the indices
into sequence.

Just give it some time. You will grow used to it.

> In case
> one should not like it, can it be changed by the user?

No.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From wbaxter at gmail.com  Fri Mar  9 13:20:01 2007
From: wbaxter at gmail.com (Bill Baxter)
Date: Sat, 10 Mar 2007 03:20:01 +0900
Subject: [SciPy-user] Python embedded in C : import scipy fails
In-Reply-To: <45F1750A.6070002@molgen.mpg.de>
References: <45F1750A.6070002@molgen.mpg.de>
Message-ID: <e86a5fd00703091020w532ab84cxf6149c758d4410ed@mail.gmail.com>

Don't know if this is the problem, but in my app, I call
"import_array1(false);" right after calling Py_Initialize().

--bb

On 3/9/07, Elisabeth Maschke-Dutz <maschke at molgen.mpg.de> wrote:
>
> Hallo,
>
> I want to use some programs that uses scipy
> in an in C embedded python program.
> import sys
> import os
> from string import join
> ..........
> all those comands work fine.
> The sys.path is found, it looks as if the whole environment
> is the same as it is when I start python in a shell.
> The directory .../python2.4/site-packages/scipy is also set
> in sys.path.
>
> But it is impossible to import scipy,
> the program terminates with a segmentation fault or the function is not
> found.
> The same problem occurs if I try to import numpy.
> ...
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070310/2b7883e0/attachment.html>

From perry at stsci.edu  Fri Mar  9 14:09:58 2007
From: perry at stsci.edu (Perry Greenfield)
Date: Fri, 9 Mar 2007 14:09:58 -0500
Subject: [SciPy-user] job opportunity at Space Telescope Science
	Institute
In-Reply-To: <80c99e790703090919l5cb4dd0mb06216c8eaccdb77@mail.gmail.com>
References: <C3DCC608-5E9A-4EAF-A9A2-3341A0B9430C@stsci.edu>
	<80c99e790703090919l5cb4dd0mb06216c8eaccdb77@mail.gmail.com>
Message-ID: <D77040EE-D9D8-4B57-8598-A2FA1334A622@stsci.edu>

I'm told this problem has been corrected (but not being on an  
external network I can't conveniently check). Please try again.

Thanks, Perry

On Mar 9, 2007, at 12:19 PM, lorenzo bolla wrote:

> I cannot access the link...
> -lorenzo
>
> On 3/9/07, Perry Greenfield <perry at stsci.edu> wrote: We are looking  
> for someone to fill a position at the Space Telescope
> Science Institute (in Baltimore, MD) to work on Python tools for
> astronomical data processing and analysis. Details can be found at:
>
> http://www.stsci.edu/institute/brc/hr/co/external/Req559.html
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user



From pwang at enthought.com  Fri Mar  9 14:18:09 2007
From: pwang at enthought.com (Peter Wang)
Date: Fri, 9 Mar 2007 13:18:09 -0600
Subject: [SciPy-user] Array Bounds
In-Reply-To: <45F1A25F.7010302@gmail.com>
References: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>
	<45F1A25F.7010302@gmail.com>
Message-ID: <6D8DF23D-66D0-41E2-9523-6103E528B2AE@enthought.com>

On Mar 9, 2007, at 12:07 PM, Robert Kern wrote:

>> In case one should not like it, can it be changed by the user?
>
> No.

Well, this is not strictly true.  The range() function is a built-in,  
but you can override it by defining your own that includes the final  
index:

def range(a,b):
     return __builtins__.range(a,b+1)

 >>> range(1,5)
[1, 2, 3, 4, 5]


Of course, you should NEVER do this.  Other people that catch you  
redefining built-in functions will probably call you all sorts of bad  
names.


-Peter


From robert.kern at gmail.com  Fri Mar  9 14:22:18 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 09 Mar 2007 13:22:18 -0600
Subject: [SciPy-user] job opportunity at Space Telescope
	Science	Institute
In-Reply-To: <D77040EE-D9D8-4B57-8598-A2FA1334A622@stsci.edu>
References: <C3DCC608-5E9A-4EAF-A9A2-3341A0B9430C@stsci.edu>	<80c99e790703090919l5cb4dd0mb06216c8eaccdb77@mail.gmail.com>
	<D77040EE-D9D8-4B57-8598-A2FA1334A622@stsci.edu>
Message-ID: <45F1B3EA.9010402@gmail.com>

Perry Greenfield wrote:
> I'm told this problem has been corrected (but not being on an  
> external network I can't conveniently check). Please try again.

Works for me.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From lanceboyle at qwest.net  Fri Mar  9 19:56:19 2007
From: lanceboyle at qwest.net (Jerry)
Date: Fri, 9 Mar 2007 17:56:19 -0700
Subject: [SciPy-user] Array Bounds
In-Reply-To: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>
References: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>
Message-ID: <19B1090D-8692-4895-AFDD-04FF662072CA@qwest.net>


On Mar 9, 2007, at 8:57 AM, Lorenzo Isella wrote:

> Dear All,
> Again a newbie question: I noticed that Python (or at least
> Numpy/Scipy) follows the C convention of labeling as zero the first
> element in an array.
> I can live with this, but I do not find very intuitive the loop  
> behavior:
>
> for i in range (1,5 )
>
> means that i actually takes values 1,2,3,4 and not 5!
> Why this choice? Is it a consequence of the array labeling? In case
> one should not like it, can it be changed by the user?
> Kind Regards
>
> Lorenzo

In my opinion, this is poor language design (along with Henry Ford- 
style array indexing--any array you want as long as it starts with  
zero) and an impediment to writing correct programs. Good design lets  
the programmer abstract his program to resemble the real-world  
problem that he is working on. Get used to it or move along--every  
time I post about this, that's the advice I get. (1) Python arrays  
will _never_ change; (2) it is blasphemous to even mention it; (3)  
someone will ask, why in the world would anyone want an array that  
doesn't start with zero; (4) someone will flame me.

Jerry



From aisaac at american.edu  Fri Mar  9 20:18:03 2007
From: aisaac at american.edu (Alan G Isaac)
Date: Fri, 9 Mar 2007 20:18:03 -0500
Subject: [SciPy-user] Array Bounds
In-Reply-To: <19B1090D-8692-4895-AFDD-04FF662072CA@qwest.net>
References: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com><19B1090D-8692-4895-AFDD-04FF662072CA@qwest.net>
Message-ID: <Mahogany-0.67.0-984-20070309-201803.00@american.edu>

On Fri, 9 Mar 2007, Jerry apparently wrote: 
> (2) it is blasphemous to even mention it; 

No, just pointless.
Surely you do not think you are the first?

Here's the opposite request for a unit-based language:
http://www-gatago.com/comp/soft-sys/matlab/30395164.html

Here is an oft-cited by not completely persuasive 
discussion:
http://www.cs.utexas.edu/users/EWD/transcriptions/EWD08xx/EWD831.html

Usually, what does it matter?
Just know what you are saying.

Cheers,
Alan Isaac




From lanceboyle at qwest.net  Fri Mar  9 21:20:41 2007
From: lanceboyle at qwest.net (Jerry)
Date: Fri, 9 Mar 2007 19:20:41 -0700
Subject: [SciPy-user] Array Bounds
In-Reply-To: <Mahogany-0.67.0-984-20070309-201803.00@american.edu>
References: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com><19B1090D-8692-4895-AFDD-04FF662072CA@qwest.net>
	<Mahogany-0.67.0-984-20070309-201803.00@american.edu>
Message-ID: <FC6010F3-EE62-43DC-AABD-FE0BD2EB7AD2@qwest.net>

The opposite request is not for unit-based arrays but for arrays of  
user-specified lower bounds.

Jerry


On Mar 9, 2007, at 6:18 PM, Alan G Isaac wrote:

>
> Here's the opposite request for a unit-based language:
> http://www-gatago.com/comp/soft-sys/matlab/30395164.html
>



From robert.kern at gmail.com  Fri Mar  9 22:20:16 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 09 Mar 2007 21:20:16 -0600
Subject: [SciPy-user] Array Bounds
In-Reply-To: <19B1090D-8692-4895-AFDD-04FF662072CA@qwest.net>
References: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>
	<19B1090D-8692-4895-AFDD-04FF662072CA@qwest.net>
Message-ID: <45F223F0.8030403@gmail.com>

Jerry wrote:

> In my opinion, this is poor language design (along with Henry Ford- 
> style array indexing--any array you want as long as it starts with  
> zero) and an impediment to writing correct programs. Good design lets  
> the programmer abstract his program to resemble the real-world  
> problem that he is working on. Get used to it or move along--every  
> time I post about this, that's the advice I get. (1) Python arrays  
> will _never_ change; (2) it is blasphemous to even mention it; (3)  
> someone will ask, why in the world would anyone want an array that  
> doesn't start with zero; (4) someone will flame me.

(0) Someone will tell you that you can define whatever classes you like with
whatever index semantics that you like. It's all very well to *say* that things
would be better if one could change the lower index bound at will, but you have
the ability to *show* us, so why don't you? Try it out! Work out the inevitable
kinks with writing generic code that is agnostic to the heterogeneous index
semantics. Show us the code!

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From perry at stsci.edu  Fri Mar  9 22:30:51 2007
From: perry at stsci.edu (Perry Greenfield)
Date: Fri, 9 Mar 2007 22:30:51 -0500
Subject: [SciPy-user] Array Bounds
In-Reply-To: <19B1090D-8692-4895-AFDD-04FF662072CA@qwest.net>
References: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>
	<19B1090D-8692-4895-AFDD-04FF662072CA@qwest.net>
Message-ID: <5BC5B737-7E08-451F-A0DA-D6E2F809F9E5@stsci.edu>


On Mar 9, 2007, at 7:56 PM, Jerry wrote:

>
> In my opinion, this is poor language design (along with Henry Ford-
> style array indexing--any array you want as long as it starts with
> zero) and an impediment to writing correct programs. Good design lets

Me wonders why you waste your time here. You should be using one of  
those good language designs instead of getting fools to fix their  
stupid language. It would be much more productive (and then show all  
what we should be using by showing us how much better your  
applications are).

> the programmer abstract his program to resemble the real-world
> problem that he is working on. Get used to it or move along--every
> time I post about this, that's the advice I get. (1) Python arrays
> will _never_ change; (2) it is blasphemous to even mention it; (3)
> someone will ask, why in the world would anyone want an array that
> doesn't start with zero; (4) someone will flame me.
>                          ^^^^^^^^^^^^^^^^^^^^^^^^^^
Er, that's what you were hoping for, wasn't it? ;-)

> Jerry
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user



From perry at stsci.edu  Fri Mar  9 22:59:55 2007
From: perry at stsci.edu (Perry Greenfield)
Date: Fri, 9 Mar 2007 22:59:55 -0500
Subject: [SciPy-user] Array Bounds
In-Reply-To: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>
References: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>
Message-ID: <20FCA004-7DB8-431C-B666-EE29DF9C70F4@stsci.edu>


On Mar 9, 2007, at 10:57 AM, Lorenzo Isella wrote:

> Dear All,
> Again a newbie question: I noticed that Python (or at least
> Numpy/Scipy) follows the C convention of labeling as zero the first
> element in an array.
> I can live with this, but I do not find very intuitive the loop  
> behavior:
>
> for i in range (1,5 )
>
> means that i actually takes values 1,2,3,4 and not 5!
> Why this choice? Is it a consequence of the array labeling? In case
> one should not like it, can it be changed by the user?
> Kind Regards
>
> Lorenzo

Your reaction is quite common for those coming from a language that  
uses 1-based indexing. But keep in mind there are also advantages for  
0-based indexing that you aren't aware of since you haven't used it.  
There are pros and cons to both (and yes, I've used both). Since  
Python uses 0-based, it generally isn't a good idea to adopt  
different conventions than the language uses for arrays. It may make  
your array usage more comfortable, but will lead to all sorts of  
problems when you use Python lists and tuples. (Note that you mention  
arrays, but your example is actually due to Python list behavior and  
the built-in range function, which has nothing to do with numpy; in  
principle array indexing behavior could be made different than Python  
list indexing behavior, but that would be a really bad thing to do).

Actually, your complaint is not directly due to that issue (though  
it's related). As I read it, it's because you are surprised that the  
end of the range does not include the number that's used for the end  
of the range. And that's a natural enough reaction. It's not  
intuitive to many people. But what's intuitive is not always the best  
choice. It works that way (I'm guessing here) because slices in  
Python work that way (e.g., mylist[1:5] doesn't include mylist[5]).  
Ok, why don't slices work that way? For various reasons, but one is  
that when you break a list into two pieces, e.g., mylist = mylist[:7]  
+ mylist[7:], you don't have to worry about dealing with those  
annoying +1 or -1 offsets to get things right. In this case, the  
number of items sliced is simply the difference of the two indices.  
It may seem like a small point, but a lot of errors arise from  
getting stuff like this wrong, especially the edge cases of when the  
slice point is at 0 or the end of an array. The rules Python has  
chosen for this stuff are consistent, and they do work pretty well.

So it isn't the most intuitive, especially to scientific or  
engineering types, but it's not that hard to get used to, and there  
are definite advantages to this approach.



From as8ca at virginia.edu  Sat Mar 10 01:31:07 2007
From: as8ca at virginia.edu (Alok Singhal)
Date: Sat, 10 Mar 2007 01:31:07 -0500
Subject: [SciPy-user] Array Bounds
In-Reply-To: <20FCA004-7DB8-431C-B666-EE29DF9C70F4@stsci.edu>
References: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>
	<20FCA004-7DB8-431C-B666-EE29DF9C70F4@stsci.edu>
Message-ID: <20070310063107.GB5256@virginia.edu>

On 09/03/07: 22:59, Perry Greenfield wrote:
> 
> Actually, your complaint is not directly due to that issue (though  
> it's related). As I read it, it's because you are surprised that the  
> end of the range does not include the number that's used for the end  
> of the range. And that's a natural enough reaction. It's not  
> intuitive to many people. But what's intuitive is not always the best  
> choice. It works that way (I'm guessing here) because slices in  
> Python work that way (e.g., mylist[1:5] doesn't include mylist[5]).  

It might also be (I am guessing too) because range(n) should,
intuitively, return n numbers.  Given 0-based indices, it makes (some)
sense to start the numbers at 0, and hence, stop at n-1.  So, then by
generalization, range(m,n) should generate integers m, m+1, .., n-1.
That is the least surprising result.

-Alok

-- 
Alok Singhal                               *   *          
Graduate Student, dept. of Astronomy   *           *     *
University of Virginia                                    
http://www.astro.virginia.edu/~as8ca/              *    * 


From hasslerjc at comcast.net  Sat Mar 10 10:23:05 2007
From: hasslerjc at comcast.net (John Hassler)
Date: Sat, 10 Mar 2007 10:23:05 -0500
Subject: [SciPy-user] Array Bounds
In-Reply-To: <20070310063107.GB5256@virginia.edu>
References: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>	<20FCA004-7DB8-431C-B666-EE29DF9C70F4@stsci.edu>
	<20070310063107.GB5256@virginia.edu>
Message-ID: <45F2CD59.3090704@comcast.net>

An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070310/b147f108/attachment.html>

From steve at shrogers.com  Sat Mar 10 13:36:21 2007
From: steve at shrogers.com (Steven H. Rogers)
Date: Sat, 10 Mar 2007 11:36:21 -0700
Subject: [SciPy-user] Array Bounds
In-Reply-To: <45F2CD59.3090704@comcast.net>
References: <a2b3004b0703090757w5f8d997dief70c9a9ac010792@mail.gmail.com>	<20FCA004-7DB8-431C-B666-EE29DF9C70F4@stsci.edu>	<20070310063107.GB5256@virginia.edu>
	<45F2CD59.3090704@comcast.net>
Message-ID: <45F2FAA5.2090905@shrogers.com>

John Hassler wrote:
> There is no way to make everybody happy, I'm afraid.  Early Fortran was 
> one-based.  A text I used to use (Carnahan, et al, "Applied Numerical 
> Methods," 1969) warns of the difficulty of translating mathematical sums 
> starting with zero to Fortran indices starting with 1.  On the other 
> hand, Press, et al., "Numerical Recipes in C," 1988, 1992, spend an 
> entire page justifying their practice of converting all C arrays to 
> one-based.  (This, by the way, has been roundly criticized by "real" C 
> programmers.)
> 
> Pascal, Fortran 77, and perhaps other 'modern' languages, allow 
> specifying both ends of the range, as in x[2:7] or something.  Some 
> people like this.  I find that it's just one more thing to keep track 
> of.  (By this I mean, if you see x[4] somewhere in a program, you have 
> to look back at the declaration to know what it means.)  I suppose that 
> something like this in Python would also require that one ALWAYS specify 
> both ends of range().
> 

For casual use, many would find it convenient to be able to change from 
0-based to 1-based indexing (APL allows this).  For larger programs 
touched by multiple programmers over time, it's a maintenance issue.

# Steve


From steve at shrogers.com  Sat Mar 10 15:30:02 2007
From: steve at shrogers.com (Steven H. Rogers)
Date: Sat, 10 Mar 2007 13:30:02 -0700
Subject: [SciPy-user] Teaching Array Languages
Message-ID: <45F3154A.5070808@shrogers.com>

Thanks to all who responded to my question about teaching array 
programming.  I've compiled a brief summary of the responses.

NumPy
=====
* Subject
   - Physics/Astronomy 3
   - Biotechnology 1
   - Engineering 2
   - Microeconomics 1
* Level
   - College/University 7

J
=
* Subject
   - Math 1
   - Engineering 1
   - Technical English
* Level
   - College/University 3
   -

Mathematica
===========
* Subject
   - Math 1
* Level
   - College/University 1

Matlab
======
* Subject
   - Engineering 1

IDL
===
* Subject
   - Physics/Astronomy 2
* Level
   - College/University 2

Discussion
==========
While this informal survey is neither precise nor comprehensive, I think 
it is interesting.  I queried the NumPy/SciPy and J mailing lists 
because those are the ones that I follow.  A more rigorous and broadly 
distributed survey may be in order.

Regards,
Steve



From nwagner at iam.uni-stuttgart.de  Mon Mar 12 08:30:33 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Mon, 12 Mar 2007 13:30:33 +0100
Subject: [SciPy-user] Structured matrices
Message-ID: <45F547E9.3050208@iam.uni-stuttgart.de>

Hi all,

I was wondering if the matrix family (see below) has a special name ?
And/or is there a way to construct this matrix via special matrices
(like Hankel, Toeplitz, etc.) ?


[[ 3.  1.]
 [ 1.  1.]]

[[ 5.  3.  1.]
 [ 3.  3.  1.]
 [ 1.  1.  1.]]

[[ 7.  5.  3.  1.]
 [ 5.  5.  3.  1.]
 [ 3.  3.  3.  1.]
 [ 1.  1.  1.  1.]]

[[ 9.  7.  5.  3.  1.]
 [ 7.  7.  5.  3.  1.]
 [ 5.  5.  5.  3.  1.]
 [ 3.  3.  3.  3.  1.]
 [ 1.  1.  1.  1.  1.]]

Nils

-------------- next part --------------
A non-text attachment was scrubbed...
Name: structure.py
Type: text/x-python
Size: 196 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070312/18705392/attachment.py>

From mwojc at p.lodz.pl  Mon Mar 12 10:09:25 2007
From: mwojc at p.lodz.pl (Marek Wojciechowski)
Date: Mon, 12 Mar 2007 15:09:25 +0100
Subject: [SciPy-user] Index of a value
In-Reply-To: <mailman.11.1173636017.31664.scipy-user@scipy.org>
References: <mailman.11.1173636017.31664.scipy-user@scipy.org>
Message-ID: <200703121509.26269.mwojc@p.lodz.pl>

Hallo!
Is there any shorthand for:

maxvalue = A.max()
maxindex = A.tolist().index(maxvalue)

where A is a 1-dimensional numpy array.

Thanks
Marek


From sgarcia at olfac.univ-lyon1.fr  Mon Mar 12 10:17:05 2007
From: sgarcia at olfac.univ-lyon1.fr (Samuel GARCIA)
Date: Mon, 12 Mar 2007 15:17:05 +0100
Subject: [SciPy-user] Index of a value
In-Reply-To: <200703121509.26269.mwojc@p.lodz.pl>
References: <mailman.11.1173636017.31664.scipy-user@scipy.org>
	<200703121509.26269.mwojc@p.lodz.pl>
Message-ID: <45F560E1.20308@olfac.univ-lyon1.fr>

Hi,
argmax    argmin

Sam

Marek Wojciechowski wrote:
> Hallo!
> Is there any shorthand for:
>
> maxvalue = A.max()
> maxindex = A.tolist().index(maxvalue)
>
> where A is a 1-dimensional numpy array.
>
> Thanks
> Marek
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Samuel Garcia
Laboratoire de Neurosciences Sensorielles, Comportement, Cognition.
CNRS - UMR5020 - Universite Claude Bernard LYON 1
Equipe logisique et technique
50, avenue Tony Garnier
69366 LYON Cedex 07
FRANCE
T?l : 04 37 28 74 64
Fax : 04 37 28 76 01
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~



From peridot.faceted at gmail.com  Mon Mar 12 13:52:09 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Mon, 12 Mar 2007 13:52:09 -0400
Subject: [SciPy-user] Structured matrices
In-Reply-To: <45F547E9.3050208@iam.uni-stuttgart.de>
References: <45F547E9.3050208@iam.uni-stuttgart.de>
Message-ID: <ce557a360703121052g59e7c2f0y364ea622d7b9fb55@mail.gmail.com>

On 12/03/07, Nils Wagner <nwagner at iam.uni-stuttgart.de> wrote:
> Hi all,
>
> I was wondering if the matrix family (see below) has a special name ?
> And/or is there a way to construct this matrix via special matrices
> (like Hankel, Toeplitz, etc.) ?

Well, I've never come across them before, but they're easy to make directly:

In [8]: minimum.outer(r_[7:-1:-2],r_[7:-1:-2])
Out[8]:
array([[7, 5, 3, 1],
       [5, 5, 3, 1],
       [3, 3, 3, 1],
       [1, 1, 1, 1]])

Anne


From bratona at yahoo.co.uk  Tue Mar 13 12:39:59 2007
From: bratona at yahoo.co.uk (Adam Malinowski)
Date: Tue, 13 Mar 2007 16:39:59 +0000 (UTC)
Subject: [SciPy-user] difficulties compiling the arpack module from the
	sandbox
Message-ID: <loom.20070313T171549-373@post.gmane.org>

Hi all,

I am having a lot of difficulty building the arpack module from the scipy 
sandbox, running windows 2000 and using mingw with cygwin. I have compiled 
numpy and scipy (latest svn) with no problems, and have sucessfully installed a 
number of scipy sandboxed modules by running the command: 'python setup.py 
build --compiler=mingw32 install', in the root of the module.

However, when I try this command for arpack, I get the following error:

---
No module named msvccompiler in numpy.distutils, trying from distutils..
  FOUND:
    libraries = ['lapack', 'lapack', 'f77blas', 'cblas', 'atlas']
    library_dirs = ['c:\\BlasLapackLibs']
    language = f77
    define_macros = [('ATLAS_INFO', '"\\"?.?.?\\""')]

running build
running config_fc
running build_src
building library "arpack" sources
building extension "arpack._arpack" sources
f2py options: []
  adding 'build\src.win32-2.5\fortranobject.c' to sources.
  adding 'build\src.win32-2.5' to include_dirs.
  adding 'build\src.win32-2.5\build\src.win32-2.5\_arpack-f2pywrappers.f' to sou
rces.
building data_files sources
running build_py
running build_clib
customize Mingw32CCompiler
customize Mingw32CCompiler using build_clib
0
Could not locate executable gfortran
Could not locate executable f95
customize GnuFCompiler
customize GnuFCompiler
customize GnuFCompiler using build_clib
building 'arpack' library
compiling Fortran sources
Fortran f77 compiler: C:\cygwin\bin\g77.exe -g -Wall -fno-second-underscore -mno
-cygwin -O3 -funroll-loops -march=pentium4 -mmmx -msse2 -msse -fomit-frame-point
er -malign-double
compile options: '-c'
g77.exe:f77: .\ARPACK\SRC\cgetv0.f
.\ARPACK\SRC\cgetv0.f: In subroutine `cgetv0':
.\ARPACK\SRC\cgetv0.f:123:
         include   'debug.h'
         ^
Unable to open INCLUDE file `debug.h' at (^)
.\ARPACK\SRC\cgetv0.f:124:
         include   'stat.h'
         ^
Unable to open INCLUDE file `stat.h' at (^)
error: Command "C:\cygwin\bin\g77.exe -g -Wall -fno-second-underscore -mno-cygwi
n -O3 -funroll-loops -march=pentium4 -mmmx -msse2 -msse -fomit-frame-pointer -ma
lign-double -c -c .\ARPACK\SRC\cgetv0.f -o build\temp.win32-2.5\ARPACK\SRC\cgetv
0.o" failed with exit status 1
---

I have also tried using the 'build_ext' command instead, which gives this error:

----
No module named msvccompiler in numpy.distutils, trying from distutils..
  FOUND:
    libraries = ['lapack', 'lapack', 'f77blas', 'cblas', 'atlas']
    library_dirs = ['c:\\BlasLapackLibs']
    language = f77
    define_macros = [('ATLAS_INFO', '"\\"?.?.?\\""')]

running build_ext
customize Mingw32CCompiler
customize Mingw32CCompiler using build_ext
0
Could not locate executable gfortran
Could not locate executable f95
customize GnuFCompiler
customize GnuFCompiler
customize GnuFCompiler using build_ext
building 'arpack._arpack' extension
compiling C sources
C compiler: gcc -mno-cygwin -O2 -Wall -Wstrict-prototypes

error: unknown file type '.src' (from '.\arpack.pyf.src')
---

Both of these errors have completely stumped me, as I'm quite new to fortran/
python compiling. Could it be a problem with the ATLAS or LAPACK setup? Or 
perhaps its a problem with numpy distutils? 

Any help would very much appreciated.

Thanks,
Adam



From ryanlists at gmail.com  Tue Mar 13 10:00:55 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Tue, 13 Mar 2007 09:00:55 -0500
Subject: [SciPy-user] dumb install question
Message-ID: <c5b438120703130700y5c804ea6yaf0ae7b92a42ada6@mail.gmail.com>

I have a dumb question.  I am having some troubles getting
scipy/numpy/matplotlib installed in Ubuntu Edgy (I have done it from
source in Breezy and used Andrew Straw's packages for Dapper), and I
read on http://scipy.org/Installing_SciPy/Linux?highlight=%28%28----%28-%2A%29%28%5Cr%29%3F%5Cn%29%28.%2A%29CategoryInstallation%5Cb%29
that I need to install either atlas-3dnow-dev or atlas-sse2-dev or
atlas-sse3-dev depending on my system.  If I have an Athlon X2
processor (running 32-bit Ubuntu), which do I need?  I think it is
sse2, but I am not completely sure.

Thanks,

Ryan


From ryanlists at gmail.com  Tue Mar 13 10:03:59 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Tue, 13 Mar 2007 09:03:59 -0500
Subject: [SciPy-user] problem with _num.seterr when importing scipy
In-Reply-To: <200702091531.11770.krlong@sandia.gov>
References: <200702091522.01125.krlong@sandia.gov> <45CCE67E.1090508@gmail.com>
	<200702091531.11770.krlong@sandia.gov>
Message-ID: <c5b438120703130703g79e10145l3e7fdbfb4bce8685@mail.gmail.com>

I was having the same problem.  Did this ever get solved?

Thanks,

Ryan

On 2/9/07, Kevin Long <krlong at sandia.gov> wrote:
>
> Hi Robert,
>
> Numpy 1.0b5 and scipy 0.5.2.
>
> - kevin
>
> On Friday 09 February 2007 15:24, Robert Kern wrote:
> > Kevin Long wrote:
> > > Hello,
> > >
> > > I'm getting an error message about "_num.seterr" when importing scipy.
> > > Output is below.
> > >
> > > <diogenes:/home/krlong> python
> > > Python 2.4.4 (#1, Feb  9 2007, 14:45:36)
> > > [GCC 3.4.6] on linux2
> > > Type "help", "copyright", "credits" or "license" for more information.
> > >
> > >>>> import scipy
> > >
> > > Traceback (most recent call last):
> > >   File "<stdin>", line 1, in ?
> > >   File "/usr/local/lib/python2.4/site-packages/scipy/__init__.py", line
> > > 37, in ?
> > >     _num.seterr(all='ignore')
> > > TypeError: seterr() got an unexpected keyword argument 'all'
> >
> > What versions of numpy and scipy do you have installed? E.g.:
> > >>> import numpy
> > >>> print numpy.__version__
> >
> > 1.0.2.dev3521
> >
> > >>> import scipy
> > >>> print scipy.__version__
> >
> > 0.5.3.dev2620
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From robert.kern at gmail.com  Wed Mar 14 00:32:58 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 13 Mar 2007 23:32:58 -0500
Subject: [SciPy-user] difficulties compiling the arpack module from the
 sandbox
In-Reply-To: <loom.20070313T171549-373@post.gmane.org>
References: <loom.20070313T171549-373@post.gmane.org>
Message-ID: <45F77AFA.5030506@gmail.com>

Adam Malinowski wrote:
> Hi all,
> 
> I am having a lot of difficulty building the arpack module from the scipy 
> sandbox, running windows 2000 and using mingw with cygwin. I have compiled 
> numpy and scipy (latest svn) with no problems, and have sucessfully installed a 
> number of scipy sandboxed modules by running the command: 'python setup.py 
> build --compiler=mingw32 install', in the root of the module.

That won't work. build doesn't take a --compiler option. Try these (pardon the
linebreaks):

  $ python setup.py build_src build_clib --compiler=mingw32 build_ext
--compiler=mingw32 build
  $ python setup.py install

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From robert.kern at gmail.com  Wed Mar 14 00:39:16 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 13 Mar 2007 23:39:16 -0500
Subject: [SciPy-user] problem with _num.seterr when importing scipy
In-Reply-To: <c5b438120703130703g79e10145l3e7fdbfb4bce8685@mail.gmail.com>
References: <200702091522.01125.krlong@sandia.gov>
	<45CCE67E.1090508@gmail.com>	<200702091531.11770.krlong@sandia.gov>
	<c5b438120703130703g79e10145l3e7fdbfb4bce8685@mail.gmail.com>
Message-ID: <45F77C74.1090901@gmail.com>

Ryan Krauss wrote:
> I was having the same problem.  Did this ever get solved?

The 'all' keyword was introduced after version 1.0b5, the version that Kevin
Long was using. That was the problem.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From bldrake at adaptcs.com  Tue Mar 13 13:44:00 2007
From: bldrake at adaptcs.com (Barry Drake)
Date: Tue, 13 Mar 2007 10:44:00 -0700 (PDT)
Subject: [SciPy-user] Structured matrices
In-Reply-To: <45F547E9.3050208@iam.uni-stuttgart.de>
Message-ID: <759494.19283.qm@web414.biz.mail.mud.yahoo.com>

Nils,
Does this matrix come from a particular application?  I'm working on algorithms for the non-negative matrix factorization (NMF).  With this matrix as input I'm getting some very strange results.  So I'm curious about potential applications.

Cheers!
Barry L. Drake
GA Tech

Nils Wagner wrote: Hi all,

I was wondering if the matrix family (see below) has a special name ?
And/or is there a way to construct this matrix via special matrices
(like Hankel, Toeplitz, etc.) ?


[[ 3.  1.]
 [ 1.  1.]]

[[ 5.  3.  1.]
 [ 3.  3.  1.]
 [ 1.  1.  1.]]

[[ 7.  5.  3.  1.]
 [ 5.  5.  3.  1.]
 [ 3.  3.  3.  1.]
 [ 1.  1.  1.  1.]]

[[ 9.  7.  5.  3.  1.]
 [ 7.  7.  5.  3.  1.]
 [ 5.  5.  5.  3.  1.]
 [ 3.  3.  3.  3.  1.]
 [ 1.  1.  1.  1.  1.]]

Nils

from scipy import *

def M(n):

    tmp = ones((n,n))
    for k in arange(n,0,-1):
        tmp[:k-1,:k-1] = tmp[:k-1,:k-1] + 2
    return tmp

for n in arange(2,8):
    print M(n)
    print
     
_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070313/522cc81a/attachment.html>

From fperez.net at gmail.com  Wed Mar 14 00:56:22 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Tue, 13 Mar 2007 22:56:22 -0600
Subject: [SciPy-user] dumb install question
In-Reply-To: <c5b438120703130700y5c804ea6yaf0ae7b92a42ada6@mail.gmail.com>
References: <c5b438120703130700y5c804ea6yaf0ae7b92a42ada6@mail.gmail.com>
Message-ID: <db6b5ecc0703132156r247c87e8o37bf6c2165f878d5@mail.gmail.com>

On 3/13/07, Ryan Krauss <ryanlists at gmail.com> wrote:
> I have a dumb question.  I am having some troubles getting
> scipy/numpy/matplotlib installed in Ubuntu Edgy (I have done it from
> source in Breezy and used Andrew Straw's packages for Dapper), and I
> read on http://scipy.org/Installing_SciPy/Linux?highlight=%28%28----%28-%2A%29%28%5Cr%29%3F%5Cn%29%28.%2A%29CategoryInstallation%5Cb%29
> that I need to install either atlas-3dnow-dev or atlas-sse2-dev or
> atlas-sse3-dev depending on my system.  If I have an Athlon X2
> processor (running 32-bit Ubuntu), which do I need?  I think it is
> sse2, but I am not completely sure.

Yes (sse2).  cat /proc/cpuinfo for the gory details.

cheers,

f


From philfarm at fastmail.fm  Mon Mar 12 23:57:52 2007
From: philfarm at fastmail.fm (philfarm at fastmail.fm)
Date: Mon, 12 Mar 2007 22:57:52 -0500
Subject: [SciPy-user] Best way to structure dynamic system-of-systems
 simulations?
In-Reply-To: <1172896076.28906.1177490111@webmail.messagingengine.com>
References: <1172896076.28906.1177490111@webmail.messagingengine.com>
Message-ID: <1173758272.25956.1179135235@webmail.messagingengine.com>

I'll try again: Can anyone offer any direction for me? I've looked a bit
at PyDSTool, but it strikes me (perhaps unjustifiably) as being rather
immature.

Conversely, if anyone could confirm that this is an area of need, then
I'll go off and work on this myself. If I manage to come up with
something workable, I'll offer it back to the community.

Thanks,
Phil

On Fri, 02 Mar 2007 22:27:56 -0600, philfarm at fastmail.fm said:
> I'm looking for any pointers on the best way to to structure the
> simulation of a system composed of subsystems, i.e., I want to build my
> system model as an assemblage of component models, where some components
> may use as U values the Y values from other components. Each of the
> components can be modeled in state-space form, i.e.:
> 
> X_dot = A(X,U,t)
>     Y = B(X,U,t)
> 
> where A and B are matrix functions (not necessarily linear).
> 
> Is anyone using any sort of framework to simplify this task? What's the
> best approach?
> 
> If it makes any difference, I've been using odeint.
> 
> Thanks,
> Phil
> 
> -- 
> http://www.fastmail.fm - And now for something completely different

> 
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user

-- 
http://www.fastmail.fm - One of many happy users:
  http://www.fastmail.fm/docs/quotes.html



From bratona at yahoo.co.uk  Wed Mar 14 01:33:37 2007
From: bratona at yahoo.co.uk (Adam Malinowski)
Date: Wed, 14 Mar 2007 05:33:37 +0000 (UTC)
Subject: [SciPy-user] difficulties compiling the arpack module from the
	sandbox
References: <loom.20070313T171549-373@post.gmane.org>
	<45F77AFA.5030506@gmail.com>
Message-ID: <loom.20070314T062301-354@post.gmane.org>

Robert Kern <robert.kern <at> gmail.com> writes:

> 
> Adam Malinowski wrote:
> > Hi all,
> > 
> > I am having a lot of difficulty building the arpack module from the scipy 
> > sandbox, running windows 2000 and using mingw with cygwin. I have compiled 
> > numpy and scipy (latest svn) with no problems, and have sucessfully 
installed a 
> > number of scipy sandboxed modules by running the command: 'python setup.py 
> > build --compiler=mingw32 install', in the root of the module.
> 
> That won't work. build doesn't take a --compiler option. Try these (pardon the
> linebreaks):
> 
>   $ python setup.py build_src build_clib --compiler=mingw32 build_ext
> --compiler=mingw32 build
>   $ python setup.py install
> 


Many thanks for your help Robert, though I appear to have solved the problem 
myself.

To my embarrassment, the problem seems to be simply that somewhere the path to 
the *.h files is incorrect (as indicated in the error message "Unable to open 
INCLUDE file"). To solve the problem I copied the files into the arpack root 
directory. Not ideal, since I don't understand what I should have done, but it 
worked!

Also, the build command did work with "--compiler=mingw32", as I used it 
successfully before seeing your reply, I guess it knows to pass it to 
build_clib.

Thanks again, 
Adam



From lechtlr at yahoo.com  Tue Mar 13 14:46:44 2007
From: lechtlr at yahoo.com (lechtlr)
Date: Tue, 13 Mar 2007 11:46:44 -0700 (PDT)
Subject: [SciPy-user] Using SciPy/NumPy optimization
Message-ID: <971382.30683.qm@web57911.mail.re3.yahoo.com>


Hi there:

I am looking for a non-linear, constrained optimization tool, and thought fmin_tnc would do the job that I wanted to do.
As a starter, I tried the attached script posted here recently. However, I get the following error when I run this script: 
"TypeError: only length-1 arrays can be converted to Python scalars" 
Can any one help me to figure out what I was doing wrong ?  

Thanks,
Lex

>>> 

Traceback (most recent call last):
 
 File "C:\Python24\lib\site-packages\scipy\optimize\tnc.py", line 200, in fmin_tnc
    fmin, ftol, rescale)
  File "C:\Python24\lib\site-packages\scipy\optimize\tnc.py", line 165, in func_and_grad
    g = approx_fprime(x, func, epsilon, *args)
  File "C:\Python24\lib\site-packages\scipy\optimize\optimize.py", line 555, in approx_fprime
    grad[k] = (apply(f,(xk+ei,)+args) - f0)/epsilon
TypeError: only
 length-1 arrays can be converted to Python scalars
>>> 

from numpy import *
from scipy.optimize import fmin_tnc

class LossFunction(object):
    def __init__(self, x, y):
      self.x = x
      self.y = y

    def __call__(self, abc):
      """ A function suitable for passing to the fmin() minimizers.
      """
      a, b, c = abc
      y = a*(1.0 + b*c*self.x) ** (-1.0/b)
      dy = self.y - y
      return dy*dy

x = array([1502.0, 1513.7,1517.5,1545.5,1578.9,1587.3,1600.4,1636.1,1682.9,1697.6,1813.4,1907.5])
y = array([0.28,0.22,0.26,0.18,0.12,0.13,0.09,0.07,0.06,0.05,0.01,0.01])

lf = LossFunction(x, y)
abc0 = [10., 2.5, 0.0]  
retcode, nfeval, abc_optimal = fmin_tnc(lf, abc0,  approx_grad=True, bounds=None, epsilon=1e-008)
  
 
---------------------------------
Don't pick lemons.
See all the new 2007 cars at Yahoo! Autos.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070313/c76f2558/attachment.html>

From jelle.feringa at ezct.net  Tue Mar 13 10:34:48 2007
From: jelle.feringa at ezct.net (Jelle Feringa / EZCT Architecture & Design Research)
Date: Tue, 13 Mar 2007 15:34:48 +0100
Subject: [SciPy-user] tvtk GL2PS
In-Reply-To: <17903.37463.446065.89579@gargle.gargle.HOWL>
References: <001501c75fe9$942463e0$c000a8c0@JELLE>
	<17903.37463.446065.89579@gargle.gargle.HOWL>
Message-ID: <37fac3b90703130734q4a6e5d33m3c710f0fe6aaccc6@mail.gmail.com>

That would be major progress!
Having mayavi/tvtk being able to produce postscripts (ps/eps/pdf) output
would be terrific!
It's a very important feauture since outputting for hardcopy is such a
common operation.

# bryce, would you be so kind to signal me if the vtk egg is updated? much
appreciated!

cheers,

-jelle


On 3/8/07, Prabhu Ramachandran <prabhu at aero.iitb.ac.in> wrote:
>
> >>>>> "Jelle" == Jelle Feringa / EZCT Architecture & Design Research <
> jelle.feringa at ezct.net> writes:
>
>     Jelle> Hi, Currently I'm using the enthon (win32, 2.4) 1.0.0
>     Jelle> release, which is terrific!
>
> Thanks to Bryce and the enthought build team (or is that a one man
> army?)  for all their work!
>
>     Jelle> Though there is a minor, but to me important glitch in
>     Jelle> (t)vtk;
> [...]
>     Jelle> Saving as a vector PS/EPS/PDF/TeX file using GL2PS is
>     Jelle> either not supported by your version of VTK or you have not
>     Jelle> configured VTK to work with GL2PS -- read the documentation
>     Jelle> for the vtkGL2PSExporter class.
>
>     Jelle> Which is really too bad.
>
> This would require Bryce to build VTK with the VTK_USE_GL2PS flag set
> to on for the next release.  There are no other dependencies that this
> brings up.
>
> cheers,
> prabhu
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070313/2ed04747/attachment.html>

From nauss at lcrs.de  Wed Mar 14 02:51:36 2007
From: nauss at lcrs.de (Thomas Nauss)
Date: Wed, 14 Mar 2007 07:51:36 +0100
Subject: [SciPy-user] Sort 3-D arrays by two columns
Message-ID: <45F79B78.4070806@lcrs.de>

Dear Experts,
I want plot some satellite datasets using matplotlib with basemap. For 
that task, I need a 3-D array (or three 1-D arrays) containing latitude, 
longitude, and corresponding data values. Since basemap needs latitude 
and longitude values in increasing order, I'm looking for a function to 
sort the 3-D array by the two lat/lon columns (i.e. first by latitude, 
then by longitude but conserving the value combinations).
Thanks,
Thomas



From robert.kern at gmail.com  Wed Mar 14 03:00:57 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 14 Mar 2007 02:00:57 -0500
Subject: [SciPy-user] Using SciPy/NumPy optimization
In-Reply-To: <971382.30683.qm@web57911.mail.re3.yahoo.com>
References: <971382.30683.qm@web57911.mail.re3.yahoo.com>
Message-ID: <45F79DA9.1020005@gmail.com>

lechtlr wrote:
> Hi there:
> 
> I am looking for a non-linear, constrained optimization tool, and thought fmin_tnc would do the job that I wanted to do.
> As a starter, I tried the attached script posted here recently. However, I get the following error when I run this script: 
> "TypeError: only length-1 arrays can be converted to Python scalars" 
> Can any one help me to figure out what I was doing wrong ?  

>     def __call__(self, abc):
>       """ A function suitable for passing to the fmin() minimizers.
>       """
>       a, b, c = abc
>       y = a*(1.0 + b*c*self.x) ** (-1.0/b)
>       dy = self.y - y
>       return dy*dy

Sorry, I screwed up this example. The function needs to return a scalar.

  return dot(dy, dy)

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From wbaxter at gmail.com  Wed Mar 14 03:09:22 2007
From: wbaxter at gmail.com (Bill Baxter)
Date: Wed, 14 Mar 2007 16:09:22 +0900
Subject: [SciPy-user] Sort 3-D arrays by two columns
In-Reply-To: <45F79B78.4070806@lcrs.de>
References: <45F79B78.4070806@lcrs.de>
Message-ID: <e86a5fd00703140009n2c60102fk565ed755fe773641@mail.gmail.com>

Does lexsort  do what you need?

Also, have you seen the numpy examples list?  It's a good place to
look for this sort of thing:
http://www.scipy.org/Numpy_Example_List
http://www.scipy.org/Numpy_Example_List_With_Doc

--bb

On 3/14/07, Thomas Nauss <nauss at lcrs.de> wrote:
> Dear Experts,
> I want plot some satellite datasets using matplotlib with basemap. For
> that task, I need a 3-D array (or three 1-D arrays) containing latitude,
> longitude, and corresponding data values. Since basemap needs latitude
> and longitude values in increasing order, I'm looking for a function to
> sort the 3-D array by the two lat/lon columns (i.e. first by latitude,
> then by longitude but conserving the value combinations).
> Thanks,
> Thomas
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From nwagner at iam.uni-stuttgart.de  Wed Mar 14 03:31:06 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Wed, 14 Mar 2007 08:31:06 +0100
Subject: [SciPy-user] Structured matrices
In-Reply-To: <759494.19283.qm@web414.biz.mail.mud.yahoo.com>
References: <759494.19283.qm@web414.biz.mail.mud.yahoo.com>
Message-ID: <45F7A4BA.7070302@iam.uni-stuttgart.de>

Barry Drake wrote:
> Nils,
> Does this matrix come from a particular application?  I'm working on
> algorithms for the non-negative matrix factorization (NMF).  With this
> matrix as input I'm getting some very strange results.  So I'm curious
> about potential applications.
>
> Cheers!
> Barry L. Drake
> GA Tech
>
Barry,

Yes indeed. The mass matrix of a multi-link pendulum (provided that the
equations of motion are linearized) exhibits this structure.
What do you mean by strange results ?

Here are some references

@article{Lobas,
         author={Lobas, L. G.},
         title={Generalized mathematical model of an inverted multilink
pendulum with follower force},
         journal={International Applied Mechanics},
         volume={41},
         pages={566-572},
         year={2005}}

@article{Gallina,
         author={Gallina, P.},
         title={About the stability of non-conservative undamped systems},
         journal={Journal of Sound and Vibration},
         volume={262},
         pages={977-988},
         year={2003}}

@article{Gasparini,
         author={Gasparini, A.~M. and Saetta, A.~V. and Vitaliani, R.~V.},
         title={On the stability and instability regions of non-conservative
                continuous system under partially follower forces},
         journal={Computer Methods in Applied Mechanics and Engineering},
         pages={63-78},
         volume={124},
         year={1995}}

I look forward to hearing from you.

Cheers,
                   Nils

> */Nils Wagner/* wrote:
>
>     Hi all,
>
>     I was wondering if the matrix family (see below) has a special name ?
>     And/or is there a way to construct this matrix via special matrices
>     (like Hankel, Toeplitz, etc.) ?
>
>
>     [[ 3. 1.]
>     [ 1. 1.]]
>
>     [[ 5. 3. 1.]
>     [ 3. 3. 1.]
>     [ 1. 1. 1.]]
>
>     [[ 7. 5. 3. 1.]
>     [ 5. 5. 3. 1.]
>     [ 3. 3. 3. 1.]
>     [ 1. 1. 1. 1.]]
>
>     [[ 9. 7. 5. 3. 1.]
>     [ 7. 7. 5. 3. 1.]
>     [ 5. 5. 5. 3. 1.]
>     [ 3. 3. 3. 3. 1.]
>     [ 1. 1. 1. 1. 1.]]
>
>     Nils
>
>     from scipy import *
>
>     def M(n):
>
>     tmp = ones((n,n))
>     for k in arange(n,0,-1):
>     tmp[:k-1,:k-1] = tmp[:k-1,:k-1] + 2
>     return tmp
>
>     for n in arange(2,8):
>     print M(n)
>     print
>
>     _______________________________________________
>     SciPy-user mailing list
>     SciPy-user at scipy.org
>     http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   
 


From nwagner at iam.uni-stuttgart.de  Wed Mar 14 04:05:32 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Wed, 14 Mar 2007 09:05:32 +0100
Subject: [SciPy-user] Structured matrices
In-Reply-To: <ce557a360703121052g59e7c2f0y364ea622d7b9fb55@mail.gmail.com>
References: <45F547E9.3050208@iam.uni-stuttgart.de>
	<ce557a360703121052g59e7c2f0y364ea622d7b9fb55@mail.gmail.com>
Message-ID: <45F7ACCC.8060300@iam.uni-stuttgart.de>

Anne Archibald wrote:
> On 12/03/07, Nils Wagner <nwagner at iam.uni-stuttgart.de> wrote:
>   
>> Hi all,
>>
>> I was wondering if the matrix family (see below) has a special name ?
>> And/or is there a way to construct this matrix via special matrices
>> (like Hankel, Toeplitz, etc.) ?
>>     
>
> Well, I've never come across them before, but they're easy to make directly:
>
> In [8]: minimum.outer(r_[7:-1:-2],r_[7:-1:-2])
> Out[8]:
> array([[7, 5, 3, 1],
>        [5, 5, 3, 1],
>        [3, 3, 3, 1],
>        [1, 1, 1, 1]])
>
> Anne
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   
I like one-liners. Thank you very much !

Nils
 


From pgmdevlist at gmail.com  Mon Mar 12 16:13:31 2007
From: pgmdevlist at gmail.com (Pierre GM)
Date: Mon, 12 Mar 2007 16:13:31 -0400
Subject: [SciPy-user] SVN: sandbox/pyloess
Message-ID: <200703121613.32441.pgmdevlist@gmail.com>

All,
A few weeks ago, an implementation of loess was requested on this list. I just 
uploaded a beta version in the 'pyloess' directory of the sandbox. There are 
actually one interface for two modules: lowess, the univariate  locally 
weighted linear estimation, and STL, the seasonal/trend decomposition. The 
modules are just f2py wrappers of the initial fortran sources (converted to 
double precision). I'd be happy to get your feedback, as usual.

I've been also trying to port the loess algorithm (a more generic version of 
lowess that can accept up to 8 variables). However, I've been hitting a wall 
on this one. The initial code is in C, with regular calls to fortran 
subroutines. The base object is a nested structure, and the operations are 
performed on this nested structure. 

Initially, I tried to code the nested structure in python directly, so that I 
would just have to call the C/fortran routines, but that didn't really work 
(the results from some tests didn't match the ones I was getting from the C 
code directly). I then tried to use SWIG, before realizing I was spending 
most of my time writing extensions. Eventually, I decided to stick to Pyrex. 
I wrote as many classes as sub-structure to get access to the underlying 
objects, plus an extra class for the combination of the substructures. The 
result is so far fairly disappointing, as the final object doesn't even get 
initialized properly. After one week of this "going-nowhere-fast", I'm giving 
up temporarily. 
Nevertheless, I uploaded this prototype in a 'sandbox' subfolder of the 
package, hoping that a kind soul would provide me with some pointers about 
the generic approach and some ideas about why I segfault all the time (I'm 
thinking about the pyrex specialists among us...). 

Thanks a lot in advance for your comments and feedback.

Pierre



From cimrman3 at ntc.zcu.cz  Wed Mar 14 09:21:25 2007
From: cimrman3 at ntc.zcu.cz (Robert Cimrman)
Date: Wed, 14 Mar 2007 14:21:25 +0100
Subject: [SciPy-user] sparse generalized eigenvalue solver
Message-ID: <45F7F6D5.2070500@ntc.zcu.cz>

Hi all,

I need to solve a problem of the type

D * u = nu * M * u, nu .. eigenvalues, u .. eigenvectors

where M = B * A^-1 B^T, the matrices come from a linear system of the
following structure:

A * u - B^T * d/dt p = 0
B * u + D * p = 0

For curious, A is a linear elasticity matrix, B^T, B discrete grad, div
operators and D a diffusion matrix; it is related to modeling deformable
porous medium.

I need all the eigenvalues/eigenvectors. M cannot be formed explicitely
in real situations, it will be available via its 'matvec' action. All
matrices are stored in the CSR format.

I am aware of the ARPACK being included in the scipy.sandbox - what is
its status? What other software (Python-ready or not) I should look at?

Thanks,
r.


From rshepard at appl-ecosys.com  Wed Mar 14 09:28:33 2007
From: rshepard at appl-ecosys.com (Rich Shepard)
Date: Wed, 14 Mar 2007 06:28:33 -0700 (PDT)
Subject: [SciPy-user] sparse generalized eigenvalue solver
In-Reply-To: <45F7F6D5.2070500@ntc.zcu.cz>
References: <45F7F6D5.2070500@ntc.zcu.cz>
Message-ID: <Pine.LNX.4.64.0703140628060.19694@salmo.appl-ecosys.com>

On Wed, 14 Mar 2007, Robert Cimrman wrote:

> I need all the eigenvalues/eigenvectors. M cannot be formed explicitely in
> real situations, it will be available via its 'matvec' action. All
> matrices are stored in the CSR format.

   Have you looked at eig() and found it wanting?

Rich

-- 
Richard B. Shepard, Ph.D.               |    The Environmental Permitting
Applied Ecosystem Services, Inc.        |          Accelerator(TM)
<http://www.appl-ecosys.com>     Voice: 503-667-4517      Fax: 503-667-8863


From nwagner at iam.uni-stuttgart.de  Wed Mar 14 09:34:27 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Wed, 14 Mar 2007 14:34:27 +0100
Subject: [SciPy-user] sparse generalized eigenvalue solver
In-Reply-To: <Pine.LNX.4.64.0703140628060.19694@salmo.appl-ecosys.com>
References: <45F7F6D5.2070500@ntc.zcu.cz>
	<Pine.LNX.4.64.0703140628060.19694@salmo.appl-ecosys.com>
Message-ID: <45F7F9E3.4090009@iam.uni-stuttgart.de>

Rich Shepard wrote:
> On Wed, 14 Mar 2007, Robert Cimrman wrote:
>
>   
>> I need all the eigenvalues/eigenvectors. M cannot be formed explicitely in
>> real situations, it will be available via its 'matvec' action. All
>> matrices are stored in the CSR format.
>>     
>
>    Have you looked at eig() and found it wanting?
>
> Rich
>
>   
eig is for dense matrices.
 
Nils



From nwagner at iam.uni-stuttgart.de  Wed Mar 14 09:37:51 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Wed, 14 Mar 2007 14:37:51 +0100
Subject: [SciPy-user] sparse generalized eigenvalue solver
In-Reply-To: <45F7F6D5.2070500@ntc.zcu.cz>
References: <45F7F6D5.2070500@ntc.zcu.cz>
Message-ID: <45F7FAAF.8090902@iam.uni-stuttgart.de>

Robert Cimrman wrote:
> Hi all,
>
> I need to solve a problem of the type
>
> D * u = nu * M * u, nu .. eigenvalues, u .. eigenvectors
>
> where M = B * A^-1 B^T, the matrices come from a linear system of the
> following structure:
>
> A * u - B^T * d/dt p = 0
> B * u + D * p = 0
>
> For curious, A is a linear elasticity matrix, B^T, B discrete grad, div
> operators and D a diffusion matrix; it is related to modeling deformable
> porous medium.
>
> I need all the eigenvalues/eigenvectors. M cannot be formed explicitely
> in real situations, it will be available via its 'matvec' action. All
> matrices are stored in the CSR format.
>
> I am aware of the ARPACK being included in the scipy.sandbox - what is
> its status? What other software (Python-ready or not) I should look at?
>
> Thanks,
> r.
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   


LOBPCG looks promising.
I have filed a ticket
http://projects.scipy.org/scipy/scipy/ticket/373

Trilinos includes Anasazi. Unfortunately Pytrilinos doesn't support
Anasazi right now.
http://software.sandia.gov/trilinos/packages/anasazi/

http://www.sandia.gov/~rblehou/*anasazi*-toms.pdf

There is SLEPc which is based on PETSc.
http://www.grycap.upv.es/slepc/

Cheers,


                     Nils

 


From sgarcia at olfac.univ-lyon1.fr  Wed Mar 14 09:52:12 2007
From: sgarcia at olfac.univ-lyon1.fr (Samuel GARCIA)
Date: Wed, 14 Mar 2007 14:52:12 +0100
Subject: [SciPy-user] stats.kstest
Message-ID: <45F7FE0C.2010501@olfac.univ-lyon1.fr>

Hi list,

stats.kstest permit to compare the distribution of a  sample with a
theorical distribution.

In matlab kstest exists to but there is also kstest2 that permit to
compare the distribution of two samples.
Is there a equivalent in scipy ?

Thanks

Sam

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Samuel Garcia
Laboratoire de Neurosciences Sensorielles, Comportement, Cognition.
CNRS - UMR5020 - Universite Claude Bernard LYON 1
Equipe logisique et technique
50, avenue Tony Garnier
69366 LYON Cedex 07
FRANCE
T?l : 04 37 28 74 64
Fax : 04 37 28 76 01
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~



From sgarcia at olfac.univ-lyon1.fr  Wed Mar 14 09:58:13 2007
From: sgarcia at olfac.univ-lyon1.fr (Samuel GARCIA)
Date: Wed, 14 Mar 2007 14:58:13 +0100
Subject: [SciPy-user] stats.kstest
In-Reply-To: <45F7FE0C.2010501@olfac.univ-lyon1.fr>
References: <45F7FE0C.2010501@olfac.univ-lyon1.fr>
Message-ID: <45F7FF75.9020408@olfac.univ-lyon1.fr>

I find it : stats.ks_2samp !

Samuel GARCIA wrote:
> Hi list,
>
> stats.kstest permit to compare the distribution of a  sample with a
> theorical distribution.
>
> In matlab kstest exists to but there is also kstest2 that permit to
> compare the distribution of two samples.
> Is there a equivalent in scipy ?
>
> Thanks
>
> Sam
>
>   

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Samuel Garcia
Laboratoire de Neurosciences Sensorielles, Comportement, Cognition.
CNRS - UMR5020 - Universite Claude Bernard LYON 1
Equipe logisique et technique
50, avenue Tony Garnier
69366 LYON Cedex 07
FRANCE
T?l : 04 37 28 74 64
Fax : 04 37 28 76 01
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~



From olivetti at itc.it  Wed Mar 14 10:17:25 2007
From: olivetti at itc.it (Emanuele Olivetti)
Date: Wed, 14 Mar 2007 15:17:25 +0100
Subject: [SciPy-user] clustering using custom distance
Message-ID: <45F803F5.20804@itc.it>

Hi,
I'm approaching clustering module in scipy but as far as I can see there
is only one distance measure available (euclidean distance). Is there a way
to easy embed a different distance measure between samples? Something
like correlation or in general custum functions I mean.

Thanks in advance,

Emanuele

------------------
ITC -> dall'1 marzo 2007 Fondazione Bruno Kessler
ITC -> since 1 March 2007 Fondazione Bruno Kessler
------------------



From zpincus at stanford.edu  Wed Mar 14 10:31:50 2007
From: zpincus at stanford.edu (Zachary Pincus)
Date: Wed, 14 Mar 2007 09:31:50 -0500
Subject: [SciPy-user] problems with ndimage spline filtering
Message-ID: <1755A525-A801-4F5A-8D68-64C93BCE88DD@stanford.edu>

Hello folks,

I've started to use the ndimage package in scipy, and had noticed  
that I got really ugly results from image interpolation with spline  
orders greater than 1.

I tracked this down to the 'pre-filtering' step with  
scipy.ndimage.spline_filter, which is (I presume) supposed to smooth  
out the image signal so that splines with large support can be used  
for interpolation.

Unfortunately, this function doesn't smooth anything; instead it  
appears to introduce very strong ringing artifacts (and not much  
else). I've attached example images, filtered with orders 2, 3, and  
4. The ringing is extremely pronounced on the piecewise-constant  
image (input2.png), but even for natural, continuous tone images  
(e.g. input1.png) there is a big problem.

Here are examples:
http://www.stanford.edu/~zpincus/spline.html

Note that in all cases, the image was processed as 64-bit floating  
point before clamping to 8-bit for output, so this isn't likely to be  
a precision problem. the filter command I used was:
 >>> filtered = scipy.ndimage.spline_filter(input_array, order=n)
where n varied.

Is this the expected results? Does anyone have any idea what's going  
on here?

Thanks,
Zach Pincus

Program in Biomedical Informatics and Department of Biochemistry
Stanford University School of Medicine




From rhc28 at cornell.edu  Wed Mar 14 10:47:41 2007
From: rhc28 at cornell.edu (Rob Clewley)
Date: Wed, 14 Mar 2007 10:47:41 -0400
Subject: [SciPy-user] Best way to structure dynamic system-of-systems
	simulations?
In-Reply-To: <1173758272.25956.1179135235@webmail.messagingengine.com>
References: <1172896076.28906.1177490111@webmail.messagingengine.com>
	<1173758272.25956.1179135235@webmail.messagingengine.com>
Message-ID: <a749952d0703140747l9cb4f05i686ff3e1556cb71b@mail.gmail.com>

Phil,

I'd happily give you some direction to get this working in PyDSTool,
which is the only python framework I know of that will give you this
kind of functionality. It will certainly run a lot faster than odeint.
Before branching out on your own, perhaps you'd consider talking to me
about solving your problem with our code, so that you can give us
constructive feedback about it. If people make the effort to
communicate with us we can speed up the maturation process. You can
offer us code too if you want to get more involved. We provide a
couple of example scripts for systems built from modular components in
the way you describe (PyDSTool/tests/ModelSpec_test.py and CIN.py).

Rob


From cimrman3 at ntc.zcu.cz  Wed Mar 14 11:18:59 2007
From: cimrman3 at ntc.zcu.cz (Robert Cimrman)
Date: Wed, 14 Mar 2007 16:18:59 +0100
Subject: [SciPy-user] sparse generalized eigenvalue solver
In-Reply-To: <45F7FAAF.8090902@iam.uni-stuttgart.de>
References: <45F7F6D5.2070500@ntc.zcu.cz>
	<45F7FAAF.8090902@iam.uni-stuttgart.de>
Message-ID: <45F81263.8050006@ntc.zcu.cz>

Nils Wagner wrote:
> 
> LOBPCG looks promising.
> I have filed a ticket
> http://projects.scipy.org/scipy/scipy/ticket/373
> 
> Trilinos includes Anasazi. Unfortunately Pytrilinos doesn't support
> Anasazi right now.
> http://software.sandia.gov/trilinos/packages/anasazi/
> 
> http://www.sandia.gov/~rblehou/*anasazi*-toms.pdf
> 
> There is SLEPc which is based on PETSc.
> http://www.grycap.upv.es/slepc/

Thanks for the links! LOBPCG looks small enough to be easily wrapped for
Python.

cheers,
r.


From nwagner at iam.uni-stuttgart.de  Wed Mar 14 11:25:42 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Wed, 14 Mar 2007 16:25:42 +0100
Subject: [SciPy-user] sparse generalized eigenvalue solver
In-Reply-To: <45F81263.8050006@ntc.zcu.cz>
References: <45F7F6D5.2070500@ntc.zcu.cz>	<45F7FAAF.8090902@iam.uni-stuttgart.de>
	<45F81263.8050006@ntc.zcu.cz>
Message-ID: <45F813F6.1010308@iam.uni-stuttgart.de>

Robert Cimrman wrote:
> Nils Wagner wrote:
>   
>> LOBPCG looks promising.
>> I have filed a ticket
>> http://projects.scipy.org/scipy/scipy/ticket/373
>>
>> Trilinos includes Anasazi. Unfortunately Pytrilinos doesn't support
>> Anasazi right now.
>> http://software.sandia.gov/trilinos/packages/anasazi/
>>
>> http://www.sandia.gov/~rblehou/*anasazi*-toms.pdf
>>
>> There is SLEPc which is based on PETSc.
>> http://www.grycap.upv.es/slepc/
>>     
>
> Thanks for the links! LOBPCG looks small enough to be easily wrapped for
> Python.
>
> cheers,
> r.
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   
Would you take the trouble and add LOBPCG to the sandbox ;-)  ?

Cheers,
                 Nils
 


From steveire at gmail.com  Wed Mar 14 11:33:44 2007
From: steveire at gmail.com (Stephen Kelly)
Date: Wed, 14 Mar 2007 15:33:44 +0000
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <45EEF5A5.6080202@ee.byu.edu>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
	<45ED948A.2040708@ee.byu.edu>
	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
	<45ED9A2F.3050803@ee.byu.edu>
	<e76aa17f0703070617j6aeeccf7g19f81896eba2e9d9@mail.gmail.com>
	<45EED6DF.9020502@american.edu> <45EEF5A5.6080202@ee.byu.edu>
Message-ID: <18fbbe5a0703140833o4cc196a4pa653439f60a6bd4e@mail.gmail.com>

Interesting discussion. I'd obviously prefer that the copy/paste takes
place, but I'd like to know what's going to happen now that the
discussion has taken place.

On 3/7/07, Travis Oliphant <oliphant at ee.byu.edu> wrote:
>  Alan Isaac wrote:
>
> >Matthieu Brucher wrote:
> >
> >
> >>I'm opposed to this as well, interpolation is not linear algebra, it is
> >>signal processing, and as Souheil said, why not nd-interpolation with
> >>B-Splines then ?
> >>
> >>
> >
> >I am not taking sides on this issue, but I do
> >take issue with trying to settle it with this
> >kind of "argument".  A rhetorical question does
> >not set out the issues.  For example, treating
> >a rhetorical question as a real question for a
> >moment, one might respond that a simple 1-d
> >interpolation would
> >- find wide use among people who need nothing more
> >- enhance backward compatibility
> >- impose minimum maintenance requirements
> >
> >Again, I am not taking sides on the issue, so I
> >am in no way suggesting that such points are
> >decisive.  However opposition that fails to
> >address such points is hardly decisive either.
> >
> >My point of reference would be: if it is likely
> >to drain even modest developer effort away from
> >core NumPy issues, then it is harmful. Otherwise
> >it is harmless as long as it introduces no new
> >dependencies.
> >
> >
> This is an impoirtant point for me.  I'm only considering adding interp
> because it's already done and just a matter of cut-and-paste.
>
> My take on this, is that I would rather use developer time to make scipy
> more modular and easier to install (this usually translates to binary
> builds for multiple platforms).   That way you could grab just the
> modules of interest.  Ideally, grabbing these modules is a matter of
> using setuptools (or some other package management concept) to do it
> automatically as needed.
>
> -Travis
>
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From ryanlists at gmail.com  Wed Mar 14 11:45:58 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Wed, 14 Mar 2007 10:45:58 -0500
Subject: [SciPy-user] problem with _num.seterr when importing scipy
In-Reply-To: <45F77C74.1090901@gmail.com>
References: <200702091522.01125.krlong@sandia.gov> <45CCE67E.1090508@gmail.com>
	<200702091531.11770.krlong@sandia.gov>
	<c5b438120703130703g79e10145l3e7fdbfb4bce8685@mail.gmail.com>
	<45F77C74.1090901@gmail.com>
Message-ID: <c5b438120703140845h34f5036aoccbbc8e0ba9b373b@mail.gmail.com>

So is this a manifestation of trying to run versions of numpy and
scipy that don't go together?

On 3/13/07, Robert Kern <robert.kern at gmail.com> wrote:
> Ryan Krauss wrote:
> > I was having the same problem.  Did this ever get solved?
>
> The 'all' keyword was introduced after version 1.0b5, the version that Kevin
> Long was using. That was the problem.
>
> --
> Robert Kern
>
> "I have come to believe that the whole world is an enigma, a harmless enigma
>  that is made terrible by our own mad attempt to interpret it as though it had
>  an underlying truth."
>   -- Umberto Eco
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From anand at soe.ucsc.edu  Wed Mar 14 12:06:58 2007
From: anand at soe.ucsc.edu (Anand Patil)
Date: Wed, 14 Mar 2007 09:06:58 -0700
Subject: [SciPy-user] Structured matrices
In-Reply-To: <mailman.58.1173850445.29466.scipy-user@scipy.org>
References: <mailman.58.1173850445.29466.scipy-user@scipy.org>
Message-ID: <45F81DA2.2000201@cse.ucsc.edu>

Nils, Barry,

That matrix is also the covariance of Brownian motion evaluated at times [9, 7, 5, 3, 1]. The inverse matrix bears a strong resemblance to the numerical second derivative operator. This makes sense because the second derivative in x of min(x, \xi) is a delta function concentrated at \xi, but there's probably a more important connection that I don't know about.

Maybe this helps explain the strange results? I found some related refs, but haven't looked at them: http://citeseer.ist.psu.edu/347057.html, http://adsabs.harvard.edu/abs/2003EAEJA.....2001X .

Cheers,
Anand


--------------------------

Nils,
Does this matrix come from a particular application?  I'm working on algorithms for the non-negative matrix factorization (NMF).  With this matrix as input I'm getting some very strange results.  So I'm curious about potential applications.

Cheers!
Barry L. Drake
GA Tech

Nils Wagner wrote: Hi all,

I was wondering if the matrix family (see below) has a special name ?
And/or is there a way to construct this matrix via special matrices
(like Hankel, Toeplitz, etc.) ?


[[ 3.  1.]
 [ 1.  1.]]

[[ 5.  3.  1.]
 [ 3.  3.  1.]
 [ 1.  1.  1.]]

[[ 7.  5.  3.  1.]
 [ 5.  5.  3.  1.]
 [ 3.  3.  3.  1.]
 [ 1.  1.  1.  1.]]

[[ 9.  7.  5.  3.  1.]
 [ 7.  7.  5.  3.  1.]
 [ 5.  5.  5.  3.  1.]
 [ 3.  3.  3.  3.  1.]
 [ 1.  1.  1.  1.  1.]]

Nils



From lucas.barcelos at gmail.com  Wed Mar 14 13:08:14 2007
From: lucas.barcelos at gmail.com (Lucas Barcelos de Oliveira)
Date: Wed, 14 Mar 2007 14:08:14 -0300
Subject: [SciPy-user] Help building CVXOPT for W32/Python 2.3
Message-ID: <b815eb2e0703141008s41d1e2eeob3ae9687be7abfa8@mail.gmail.com>

Hello all,

It's been a week now since I started trying to build cvxopt for python
2.3under Win XP with no success. I've build ATLAS using Cygwin
(followed the
instructions in http://www.scipy.org/Installing_SciPy/Windows) and manage to
create the lib files, but when I try to compile, using the command
/cygdrive/c/Python23/python setup.py config --compiler=cygwin build
--compiler=cygwin, I get a lot undefined references to several functions:
C:\cygwin\bin\gcc.exe -mcygwin -shared -s build\temp.win32-
2.3\Release\c\base.o build\temp.win32-2.3\Release\c\dense.obuild\temp.win32-
2.3\Release\c\sparse.o
build\temp.win32-2.3\Release\c\base.def-L/cygdrive/c/BLAS-LAPACK/
-Lc:\Python23\libs -Lc:
\Python23\PCBuild -lm -llapack -lblas -lg2c -lpython23 -o build\lib.win32-
2.3\cvxopt\base.pyd
build\temp.win32-2.3\Release\c\base.o:base.c:(.text+0x587): undefined
reference to `_dscal'
build\temp.win32-2.3\Release\c\base.o:base.c:(.text+0x62f): undefined
reference to `_zscal'
...
build\temp.win32-2.3\Release\c\dense.o:dense.c:(.text+0x57cb): undefined
reference to `_cpow'
build\temp.win32-2.3\Release\c\dense.o:dense.c:(.text+0x5914): undefined
reference to `__assert'
...
build\temp.win32-2.3\Release\c\sparse.o:sparse.c:(.text+0xf8b6): undefined
reference to `__assert'
collect2: ld returned 1 exit status

Since I couldn't find a solution for this, I tried compiling with Borland
C++ compiler, here is the result;
C:\cvxopt-0.8.2\src>python setup.py build --compiler=bcpp
running build
running build_py
running build_ext
building 'base' extension
C:\Borland\BCC55\Bin\bcc32.exe -c /tWM /O2 /q /g0 -IC:\Python23\include
-IC:\Python23\PC  -obuild\temp.win32-2.3\Release\c\dense.obj C\dense.c
C\dense.c:
Fatal F1003 c:\Borland\Bcc55\include\stdcomp.h 5: Error directive: Must use
C++ for STDCOMP.H
*** 1 errors in Compile ***
error: command 'bcc32.exe' failed with exit status 1

My next attempt was to use Visual Studio 6. The result:
C:\cvxopt-0.8.2\src>python setup.py build
running build
running build_py
running build_ext
building 'base' extension
C:\Program Files\Microsoft Visual Studio\VC98\BIN\cl.exe /c /nologo /Ox /MD
/W3 /GX /DNDEBUG -IC:\Python23\include -IC:\Python23\PC /TcC/sparse.c
/Fobuild\temp.win32-2.3\Release\C/sparse.obj
sparse.c
C/cvxopt.h(13) : fatal error C1083: Cannot open include file: 'complex.h':
No such file or directory
error: command '"C:\Program Files\Microsoft Visual Studio\VC98\BIN\cl.exe"'
failed with exit status 2

I've checked MSVC6 include dir and there really isn't a complex.h file, but
there is a COMPLEX file, if I rename it to COMPLEX.H i get the error:
C:\Program Files\Microsoft Visual Studio\VC98\INCLUDE\eh.h(32) : fatal error
C1189: #error :  "eh.h is only for C++!"
error: command '"C:\Program Files\Microsoft Visual Studio\VC98\BIN\cl.exe"'
failed with exit status 2

So I am pretty lost and desperate, I wish I could upgrade my Python version
to use a win32 installer but I am stuck at 2.3 because of the traffic
simulator I use (AIMSUN). Any help will be really appreciated!

Best regards,

-- 
----------------------------------------------------------
         Lucas Barcelos de Oliveira
P?s Controle e Automa??o - UFSC
         lucas.barcelos at gmail.com
----------------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070314/4e9cb6b2/attachment.html>

From robert.kern at gmail.com  Wed Mar 14 13:38:22 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 14 Mar 2007 12:38:22 -0500
Subject: [SciPy-user] problem with _num.seterr when importing scipy
In-Reply-To: <c5b438120703140845h34f5036aoccbbc8e0ba9b373b@mail.gmail.com>
References: <200702091522.01125.krlong@sandia.gov>
	<45CCE67E.1090508@gmail.com>	<200702091531.11770.krlong@sandia.gov>	<c5b438120703130703g79e10145l3e7fdbfb4bce8685@mail.gmail.com>	<45F77C74.1090901@gmail.com>
	<c5b438120703140845h34f5036aoccbbc8e0ba9b373b@mail.gmail.com>
Message-ID: <45F8330E.9070603@gmail.com>

Ryan Krauss wrote:
> So is this a manifestation of trying to run versions of numpy and
> scipy that don't go together?

Yes.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From robert.kern at gmail.com  Wed Mar 14 13:40:28 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 14 Mar 2007 12:40:28 -0500
Subject: [SciPy-user] Help building CVXOPT for W32/Python 2.3
In-Reply-To: <b815eb2e0703141008s41d1e2eeob3ae9687be7abfa8@mail.gmail.com>
References: <b815eb2e0703141008s41d1e2eeob3ae9687be7abfa8@mail.gmail.com>
Message-ID: <45F8338C.9050108@gmail.com>

Lucas Barcelos de Oliveira wrote:
> Hello all,
> 
> It's been a week now since I started trying to build cvxopt for python
> 2.3 under Win XP with no success.

CVXOPT is not a scipy project, and I don't know of anyone here who uses it. I
recommend asking the authors for help.

  http://www.ee.ucla.edu/~vandenbe/cvxopt/

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From lechtlr at yahoo.com  Wed Mar 14 13:52:08 2007
From: lechtlr at yahoo.com (lechtlr)
Date: Wed, 14 Mar 2007 10:52:08 -0700 (PDT)
Subject: [SciPy-user] Using SciPy/NumPy optimization
Message-ID: <929095.44776.qm@web57903.mail.re3.yahoo.com>

Robert:

Thanks for the fmin_tnc example and it's working now. However, I have trouble getting optimal parameters for the model.  I have tied varies options from 'fmin_tnc', and nothing seems to work so far.  I would  appreciate, if you can give some clues to get optimal values.  I have attached my python script.

Thanks,
Lex


from numpy import *
from scipy.optimize import fmin_tnc

class LossFunction(object):
    def __init__(self, x, y):
      self.x = x
      self.y = y

    def __call__(self, abc):
      """ A function suitable for passing to the fmin() minimizers.
      """
      a, b, c = abc
      y = a*(1.0 + b*c*self.x) ** (-1.0/b)
      dy = self.y - y
      return dot(dy,dy)

#Generating y for given set of x
a_true = 100.0
b_true = 1.0 
c_true = 10.0 

T = range(1000, 2500, 10)

data = zeros([len(T), 2], 'd')

for m in range(len(T)):
    error = rand(1)/T[m]
    r = a_true*(1.0 + b_true*c_true*T[m])**(-1.0/b_true) + error
    data[m][0] = T[m]
    data[m][1] = r
    

x = data[:,0]
y = data[:,1]

print 'numer of data points:', len(T)

lf = LossFunction(x, y)
abc0 = [10.0, 1.0, 5.0]  
retcode, nfeval, abc_optimal = fmin_tnc(lf, abc0, bounds=None, approx_grad=True)

print 'retcode:', retcode
print 'nfeval:', nfeval
print 'Optimal Parameters:', abc_optimal


   
 
---------------------------------
Expecting? Get great news right away with email Auto-Check.
Try the Yahoo! Mail Beta.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070314/2a45d35e/attachment.html>

From corzneffect at gmail.com  Wed Mar 14 14:36:49 2007
From: corzneffect at gmail.com (Cory Davis)
Date: Wed, 14 Mar 2007 18:36:49 +0000
Subject: [SciPy-user] import numpy segmentation fault
Message-ID: <4948f97c0703141136j522fd32kbe202874ff19e3cf@mail.gmail.com>

Hi there,

I have just installed numpy-1.0.1 from source, which seemed to go
fine.  However when I try to "import numpy" I get a segmentation
fault.

A have a 64 bit machine running RedHat Enterprise Linux and Python 2.34

Any clues greatly appreciated.

Cheers,
Cory.


From zpincus at stanford.edu  Wed Mar 14 14:49:10 2007
From: zpincus at stanford.edu (Zachary Pincus)
Date: Wed, 14 Mar 2007 13:49:10 -0500
Subject: [SciPy-user] [Numpy-discussion] import numpy segmentation fault
In-Reply-To: <4948f97c0703141136j522fd32kbe202874ff19e3cf@mail.gmail.com>
References: <4948f97c0703141136j522fd32kbe202874ff19e3cf@mail.gmail.com>
Message-ID: <0F58E0ED-B2C4-4CF6-8EE3-4A25BEC73854@stanford.edu>

If I recall correctly, there's a bug in numpy 1.0.1 on Linux-x86-64  
that causes this segfault. This is fixed in the latest SVN version of  
numpy, so if you can grab that, it should work.

I can't find the trac ticket, but I ran into this some weeks ago.

Zach



On Mar 14, 2007, at 1:36 PM, Cory Davis wrote:

> Hi there,
>
> I have just installed numpy-1.0.1 from source, which seemed to go
> fine.  However when I try to "import numpy" I get a segmentation
> fault.
>
> A have a 64 bit machine running RedHat Enterprise Linux and Python  
> 2.34
>
> Any clues greatly appreciated.
>
> Cheers,
> Cory.
> _______________________________________________
> Numpy-discussion mailing list
> Numpy-discussion at scipy.org
> http://projects.scipy.org/mailman/listinfo/numpy-discussion



From stefan at sun.ac.za  Wed Mar 14 15:08:39 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Wed, 14 Mar 2007 21:08:39 +0200
Subject: [SciPy-user] problems with ndimage spline filtering
In-Reply-To: <1755A525-A801-4F5A-8D68-64C93BCE88DD@stanford.edu>
References: <1755A525-A801-4F5A-8D68-64C93BCE88DD@stanford.edu>
Message-ID: <20070314190839.GH7562@mentat.za.net>

Hi Zachary

On Wed, Mar 14, 2007 at 09:31:50AM -0500, Zachary Pincus wrote:
> I've started to use the ndimage package in scipy, and had noticed  
> that I got really ugly results from image interpolation with spline  
> orders greater than 1.

I've noticed this too, hence

http://projects.scipy.org/scipy/scipy/ticket/213

Unfortunately I havn't had time to find the source of the problem.
Please add your information to the ticket, hopefully we can narrow it
down somewhat.

Cheers
St?fan


From stefan at sun.ac.za  Wed Mar 14 15:08:39 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Wed, 14 Mar 2007 21:08:39 +0200
Subject: [SciPy-user] problems with ndimage spline filtering
In-Reply-To: <1755A525-A801-4F5A-8D68-64C93BCE88DD@stanford.edu>
References: <1755A525-A801-4F5A-8D68-64C93BCE88DD@stanford.edu>
Message-ID: <20070314190839.GH7562@mentat.za.net>

Hi Zachary

On Wed, Mar 14, 2007 at 09:31:50AM -0500, Zachary Pincus wrote:
> I've started to use the ndimage package in scipy, and had noticed  
> that I got really ugly results from image interpolation with spline  
> orders greater than 1.

I've noticed this too, hence

http://projects.scipy.org/scipy/scipy/ticket/213

Unfortunately I havn't had time to find the source of the problem.
Please add your information to the ticket, hopefully we can narrow it
down somewhat.

Cheers
St?fan


From ryanlists at gmail.com  Wed Mar 14 16:38:46 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Wed, 14 Mar 2007 15:38:46 -0500
Subject: [SciPy-user] scipy.multiply and scipy.divide
Message-ID: <c5b438120703141338xe8d108dm4adef1be1f927965@mail.gmail.com>

I was about to reuse some old code of mine and noticed that it uses
scipy.multiply and scipy.divide which are ufuncs for element-wise
multiplication and division.  With the new API for numpy, is this any
different from just using a*b or a/b is a and b are vectors (i.e.
arrays with shape (n,))?

Thanks,

Ryan


From robert.kern at gmail.com  Wed Mar 14 16:51:54 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 14 Mar 2007 15:51:54 -0500
Subject: [SciPy-user] scipy.multiply and scipy.divide
In-Reply-To: <c5b438120703141338xe8d108dm4adef1be1f927965@mail.gmail.com>
References: <c5b438120703141338xe8d108dm4adef1be1f927965@mail.gmail.com>
Message-ID: <45F8606A.7000308@gmail.com>

Ryan Krauss wrote:
> I was about to reuse some old code of mine and noticed that it uses
> scipy.multiply and scipy.divide which are ufuncs for element-wise
> multiplication and division.  With the new API for numpy, is this any
> different from just using a*b or a/b is a and b are vectors (i.e.
> arrays with shape (n,))?

Not any more. I assume that you used the scipy (actually scipy_base) functions
instead of the operators because they were the faster/nan-tastic versions at the
time.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From tom.denniston at alum.dartmouth.org  Wed Mar 14 17:11:42 2007
From: tom.denniston at alum.dartmouth.org (Tom Denniston)
Date: Wed, 14 Mar 2007 16:11:42 -0500
Subject: [SciPy-user] Sort 3-D arrays by two columns
In-Reply-To: <e86a5fd00703140009n2c60102fk565ed755fe773641@mail.gmail.com>
References: <45F79B78.4070806@lcrs.de>
	<e86a5fd00703140009n2c60102fk565ed755fe773641@mail.gmail.com>
Message-ID: <d6f2d3dd0703141411h25d81027lcec73eef04bb4861@mail.gmail.com>

I think you want lexsort.  It takes a tuple of columns.

On 3/14/07, Bill Baxter <wbaxter at gmail.com> wrote:
> Does lexsort  do what you need?
>
> Also, have you seen the numpy examples list?  It's a good place to
> look for this sort of thing:
> http://www.scipy.org/Numpy_Example_List
> http://www.scipy.org/Numpy_Example_List_With_Doc
>
> --bb
>
> On 3/14/07, Thomas Nauss <nauss at lcrs.de> wrote:
> > Dear Experts,
> > I want plot some satellite datasets using matplotlib with basemap. For
> > that task, I need a 3-D array (or three 1-D arrays) containing latitude,
> > longitude, and corresponding data values. Since basemap needs latitude
> > and longitude values in increasing order, I'm looking for a function to
> > sort the 3-D array by the two lat/lon columns (i.e. first by latitude,
> > then by longitude but conserving the value combinations).
> > Thanks,
> > Thomas
> >
> > _______________________________________________
> > SciPy-user mailing list
> > SciPy-user at scipy.org
> > http://projects.scipy.org/mailman/listinfo/scipy-user
> >
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From corzneffect at gmail.com  Thu Mar 15 08:20:22 2007
From: corzneffect at gmail.com (Cory Davis)
Date: Thu, 15 Mar 2007 12:20:22 +0000
Subject: [SciPy-user] New Scipy not finding ATLAS
Message-ID: <4948f97c0703150520o7b63e71t289d43165a5d4d5d@mail.gmail.com>

Hi All,
Previously I had scipy-0.4.9 working perfectly with the complete
atlas3.6.0 libraries.  But now for some reason scipy-0.5.2 is not
seeing these during installation, where for both 0.4.9 and 0.5.2 the
location of the atlas libraries is specified by the ATLAS environment
variable.  Can anyone tell me what is going on?  What has changed
between 0.4.9 and 0.5.2 to make this happen?

If it helps, I have a 64 bit Linux machine and Python 2.3.4

Cheers,
Cory


From corzneffect at gmail.com  Thu Mar 15 12:32:40 2007
From: corzneffect at gmail.com (Cory Davis)
Date: Thu, 15 Mar 2007 16:32:40 +0000
Subject: [SciPy-user] New Scipy not finding ATLAS
In-Reply-To: <4948f97c0703150520o7b63e71t289d43165a5d4d5d@mail.gmail.com>
References: <4948f97c0703150520o7b63e71t289d43165a5d4d5d@mail.gmail.com>
Message-ID: <4948f97c0703150932x52039e75y549f9d9a73f9a710@mail.gmail.com>

Solved.  It seems you have to edit the numpy site.cfg file.

On 3/15/07, Cory Davis <corzneffect at gmail.com> wrote:
> Hi All,
> Previously I had scipy-0.4.9 working perfectly with the complete
> atlas3.6.0 libraries.  But now for some reason scipy-0.5.2 is not
> seeing these during installation, where for both 0.4.9 and 0.5.2 the
> location of the atlas libraries is specified by the ATLAS environment
> variable.  Can anyone tell me what is going on?  What has changed
> between 0.4.9 and 0.5.2 to make this happen?
>
> If it helps, I have a 64 bit Linux machine and Python 2.3.4
>
> Cheers,
> Cory
>


From ldavid at MIT.EDU  Thu Mar 15 12:36:59 2007
From: ldavid at MIT.EDU (Lawrence David)
Date: Thu, 15 Mar 2007 12:36:59 -0400
Subject: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9: _fftpack.so
	problems?
Message-ID: <0118D1DC-BB70-44F3-8B6A-1329EA267CDF@mit.edu>

Hi there,

I'm pulling the little hair I have out trying to install SciPy on my  
G4 PPC running OS 10.4.9.

I've followed the instructions on the SciPy.org website: http:// 
www.scipy.org/Installing_SciPy/Mac_OS_X.  I've got gcc 4.0.1 and  
gfortran (gnu95 -> 4.3).

Running build: "python setup.py build_src build_clib -- 
fcompiler=gnu95 build_ext --fcompiler=gnu95 build", yields a long  
error message that terminates with a litany of "Undefined symbols"  
and this last message:

collect2: ld returned 1 exit status
error: Command "/usr/local/bin/gfortran -L/sw/lib build/ 
temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/Lib/fftpack/ 
_fftpackmodule.o build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/ 
zfft.o build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/drfft.o build/ 
temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zrfft.o build/ 
temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfftnd.o build/ 
temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/ 
fortranobject.o -L/usr/local/lib -L/usr/local/lib/gcc/powerpc-apple- 
darwin8.8.0/4.3.0 -Lbuild/temp.macosx-10.3-fat-2.5 -ldfftpack -lfftw3  
-lgfortran -o build/lib.macosx-10.3-fat-2.5/scipy/fftpack/ 
_fftpack.so" failed with exit status 1

I've noticed several other folks have had similar error messages on  
this board, but none of the posted solutions worked for me.  If  
anyone could offer any suggestions on what to try, I'd love to hear  
them!

(I've even tried installing the SuperPack binary, which lies through  
its teeth when it says that it's installed correctly; import scipy  
still fails.  Perhaps my problem is that I'm using gfortran?  Can  
anyone tell me how to revert to g77?)

Much thanks!!
- lawrence

http://www.stinkpot.org


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070315/1a4b50b5/attachment.html>

From lorenzo.isella at gmail.com  Thu Mar 15 12:49:42 2007
From: lorenzo.isella at gmail.com (Lorenzo Isella)
Date: Thu, 15 Mar 2007 17:49:42 +0100
Subject: [SciPy-user] Vectorize vs Map
Message-ID: <a2b3004b0703150949r6b4ab649g104386816ec951e6@mail.gmail.com>

Dear All,
Probably another newbie question: I like quite a lot the vectorize()
command which allows me to skip iterations on functions, but the map()
command on a list performs a similar task if I am not mistaken.
Is there any reason to favour one above the other or is it just a
matter of taste?
Kind Regards

Lorenzo


From robert.kern at gmail.com  Thu Mar 15 12:58:52 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Thu, 15 Mar 2007 11:58:52 -0500
Subject: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9: _fftpack.so
 problems?
In-Reply-To: <0118D1DC-BB70-44F3-8B6A-1329EA267CDF@mit.edu>
References: <0118D1DC-BB70-44F3-8B6A-1329EA267CDF@mit.edu>
Message-ID: <45F97B4C.4080404@gmail.com>

Lawrence David wrote:
> Hi there,
> 
> I'm pulling the little hair I have out trying to install SciPy on my G4
> PPC running OS 10.4.9.
> 
> I've followed the instructions on the SciPy.org
> website: http://www.scipy.org/Installing_SciPy/Mac_OS_X.  I've got gcc
> 4.0.1 and gfortran (gnu95 -> 4.3).
> 
> Running build: "python setup.py build_src build_clib --fcompiler=gnu95
> build_ext --fcompiler=gnu95 build", yields a long error message that
> terminates with a litany of "Undefined symbols" and this last message:
> 
> collect2: ld returned 1 exit status
> error: Command "/usr/local/bin/gfortran -L/sw/lib
> build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/Lib/fftpack/_fftpackmodule.o
> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfft.o
> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/drfft.o
> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zrfft.o
> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfftnd.o
> build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/fortranobject.o
> -L/usr/local/lib -L/usr/local/lib/gcc/powerpc-apple-darwin8.8.0/4.3.0
> -Lbuild/temp.macosx-10.3-fat-2.5 -ldfftpack -lfftw3 -lgfortran -o
> build/lib.macosx-10.3-fat-2.5/scipy/fftpack/_fftpack.so" failed with
> exit status 1
> 
> I've noticed several other folks have had similar error messages on this
> board, but none of the posted solutions worked for me.  If anyone could
> offer any suggestions on what to try, I'd love to hear them!  

It looks like you have LDFLAGS defined (thus, the -L/sw/lib). That *replaces*
all of the Python-specific linker flags that distutils provides.

> (I've even tried installing the SuperPack binary, which lies through its
> teeth when it says that it's installed correctly; import scipy still
> fails.  Perhaps my problem is that I'm using gfortran?  Can anyone tell
> me how to revert to g77?)

You can't with a Universal Python. It needs gcc 4, which requires gfortran, not g77.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From oliphant at ee.byu.edu  Thu Mar 15 12:58:56 2007
From: oliphant at ee.byu.edu (Travis Oliphant)
Date: Thu, 15 Mar 2007 09:58:56 -0700
Subject: [SciPy-user] Vectorize vs Map
In-Reply-To: <a2b3004b0703150949r6b4ab649g104386816ec951e6@mail.gmail.com>
References: <a2b3004b0703150949r6b4ab649g104386816ec951e6@mail.gmail.com>
Message-ID: <45F97B50.9050109@ee.byu.edu>

Lorenzo Isella wrote:

>Dear All,
>Probably another newbie question: I like quite a lot the vectorize()
>command which allows me to skip iterations on functions, but the map()
>command on a list performs a similar task if I am not mistaken.
>Is there any reason to favour one above the other or is it just a
>matter of taste?
>  
>
map only works with sequences (i.e. 1-d arrays) and does not do 
broadcasting for multiple-valued functions.

vectorize works with multiple dimensional arrays and also does 
broadcasting.

I have not done any timings but I suspect there are lots of cases where 
map is probably faster because it does less.

-Travis





From robert.kern at gmail.com  Thu Mar 15 13:04:07 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Thu, 15 Mar 2007 12:04:07 -0500
Subject: [SciPy-user] Vectorize vs Map
In-Reply-To: <a2b3004b0703150949r6b4ab649g104386816ec951e6@mail.gmail.com>
References: <a2b3004b0703150949r6b4ab649g104386816ec951e6@mail.gmail.com>
Message-ID: <45F97C87.8080506@gmail.com>

Lorenzo Isella wrote:
> Dear All,
> Probably another newbie question: I like quite a lot the vectorize()
> command which allows me to skip iterations on functions, but the map()
> command on a list performs a similar task if I am not mistaken.
> Is there any reason to favour one above the other or is it just a
> matter of taste?

vectorize() takes a Python function and turns it into a ufunc. ufuncs do a lot
more than map() does. They can take multidimensional arrays. n-ary ufuncs can
take multiple inputs and broadcast them against each other. ufuncs have methods
like .inner() and .reduce() which are quite powerful.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From fperez.net at gmail.com  Thu Mar 15 13:15:34 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Thu, 15 Mar 2007 11:15:34 -0600
Subject: [SciPy-user] Vectorize vs Map
In-Reply-To: <45F97C87.8080506@gmail.com>
References: <a2b3004b0703150949r6b4ab649g104386816ec951e6@mail.gmail.com>
	<45F97C87.8080506@gmail.com>
Message-ID: <db6b5ecc0703151015p445872c5n744c5484eea6d589@mail.gmail.com>

On 3/15/07, Robert Kern <robert.kern at gmail.com> wrote:
> Lorenzo Isella wrote:
> > Dear All,
> > Probably another newbie question: I like quite a lot the vectorize()
> > command which allows me to skip iterations on functions, but the map()
> > command on a list performs a similar task if I am not mistaken.
> > Is there any reason to favour one above the other or is it just a
> > matter of taste?
>
> vectorize() takes a Python function and turns it into a ufunc. ufuncs do a lot
> more than map() does. They can take multidimensional arrays. n-ary ufuncs can
> take multiple inputs and broadcast them against each other. ufuncs have methods
> like .inner() and .reduce() which are quite powerful.

Mmh, isn't that what 'frompyfunc' does instead?  vectorize doesn't
seem to produce a true ufunc.  Perhaps I'm just misunderstanding
something:

In [10]: def foo(x):return x
   ....:

In [11]: vfoo = N.vectorize(foo)

In [12]: vfoo.outer?
Object `vfoo.outer` not found.

In [13]: type vfoo
-------> type(vfoo)
Out[13]: <class 'numpy.lib.function_base.vectorize'>

In [14]: ufoo = N.frompyfunc(foo,1,1)

In [15]: type ufoo
-------> type(ufoo)
Out[15]: <type 'numpy.ufunc'>

In [16]: ufoo.outer?
Type:           builtin_function_or_method
Base Class:     <type 'builtin_function_or_method'>
Namespace:      Interactive
Docstring:
    <no docstring>


I'm actually not exactly sure why the two do exist, to be honest.  I
realize frompyfunc has (probably by necessity) a more complicated API,
but it does return a true ufunc, which vectorize() doesn't.

Insight welcome.

Cheers,

f


From robert.kern at gmail.com  Thu Mar 15 13:17:24 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Thu, 15 Mar 2007 12:17:24 -0500
Subject: [SciPy-user] Vectorize vs Map
In-Reply-To: <db6b5ecc0703151015p445872c5n744c5484eea6d589@mail.gmail.com>
References: <a2b3004b0703150949r6b4ab649g104386816ec951e6@mail.gmail.com>	<45F97C87.8080506@gmail.com>
	<db6b5ecc0703151015p445872c5n744c5484eea6d589@mail.gmail.com>
Message-ID: <45F97FA4.6050902@gmail.com>

Fernando Perez wrote:
> On 3/15/07, Robert Kern <robert.kern at gmail.com> wrote:
>> Lorenzo Isella wrote:
>>> Dear All,
>>> Probably another newbie question: I like quite a lot the vectorize()
>>> command which allows me to skip iterations on functions, but the map()
>>> command on a list performs a similar task if I am not mistaken.
>>> Is there any reason to favour one above the other or is it just a
>>> matter of taste?
>> vectorize() takes a Python function and turns it into a ufunc. ufuncs do a lot
>> more than map() does. They can take multidimensional arrays. n-ary ufuncs can
>> take multiple inputs and broadcast them against each other. ufuncs have methods
>> like .inner() and .reduce() which are quite powerful.
> 
> Mmh, isn't that what 'frompyfunc' does instead?  vectorize doesn't
> seem to produce a true ufunc.  Perhaps I'm just misunderstanding
> something:

No, just my poor memory. I don't use either.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From cscheid at sci.utah.edu  Thu Mar 15 14:04:24 2007
From: cscheid at sci.utah.edu (Carlos Scheidegger)
Date: Thu, 15 Mar 2007 12:04:24 -0600
Subject: [SciPy-user] should vectorize honor array/matrix distinctions?
Message-ID: <45F98AA8.30408@sci.utah.edu>

Hi,

should scipy.vectorize honor array/matrix distinctions?

http://www.scipy.org/NumPy_for_Matlab_Users claims "[returning an array when
given a matrix] shouldn't happen with NumPy functions (if it does it's a
bug)". However, this is what I get on my amd64 ubuntu edgy box:

$ python
Python 2.4.4c1 (#2, Oct 11 2006, 20:00:03)
[GCC 4.1.2 20060928 (prerelease) (Ubuntu 4.1.1-13ubuntu5)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import scipy
>>> import numpy
>>> scipy.__version__
'0.5.1'
>>> numpy.__version__
'1.0rc1'
>>> import math
>>> def v(x): return int(math.log(abs(x) * 10.0, 10.0))
...
>>> m = scipy.matrix([[1.0, 5.0], [5.0, 10.0]])
>>> a = scipy.array([[1.0, 5.0], [5.0, 10.0]])
>>> vv = scipy.vectorize(v)
>>> vv(m)
array([[1, 1],
       [1, 2]])
>>> vv(a)
array([[1, 1],
       [1, 2]])
>>> type(m)
<class 'numpy.core.defmatrix.matrix'>
>>> type(a)
<type 'numpy.ndarray'>
>>> type(vv(m))
<type 'numpy.ndarray'>
>>> type(vv(a))
<type 'numpy.ndarray'>

Is this expected behavior? I know there are trivial workarounds, but I just
wanted to clarify.

FWIW, I'm using the universe numpy package. Thank you very much for your time,
-carlos


From ldavid at MIT.EDU  Thu Mar 15 15:07:26 2007
From: ldavid at MIT.EDU (Lawrence David)
Date: Thu, 15 Mar 2007 15:07:26 -0400
Subject: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9: _fftpack.so
	problems?
In-Reply-To: <45F97B4C.4080404@gmail.com>
References: <0118D1DC-BB70-44F3-8B6A-1329EA267CDF@mit.edu>
	<45F97B4C.4080404@gmail.com>
Message-ID: <77114BB9-AB85-4BB2-85F5-DE324676EC3C@mit.edu>

Hot darn Batman - you nailed it!!

Much thanks,
- lawrence

http://www.stinkpot.org


On Mar 15, 2007, at 12:58 PM, Robert Kern wrote:

> Lawrence David wrote:
>> Hi there,
>>
>> I'm pulling the little hair I have out trying to install SciPy on  
>> my G4
>> PPC running OS 10.4.9.
>>
>> I've followed the instructions on the SciPy.org
>> website: http://www.scipy.org/Installing_SciPy/Mac_OS_X.  I've got  
>> gcc
>> 4.0.1 and gfortran (gnu95 -> 4.3).
>>
>> Running build: "python setup.py build_src build_clib -- 
>> fcompiler=gnu95
>> build_ext --fcompiler=gnu95 build", yields a long error message that
>> terminates with a litany of "Undefined symbols" and this last  
>> message:
>>
>> collect2: ld returned 1 exit status
>> error: Command "/usr/local/bin/gfortran -L/sw/lib
>> build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/Lib/ 
>> fftpack/_fftpackmodule.o
>> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfft.o
>> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/drfft.o
>> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zrfft.o
>> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfftnd.o
>> build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/ 
>> fortranobject.o
>> -L/usr/local/lib -L/usr/local/lib/gcc/powerpc-apple-darwin8.8.0/4.3.0
>> -Lbuild/temp.macosx-10.3-fat-2.5 -ldfftpack -lfftw3 -lgfortran -o
>> build/lib.macosx-10.3-fat-2.5/scipy/fftpack/_fftpack.so" failed with
>> exit status 1
>>
>> I've noticed several other folks have had similar error messages  
>> on this
>> board, but none of the posted solutions worked for me.  If anyone  
>> could
>> offer any suggestions on what to try, I'd love to hear them!
>
> It looks like you have LDFLAGS defined (thus, the -L/sw/lib). That  
> *replaces*
> all of the Python-specific linker flags that distutils provides.
>
>> (I've even tried installing the SuperPack binary, which lies  
>> through its
>> teeth when it says that it's installed correctly; import scipy still
>> fails.  Perhaps my problem is that I'm using gfortran?  Can anyone  
>> tell
>> me how to revert to g77?)
>
> You can't with a Universal Python. It needs gcc 4, which requires  
> gfortran, not g77.
>
> -- 
> Robert Kern
>
> "I have come to believe that the whole world is an enigma, a  
> harmless enigma
>  that is made terrible by our own mad attempt to interpret it as  
> though it had
>  an underlying truth."
>   -- Umberto Eco
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070315/f41dfcf0/attachment.html>

From lbolla at gmail.com  Fri Mar 16 05:37:21 2007
From: lbolla at gmail.com (lorenzo bolla)
Date: Fri, 16 Mar 2007 10:37:21 +0100
Subject: [SciPy-user] Vectorize vs Map
In-Reply-To: <45F97FA4.6050902@gmail.com>
References: <a2b3004b0703150949r6b4ab649g104386816ec951e6@mail.gmail.com>
	<45F97C87.8080506@gmail.com>
	<db6b5ecc0703151015p445872c5n744c5484eea6d589@mail.gmail.com>
	<45F97FA4.6050902@gmail.com>
Message-ID: <80c99e790703160237n1367c895pba2ba31c9a3bf9d@mail.gmail.com>

here is a simple timing on my machine.

---------------------------------------------------

import scipy as S
from timeit import Timer

def f(x):
    return S.sqrt(S.absolute(x)**2)

x = S.array(S.rand(1000))
fv = S.vectorize(f)
fu = S.frompyfunc(f,1,1)

def test():
    #f(x)
    #fv(x)
    fu(x)

if __name__ == '__main__':
    t = Timer('test()', 'from __main__ import test')
    n = 100
    print "%.2f usec/pass" % (1e6*t.timeit(number=n)/n)

---------------------------------------------------

I get:
 229.84 usec/pass for f(x) 119410.40 usec/pass for fv(x) 114513.80 usec/pass
for fu(x)

vectorize and frompyfunc create functions roughly 500 times slower than the
one using ndarrays arithmetics (even if it cannot operate on lists, just
ndarrays).

lorenzo.


On 3/15/07, Robert Kern <robert.kern at gmail.com> wrote:
>
> Fernando Perez wrote:
> > On 3/15/07, Robert Kern <robert.kern at gmail.com> wrote:
> >> Lorenzo Isella wrote:
> >>> Dear All,
> >>> Probably another newbie question: I like quite a lot the vectorize()
> >>> command which allows me to skip iterations on functions, but the map()
> >>> command on a list performs a similar task if I am not mistaken.
> >>> Is there any reason to favour one above the other or is it just a
> >>> matter of taste?
> >> vectorize() takes a Python function and turns it into a ufunc. ufuncs
> do a lot
> >> more than map() does. They can take multidimensional arrays. n-ary
> ufuncs can
> >> take multiple inputs and broadcast them against each other. ufuncs have
> methods
> >> like .inner() and .reduce() which are quite powerful.
> >
> > Mmh, isn't that what 'frompyfunc' does instead?  vectorize doesn't
> > seem to produce a true ufunc.  Perhaps I'm just misunderstanding
> > something:
>
> No, just my poor memory. I don't use either.
>
> --
> Robert Kern
>
> "I have come to believe that the whole world is an enigma, a harmless
> enigma
> that is made terrible by our own mad attempt to interpret it as though it
> had
> an underlying truth."
> -- Umberto Eco
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070316/769f01d2/attachment.html>

From sgarcia at olfac.univ-lyon1.fr  Fri Mar 16 05:40:47 2007
From: sgarcia at olfac.univ-lyon1.fr (Samuel GARCIA)
Date: Fri, 16 Mar 2007 10:40:47 +0100
Subject: [SciPy-user] Problem with weave blitz win32
Message-ID: <45FA661F.9020004@olfac.univ-lyon1.fr>

Hi list,

I have problem running weave on windowsXP.

Version :
Python 2.5
MiGW 3.4.2
Scipy 0.5.2
Numpy 1.0.1


array3d.py give me :

numpy:
[[[ 0  1  2  3]
  [ 4  5  6  7]
  [ 8  9 10 11]]

 [[12 13 14 15]
  [16 17 18 19]
  [20 21 22 23]]]
Pure Inline:
<weave: compiling>
img[  0][  0]=   0   1   2   3
img[  0][  1]=   4   5   6   7
img[  0][  2]=   8   9  10  11
img[  1][  0]=  12  13  14  15
img[  1][  1]=  16  17  18  19
img[  1][  2]=  20  21  22  23
Blitz Inline:
<weave: compiling>
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:5: 
warning: ignoring #pragma warning
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:6: 
warning: ignoring #pragma warning
In file included from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/applics.h:400,
                 from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/vecexpr.h:32,
                 from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/vecpick.cc:16,
                 from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/vecpick.h:293,
                 from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/vector.h:449,
                 from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/tinyvec.h:430,
                 from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array-impl.h:44,
                 from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array.h:32,
                 from 
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:9:
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/mathfunc.h: In 
static member function `static double 
blitz::_bz_expm1<P_numtype1>::apply(P_numtype1)':
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/mathfunc.h:1353: 
error: `::expm1' has not been declared
In file included from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array/funcs.h:29,
                 from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array/newet.h:29,
                 from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array/et.h:27,
                 from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array-impl.h:2515,
                 from 
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array.h:32,
                 from 
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:9:
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/funcs.h: In static 
member function `static T_numtype1 
blitz::Fn_expm1<T_numtype1>::apply(T_numtype1)':
G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/funcs.h:113: 
error: `::expm1' has not been declared
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp: 
In function `PyObject* file_to_py(FILE*, char*, char*)':
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:404: 
warning: unused variable 'py_obj'
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp: 
In function `blitz::Array<P_numtype, N_rank> 
convert_to_blitz(PyArrayObject*, const char*) [with T = int, int N = 3]':
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:708:   
instantiated from here
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:657: 
warning: unused variable 'stride_acc'
Traceback (most recent call last):
  File "array3d.py", line 105, in <module>
    main()
  File "array3d.py", line 101, in main
    blitz_inline(arr)
  File "array3d.py", line 89, in blitz_inline
    weave.inline(code, ['arr'], type_converters=converters.blitz)
  File "G:\Python25\Lib\site-packages\scipy\weave\inline_tools.py", line 
339, in inline
    **kw)
  File "G:\Python25\Lib\site-packages\scipy\weave\inline_tools.py", line 
447, in compile_function
    verbose=verbose, **kw)
  File "G:\Python25\Lib\site-packages\scipy\weave\ext_tools.py", line 
365, in compile
    verbose = verbose, **kw)
  File "G:\Python25\Lib\site-packages\scipy\weave\build_tools.py", line 
269, in build_extension
    setup(name = module_name, ext_modules = [ext],verbose=verb)
  File "G:\Python25\Lib\site-packages\numpy\distutils\core.py", line 
174, in setup
    return old_setup(**new_attr)
  File "G:\Python25\lib\distutils\core.py", line 168, in setup
    raise SystemExit, "error: " + str(msg)
distutils.errors.CompileError: error: Command "g++ -mno-cygwin -O2 -Wall 
-IG:\Python25\lib\site-packages\scipy\weave 
-IG:\Python25\lib\site-packages\scipy\weave\scxx 
-IG:\Python25\lib\site-packages\scipy\weave\blitz 
-IG:\Python25\lib\site-packages\numpy\core\include -IG:\Python25\include 
-IG:\Python25\PC -c 
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp 
-o 
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_intermediate\compiler_894ad5ed761bb51736c6d2b7872dc212\Release\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.o" 
failed with exit status 1


And weave.test() give me :


Found 0 tests for scipy.weave.c_spec
  Found 2 tests for scipy.weave.blitz_tools
building extensions here: 
g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\m7
  Found 1 tests for scipy.weave.ext_tools
  Found 9 tests for scipy.weave.build_tools
  Found 0 tests for scipy.weave.inline_tools
  Found 1 tests for scipy.weave.ast_tools
Warning: FAILURE importing tests for <module 'scipy.weave.wx_spec' from 
'...ckages\\scipy\\weave\\wx_spec.pyc'>
G:\Python25\Lib\site-packages\scipy\weave\tests\test_wx_spec.py:16: 
ImportError: No module named wxPython (in <module>)
  Found 3 tests for scipy.weave.standard_array_spec
  Found 74 tests for scipy.weave.size_check
  Found 26 tests for scipy.weave.catalog
  Found 16 tests for scipy.weave.slice_handler
  Found 0 tests for __main__
...warning: specified build_dir '_bad_path_' does not exist or is not 
writable. Trying default locations
.....warning: specified build_dir '_bad_path_' does not exist or is not 
writable. Trying default locations
....................................F..F..................................................................removing 
'g:\docume~1\sgarcia\locals~1\temp\tmpga71mmcat_test' (and everything 
under it)
error removing g:\docume~1\sgarcia\locals~1\temp\tmpga71mmcat_test: 
g:\docume~1\sgarcia\locals~1\temp\tmpga71mmcat_test: Le r?pertoire n'est 
pas vide
.removing 'g:\docume~1\sgarcia\locals~1\temp\tmpimmvs6cat_test' (and 
everything under it)
.................
======================================================================
FAIL: check_1d_3 
(scipy.weave.tests.test_size_check.test_dummy_array_indexing)
----------------------------------------------------------------------
Traceback (most recent call last):
  File 
"G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py", 
line 168, in check_1d_3
    self.generic_1d('a[-11:]')
  File 
"G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py", 
line 135, in generic_1d
    self.generic_wrap(a,expr)
  File 
"G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py", 
line 127, in generic_wrap
    self.generic_test(a,expr,desired)
  File 
"G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py", 
line 123, in generic_test
    assert_array_equal(actual,desired, expr)
  File "G:\Python25\lib\site-packages\numpy\testing\utils.py", line 223, 
in assert_array_equal
    verbose=verbose, header='Arrays are not equal')
  File "G:\Python25\lib\site-packages\numpy\testing\utils.py", line 215, 
in assert_array_compare
    assert cond, msg
AssertionError:
Arrays are not equal
a[-11:]
(mismatch 100.0%)
 x: array([1])
 y: array([10])

======================================================================
FAIL: check_1d_6 
(scipy.weave.tests.test_size_check.test_dummy_array_indexing)
----------------------------------------------------------------------
Traceback (most recent call last):
  File 
"G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py", 
line 174, in check_1d_6
    self.generic_1d('a[:-11]')
  File 
"G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py", 
line 135, in generic_1d
    self.generic_wrap(a,expr)
  File 
"G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py", 
line 127, in generic_wrap
    self.generic_test(a,expr,desired)
  File 
"G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py", 
line 123, in generic_test
    assert_array_equal(actual,desired, expr)
  File "G:\Python25\lib\site-packages\numpy\testing\utils.py", line 223, 
in assert_array_equal
    verbose=verbose, header='Arrays are not equal')
  File "G:\Python25\lib\site-packages\numpy\testing\utils.py", line 215, 
in assert_array_compare
    assert cond, msg
AssertionError:
Arrays are not equal
a[:-11]
(mismatch 100.0%)
 x: array([9])
 y: array([0])

----------------------------------------------------------------------
Ran 132 tests in 2.266s

FAILED (failures=2)
 >Exit code: 0


And of course my main problem is that some of my code that work on linux 
debian does'nt work on win32.
(But it used to work on the old enthon distribution version!)

Any idea ?

Thank

Sam



-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Samuel Garcia
Laboratoire de Neurosciences Sensorielles, Comportement, Cognition.
CNRS - UMR5020 - Universite Claude Bernard LYON 1
Equipe logisique et technique
50, avenue Tony Garnier
69366 LYON Cedex 07
FRANCE
T?l : 04 37 28 74 64
Fax : 04 37 28 76 01
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070316/bec1d904/attachment.html>

From nauss at lcrs.de  Fri Mar 16 06:16:55 2007
From: nauss at lcrs.de (Thomas Nauss)
Date: Fri, 16 Mar 2007 11:16:55 +0100
Subject: [SciPy-user] Sort 3-D arrays by two columns
In-Reply-To: <d6f2d3dd0703141411h25d81027lcec73eef04bb4861@mail.gmail.com>
References: <45F79B78.4070806@lcrs.de>	<e86a5fd00703140009n2c60102fk565ed755fe773641@mail.gmail.com>
	<d6f2d3dd0703141411h25d81027lcec73eef04bb4861@mail.gmail.com>
Message-ID: <45FA6E97.5060505@lcrs.de>

Thank's a lot! That helps.
Thomas

Tom Denniston wrote:
> I think you want lexsort.  It takes a tuple of columns.
> 
> On 3/14/07, Bill Baxter <wbaxter at gmail.com> wrote:
>> Does lexsort  do what you need?
>>
>> Also, have you seen the numpy examples list?  It's a good place to
>> look for this sort of thing:
>> http://www.scipy.org/Numpy_Example_List
>> http://www.scipy.org/Numpy_Example_List_With_Doc
>>
>> --bb
>>
>> On 3/14/07, Thomas Nauss <nauss at lcrs.de> wrote:
>>> Dear Experts,
>>> I want plot some satellite datasets using matplotlib with basemap. For
>>> that task, I need a 3-D array (or three 1-D arrays) containing latitude,
>>> longitude, and corresponding data values. Since basemap needs latitude
>>> and longitude values in increasing order, I'm looking for a function to
>>> sort the 3-D array by the two lat/lon columns (i.e. first by latitude,
>>> then by longitude but conserving the value combinations).
>>> Thanks,
>>> Thomas
>>>
>>> _______________________________________________
>>> SciPy-user mailing list
>>> SciPy-user at scipy.org
>>> http://projects.scipy.org/mailman/listinfo/scipy-user
>>>
>> _______________________________________________
>> SciPy-user mailing list
>> SciPy-user at scipy.org
>> http://projects.scipy.org/mailman/listinfo/scipy-user
>>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
> 

-- 

_____________________________________________
Dr. Thomas Nauss
Philipps-University of Marburg
Department of Geography
Laboratory for Climatology and Remote Sensing
Deutschhausstr. 10
D-35032 Marburg
fon:   ++49(6421)2824252
fax:   ++49(6421)2828950
mobile:++49(163)1462714
mail:  nauss at lcrs.de
web:   http://www.lcrs.de


From ryanlists at gmail.com  Fri Mar 16 10:18:16 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Fri, 16 Mar 2007 09:18:16 -0500
Subject: [SciPy-user] install from source problem
Message-ID: <c5b438120703160718g70d65eb4g78e8aa65abe830d1@mail.gmail.com>

I am trying to install from source and am running into some problems.
The first problem (that I don't know if it is really a problem) is
that the svn check out seems to go well, but at the very end I get
this message:
svn: REPORT request failed on '/svn/numpy/!svn/vcc/default'
svn: REPORT of '/svn/numpy/!svn/vcc/default': 200 OK (http://svn.scipy.org)

Is that an issue?  I am running svn version 1.3.1 if that matters.

This may be a problem because the svn check out only has the folders
f2py  fft  random
under numpy, so that when I try "python setup.py build" I get:
Traceback (most recent call last):
  File "setup.py", line 89, in ?
    setup_package()
  File "setup.py", line 59, in setup_package
    from numpy.distutils.core import setup
ImportError: No module named numpy.distutils.core

Am I doing something wrong?

Thanks,

Ryan


From stefan at sun.ac.za  Fri Mar 16 10:43:55 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Fri, 16 Mar 2007 16:43:55 +0200
Subject: [SciPy-user] install from source problem
In-Reply-To: <c5b438120703160718g70d65eb4g78e8aa65abe830d1@mail.gmail.com>
References: <c5b438120703160718g70d65eb4g78e8aa65abe830d1@mail.gmail.com>
Message-ID: <20070316144355.GL6168@mentat.za.net>

Hi Ryan

On Fri, Mar 16, 2007 at 09:18:16AM -0500, Ryan Krauss wrote:
> I am trying to install from source and am running into some problems.
> The first problem (that I don't know if it is really a problem) is
> that the svn check out seems to go well, but at the very end I get
> this message:
> svn: REPORT request failed on '/svn/numpy/!svn/vcc/default'
> svn: REPORT of '/svn/numpy/!svn/vcc/default': 200 OK
> (http://svn.scipy.org)

Sounds like a proxy problem.  Try

https://svn.scipy.org

instead of

http://svn.scipy.org

and see if that helps?

Cheers
St?fan


From ryanlists at gmail.com  Fri Mar 16 11:02:41 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Fri, 16 Mar 2007 10:02:41 -0500
Subject: [SciPy-user] install from source problem
In-Reply-To: <c5b438120703160718g70d65eb4g78e8aa65abe830d1@mail.gmail.com>
References: <c5b438120703160718g70d65eb4g78e8aa65abe830d1@mail.gmail.com>
Message-ID: <c5b438120703160802j30a6b4a1ofafdb9e1028ebba3@mail.gmail.com>

Thank you Stefan, that did it.

On 3/16/07, Ryan Krauss <ryanlists at gmail.com> wrote:
> I am trying to install from source and am running into some problems.
> The first problem (that I don't know if it is really a problem) is
> that the svn check out seems to go well, but at the very end I get
> this message:
> svn: REPORT request failed on '/svn/numpy/!svn/vcc/default'
> svn: REPORT of '/svn/numpy/!svn/vcc/default': 200 OK (http://svn.scipy.org)
>
> Is that an issue?  I am running svn version 1.3.1 if that matters.
>
> This may be a problem because the svn check out only has the folders
> f2py  fft  random
> under numpy, so that when I try "python setup.py build" I get:
> Traceback (most recent call last):
>   File "setup.py", line 89, in ?
>     setup_package()
>   File "setup.py", line 59, in setup_package
>     from numpy.distutils.core import setup
> ImportError: No module named numpy.distutils.core
>
> Am I doing something wrong?
>
> Thanks,
>
> Ryan
>


From ryanlists at gmail.com  Fri Mar 16 11:09:59 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Fri, 16 Mar 2007 10:09:59 -0500
Subject: [SciPy-user] clapack and cblas modules are empty
In-Reply-To: <4570D4D8.30806@gmail.com>
References: <1165001752.5095.20.camel@jrpirone-desktop>
	<4570D4D8.30806@gmail.com>
Message-ID: <c5b438120703160809y5ab0b07as83d39d795ed3943b@mail.gmail.com>

I think I have this exact situation and don't understand the final
answer.  Do I need to be concerned with the warning that my clapack
and cblas modules are empty?

ryan at ubuntu:~$ ldd /usr/lib/python2.4/site-packages/numpy/linalg/lapack_lite.so
        linux-gate.so.1 =>  (0xffffe000)
        liblapack.so.3 => /usr/lib/atlas/sse2/liblapack.so.3 (0xb7945000)
        libblas.so.3 => /usr/lib/atlas/sse2/libblas.so.3 (0xb73a8000)
        libg2c.so.0 => /usr/lib/libg2c.so.0 (0xb7381000)
        libm.so.6 => /lib/tls/i686/cmov/libm.so.6 (0xb735f000)
        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0xb7355000)
        libc.so.6 => /lib/tls/i686/cmov/libc.so.6 (0xb7226000)
        /lib/ld-linux.so.2 (0x80000000)

ryan at ubuntu:~$ locate *clapack.so
/usr/lib/python2.4/site-packages/scipy/lib/lapack/clapack.so
/usr/lib/python2.4/site-packages/scipy/linalg/clapack.so

ryan at ubuntu:/usr/lib/python2.4/site-packages/scipy/lib/lapack$ ls -al
total 704
drwxr-xr-x 3 root root   4096 2007-03-16 09:50 .
drwxr-xr-x 4 root root   4096 2007-03-16 09:50 ..
-rwxr-xr-x 1 root root  13022 2007-03-16 09:49 atlas_version.so
-rwxr-xr-x 1 root root 103466 2007-03-16 09:49 calc_lwork.so
-rwxr-xr-x 1 root root  48398 2007-03-16 09:49 clapack.so
-rwxr-xr-x 1 root root 494047 2007-03-16 09:49 flapack.so
.
.
.


What does it all mean?

Thanks,

Ryan



On 12/1/06, Robert Kern <robert.kern at gmail.com> wrote:
> jrpirone wrote:
> > Hello,
> >
> > I've installed scipy and numpy from Andrew Straw's repository, all other
> > libraries are from the standard Ubuntu repositories (6.10).  When I
> > execute scipy.test(), the following warnings appear:
> >
> > ****************************************************************
> > WARNING: clapack module is empty
> > -----------
> > See scipy/INSTALL.txt for troubleshooting.
> > Notes:
> > * If atlas library is not found by numpy/distutils/system_info.py,
> >   then scipy uses flapack instead of clapack.
> > ****************************************************************
> >
> > ****************************************************************
> > WARNING: cblas module is empty
> > -----------
> > See scipy/INSTALL.txt for troubleshooting.
> > Notes:
> > * If atlas library is not found by numpy/distutils/system_info.py,
> >   then scipy uses fblas instead of cblas.
> > ****************************************************************
> >
> > Running the command:
> > ldd /usr/lib/python2.4/site-packages/numpy/linalg/lapack_lite.so
> >
> > Gives me:
> >         linux-gate.so.1 =>  (0xffffe000)
> >         liblapack.so.3 => /usr/lib/atlas/sse2/liblapack.so.3
> > (0xb7958000)
> >         libblas.so.3 => /usr/lib/atlas/sse2/libblas.so.3 (0xb7387000)
> >         libg2c.so.0 => /usr/lib/libg2c.so.0 (0xb735f000)
> >         libm.so.6 => /lib/tls/i686/cmov/libm.so.6 (0xb7339000)
> >         libgcc_s.so.1 => /lib/libgcc_s.so.1 (0xb732e000)
> >         libc.so.6 => /lib/tls/i686/cmov/libc.so.6 (0xb71fa000)
> >         /lib/ld-linux.so.2 (0x80000000)
> >
> > Is there a problem with the libraries I have installed?  Should I be
> > worried about these warnings, or am I just being compulsive?
>
> They're not *necessary* but they are nice to have. It demonstrates that scipy
> was not compiled with ATLAS.
>
> --
> Robert Kern
>
> "I have come to believe that the whole world is an enigma, a harmless enigma
>  that is made terrible by our own mad attempt to interpret it as though it had
>  an underlying truth."
>   -- Umberto Eco
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From hirani at cs.uiuc.edu  Fri Mar 16 11:10:09 2007
From: hirani at cs.uiuc.edu (Anil N. Hirani)
Date: Fri, 16 Mar 2007 10:10:09 -0500
Subject: [SciPy-user] Question about test failures
Message-ID: <899620A2-9F3A-4CE0-B089-F0D26DD0F6E5@cs.uiuc.edu>

I built and installed scipy from the svn respository (svn co http:// 
svn.scipy.org/svn/scipy/trunk scipy)
on March 15, on a Mac OS X 10.4.8 with python2.5. Running the scipy  
test(1,10) gave 3 faiilures.
What is curious is that even though the dot product fails in the test  
(see second and third FAIL below),
the same expression gives the right result when I typed it in :

 >>> dot([3j,-4,3-4j],[2,3,1])
(-9+2j)

Can someone please let me know the possible reasons, fixes and  
consequences of the
test failures listed below.

Regards
Anil Hirani



======================================================================
FAIL: check loadmat case sparse
----------------------------------------------------------------------
Traceback (most recent call last):
   File "/sw/lib/python2.5/site-packages/scipy/io/tests/test_mio.py",  
line 85, in cc
     self._check_case(name, files, expected)
   File "/sw/lib/python2.5/site-packages/scipy/io/tests/test_mio.py",  
line 80, in _check_case
     self._check_level(k_label, expected, matdict[k])
   File "/sw/lib/python2.5/site-packages/scipy/io/tests/test_mio.py",  
line 63, in _check_level
     decimal = 5)
   File "/sw/lib/python2.5/site-packages/numpy/testing/utils.py",  
line 230, in assert_array_almost_equal
     header='Arrays are not almost equal')
   File "/sw/lib/python2.5/site-packages/numpy/testing/utils.py",  
line 215, in assert_array_compare
     assert cond, msg
AssertionError:
Arrays are not almost equal
test sparse; file /sw/lib/python2.5/site-packages/scipy/io/tests/./ 
data/testsparse_6.5.1_GLNX86.mat, variable testsparse
(mismatch 46.6666666667%)
x: array([[  3.03865194e-319,   3.16202013e-322,   1.04346664e-320,
           2.05531309e-320,   2.56123631e-320],
        [  3.16202013e-322,   0.00000000e+000,   0.00000000e+000,...
y: array([[ 1.,  2.,  3.,  4.,  5.],
        [ 2.,  0.,  0.,  0.,  0.],
        [ 3.,  0.,  0.,  0.,  0.]])

======================================================================
FAIL: check_dot (scipy.lib.tests.test_blas.test_fblas1_simple)
----------------------------------------------------------------------
Traceback (most recent call last):
   File "/sw/lib/python2.5/site-packages/scipy/lib/blas/tests/ 
test_blas.py", line 76, in check_dot
     assert_almost_equal(f([3j,-4,3-4j],[2,3,1]),-9+2j)
   File "/sw/lib/python2.5/site-packages/numpy/testing/utils.py",  
line 156, in assert_almost_equal
     assert round(abs(desired - actual),decimal) == 0, msg
AssertionError:
Items are not equal:
ACTUAL: (-1.998711109161377+1.0175286878527778e-36j)
DESIRED: (-9+2j)

======================================================================
FAIL: check_dot (scipy.linalg.tests.test_blas.test_fblas1_simple)
----------------------------------------------------------------------
Traceback (most recent call last):
   File "/sw/lib/python2.5/site-packages/scipy/linalg/tests/ 
test_blas.py", line 75, in check_dot
     assert_almost_equal(f([3j,-4,3-4j],[2,3,1]),-9+2j)
   File "/sw/lib/python2.5/site-packages/numpy/testing/utils.py",  
line 156, in assert_almost_equal
     assert round(abs(desired - actual),decimal) == 0, msg
AssertionError:
Items are not equal:
ACTUAL: (-1.998711109161377+1.3368407079941133e-36j)
DESIRED: (-9+2j)



From oliphant at ee.byu.edu  Fri Mar 16 12:13:45 2007
From: oliphant at ee.byu.edu (Travis Oliphant)
Date: Fri, 16 Mar 2007 09:13:45 -0700
Subject: [SciPy-user] Vectorize vs Map
In-Reply-To: <80c99e790703160237n1367c895pba2ba31c9a3bf9d@mail.gmail.com>
References: <a2b3004b0703150949r6b4ab649g104386816ec951e6@mail.gmail.com>	<45F97C87.8080506@gmail.com>	<db6b5ecc0703151015p445872c5n744c5484eea6d589@mail.gmail.com>	<45F97FA4.6050902@gmail.com>
	<80c99e790703160237n1367c895pba2ba31c9a3bf9d@mail.gmail.com>
Message-ID: <45FAC239.9030803@ee.byu.edu>

lorenzo bolla wrote:

> here is a simple timing on my machine.
>  
> ---------------------------------------------------
>  
> import scipy as S
> from timeit import Timer
>
> def f(x):
>     return S.sqrt(S.absolute(x)**2)
>
> x = S.array(S.rand(1000))
> fv = S.vectorize(f)
> fu = S.frompyfunc(f,1,1)
>
> def test():
>     #f(x)
>     #fv(x)
>     fu(x)
>
> if __name__ == '__main__':
>     t = Timer('test()', 'from __main__ import test')
>     n = 100
>     print "%.2f usec/pass" % (1e6*t.timeit(number=n)/n)
>
>
> ---------------------------------------------------
>  
> I get:
> 229.84 usec/pass for f(x)
> 119410.40 usec/pass for fv(x)
> 114513.80 usec/pass for fu(x)
>  
> vectorize and frompyfunc create functions roughly 500 times slower 
> than the one using ndarrays arithmetics (even if it cannot operate on 
> lists, just ndarrays).
>  


There is nothing new here.  It is not surprising at all.   From pyfunc 
creates a ufunc out of a python function.   The python function is 
called  at each element-by-element calculation.  In this case, the 
element-by-element calculation is also using ufuncs to compute the 
result (ufuncs are a very slow way to compute a single scalar operation). 

Because a Python function is called, the ufunc uses object arrays as well.

All of these intermediate things means the result will be much slower.   
Vectorize and frompyfunc are only useful when you want a quick ufunc 
that you can't figure out how to vectorize yourself. 

If you can vectorize it, it is always better to do it than to use 
vectorize.

-Travis



From alexander.borghgraef.rma at gmail.com  Fri Mar 16 12:42:42 2007
From: alexander.borghgraef.rma at gmail.com (Alexander Borghgraef)
Date: Fri, 16 Mar 2007 17:42:42 +0100
Subject: [SciPy-user] Gaussian pyramid using ndimage
Message-ID: <9e8c52a20703160942j3d1b4ca5ubbfab6dabe39f7c0@mail.gmail.com>

Hi all,

 I've been looking into the ndimage module functionality, and I was
wondering whether there were any downsampling filters
implemented within it, like used in gaussian pyramid implementations.
Obviously, convolution followed by downsampling would
do the trick, but that requires an intermediate storing of a full-sized
image, which isn't all that memory-efficient. Any suggestions?

-- 
Alex Borghgraef
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070316/d211d20d/attachment.html>

From bldrake at adaptcs.com  Fri Mar 16 12:45:06 2007
From: bldrake at adaptcs.com (Barry Drake)
Date: Fri, 16 Mar 2007 09:45:06 -0700 (PDT)
Subject: [SciPy-user] Structured matrices
In-Reply-To: <45F81DA2.2000201@cse.ucsc.edu>
Message-ID: <363507.29865.qm@web408.biz.mail.mud.yahoo.com>

Anand,
Since the matrix is not symmetric, how can it be a covariance matrix?

My area is signal processing and I haven't seen a non symmetric covariance matrix in my work.

Regards,
Barry L Drake

Anand Patil  wrote: Nils, Barry,

That matrix is also the covariance of Brownian motion evaluated at times [9, 7, 5, 3, 1]. The inverse matrix bears a strong resemblance to the numerical second derivative operator. This makes sense because the second derivative in x of min(x, \xi) is a delta function concentrated at \xi, but there's probably a more important connection that I don't know about.

Maybe this helps explain the strange results? I found some related refs, but haven't looked at them: http://citeseer.ist.psu.edu/347057.html, http://adsabs.harvard.edu/abs/2003EAEJA.....2001X .

Cheers,
Anand


--------------------------

Nils,
Does this matrix come from a particular application?  I'm working on algorithms for the non-negative matrix factorization (NMF).  With this matrix as input I'm getting some very strange results.  So I'm curious about potential applications.

Cheers!
Barry L. Drake
GA Tech

Nils Wagner wrote: Hi all,

I was wondering if the matrix family (see below) has a special name ?
And/or is there a way to construct this matrix via special matrices
(like Hankel, Toeplitz, etc.) ?


[[ 3.  1.]
 [ 1.  1.]]

[[ 5.  3.  1.]
 [ 3.  3.  1.]
 [ 1.  1.  1.]]

[[ 7.  5.  3.  1.]
 [ 5.  5.  3.  1.]
 [ 3.  3.  3.  1.]
 [ 1.  1.  1.  1.]]

[[ 9.  7.  5.  3.  1.]
 [ 7.  7.  5.  3.  1.]
 [ 5.  5.  5.  3.  1.]
 [ 3.  3.  3.  3.  1.]
 [ 1.  1.  1.  1.  1.]]

Nils

_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070316/83a4d7b3/attachment.html>

From zpincus at stanford.edu  Fri Mar 16 12:54:32 2007
From: zpincus at stanford.edu (Zachary Pincus)
Date: Fri, 16 Mar 2007 11:54:32 -0500
Subject: [SciPy-user] Gaussian pyramid using ndimage
In-Reply-To: <9e8c52a20703160942j3d1b4ca5ubbfab6dabe39f7c0@mail.gmail.com>
References: <9e8c52a20703160942j3d1b4ca5ubbfab6dabe39f7c0@mail.gmail.com>
Message-ID: <BCD52B8D-1FE6-4616-85BF-4BE4103F02B9@stanford.edu>

I *think* that all of the filters in ndimage proceed by pre-filtering  
and then doing whatever task is required, so even if there was a  
downsampling filter (maybe 'zoom' would work), it wouldn't be memory- 
efficient in the way you were hoping.

Also, the spline filters used by ndimage for pre-filtering seem  
somewhat broken:
http://projects.scipy.org/scipy/scipy/ticket/213

Zach



On Mar 16, 2007, at 11:42 AM, Alexander Borghgraef wrote:

> Hi all,
>
>  I've been looking into the ndimage module functionality, and I was  
> wondering whether there were any downsampling filters
> implemented within it, like used in gaussian pyramid  
> implementations. Obviously, convolution followed by downsampling would
> do the trick, but that requires an intermediate storing of a full- 
> sized image, which isn't all that memory-efficient. Any suggestions?
>
> -- 
> Alex Borghgraef
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user



From fperez.net at gmail.com  Fri Mar 16 14:02:28 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Fri, 16 Mar 2007 12:02:28 -0600
Subject: [SciPy-user] Structured matrices
In-Reply-To: <363507.29865.qm@web408.biz.mail.mud.yahoo.com>
References: <45F81DA2.2000201@cse.ucsc.edu>
	<363507.29865.qm@web408.biz.mail.mud.yahoo.com>
Message-ID: <db6b5ecc0703161102x8c9a148sdc3fc5bf8963c7c1@mail.gmail.com>

On 3/16/07, Barry Drake <bldrake at adaptcs.com> wrote:
> Anand,
> Since the matrix is not symmetric, how can it be a covariance matrix?

Why do you say it's not symmetric?  Using Anne's beautiful one-liner:

In [223]: ss = lambda s: minimum.outer(r_[s:-1:-2],r_[s:-1:-2])

Checking for s=9:

In [224]: ss(9)
Out[224]:
array([[9, 7, 5, 3, 1],
       [7, 7, 5, 3, 1],
       [5, 5, 5, 3, 1],
       [3, 3, 3, 3, 1],
       [1, 1, 1, 1, 1]])


Structurally it seems pretty obvious they must be symmetric always,
and it's also straightforward to do an explicit check over a number of
sizes:

In [225]: sszero = lambda s: abs(ss(s)-ss(s).T).max()

In [226]: [sszero(n) for n in range(3,45,2)]
Out[226]: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]


Am I missing something?

Cheers,

f


From nvf at uwm.edu  Fri Mar 16 16:16:33 2007
From: nvf at uwm.edu (Nick Fotopoulos)
Date: Fri, 16 Mar 2007 15:16:33 -0500
Subject: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9: _fftpack.so
	problems?
Message-ID: <f687be130703161316i9933da8t17cf6b0c7a992a80@mail.gmail.com>

> Message: 5
> Date: Thu, 15 Mar 2007 15:07:26 -0400
> From: Lawrence David <ldavid at MIT.EDU>
> Subject: Re: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9:
>         _fftpack.so     problems?
> To: SciPy Users List <scipy-user at scipy.org>
> Message-ID: <77114BB9-AB85-4BB2-85F5-DE324676EC3C at mit.edu>
> Content-Type: text/plain; charset="us-ascii"
>
> Hot darn Batman - you nailed it!!
>
> Much thanks,
> - lawrence
>
> http://www.stinkpot.org
>
>
> On Mar 15, 2007, at 12:58 PM, Robert Kern wrote:
>
> > Lawrence David wrote:
> >> Hi there,
> >>
> >> I'm pulling the little hair I have out trying to install SciPy on
> >> my G4
> >> PPC running OS 10.4.9.
> >>
> >> I've followed the instructions on the SciPy.org
> >> website: http://www.scipy.org/Installing_SciPy/Mac_OS_X.  I've got
> >> gcc
> >> 4.0.1 and gfortran (gnu95 -> 4.3).
> >>
> >> Running build: "python setup.py build_src build_clib --
> >> fcompiler=gnu95
> >> build_ext --fcompiler=gnu95 build", yields a long error message that
> >> terminates with a litany of "Undefined symbols" and this last
> >> message:
> >>
> >> collect2: ld returned 1 exit status
> >> error: Command "/usr/local/bin/gfortran -L/sw/lib
> >> build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/Lib/
> >> fftpack/_fftpackmodule.o
> >> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfft.o
> >> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/drfft.o
> >> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zrfft.o
> >> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfftnd.o
> >> build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/
> >> fortranobject.o
> >> -L/usr/local/lib -L/usr/local/lib/gcc/powerpc-apple-darwin8.8.0/4.3.0
> >> -Lbuild/temp.macosx-10.3-fat-2.5 -ldfftpack -lfftw3 -lgfortran -o
> >> build/lib.macosx-10.3-fat-2.5/scipy/fftpack/_fftpack.so" failed with
> >> exit status 1
> >>
> >> I've noticed several other folks have had similar error messages
> >> on this
> >> board, but none of the posted solutions worked for me.  If anyone
> >> could
> >> offer any suggestions on what to try, I'd love to hear them!
> >
> > It looks like you have LDFLAGS defined (thus, the -L/sw/lib). That
> > *replaces*
> > all of the Python-specific linker flags that distutils provides.

The symptoms sound exactly like mine, except that in my case LDFLAGS
is definitely not defined, so the solution does not apply to me.  The
big thread http://groups.google.com/group/comp.lang.python/browse_thread/thread/1fadaffda3a768a7/9a76e6cb3482d4a7?lnk=st&q=distutils+build+flags+problems&rnum=14&hl=en#9a76e6cb3482d4a7
says more or less the same thing.

nvf at dirac:~$ echo $LD<tab><tab>
$LDG_DIRECTORY          $LDG_LOCATION           $LD_LIBRARY_PATH
$LDG_INSTALL_LOG        $LDG_SOFTWARE_LOCATION

Is there anything else that could cause missing flags?  Scipy builds for the
Apple-supplied Python 2.3, but does not build for MacPython 2.4 or
2.5.

I eventually got it to build using "python setup.py build_src
build_clib --fcompiler=gnu95 build_ext -lpython --fcompiler=gnu95
build", but unsurprisingly, it doesn't work.  I think it's picking up
the system python's libpython:

nvf at dirac:~$ python -c "import scipy; scipy.test(10,10)"
Fatal Python error: Interpreter not initialized (version mismatch?)
Abort trap
nvf at dirac:~$ locate libpython
/Developer/SDKs/MacOSX10.4u.sdk/usr/lib/libpython.dylib
/Developer/SDKs/MacOSX10.4u.sdk/usr/lib/libpython2.3.dylib
/Developer/SDKs/MacOSX10.4u.sdk/usr/lib/libpython2.dylib
/Library/Frameworks/Python.framework/Versions/2.4/lib/python2.4/config/libpython2.4.a
/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/config/libpython2.5.a
/usr/lib/libpython.dylib
/usr/lib/libpython2.3.dylib
/usr/lib/libpython2.dylib

Those libpython.dylibs just simlink to the libpython2.3.dylib.  I
would have thought that the MacPython installer would provide dynamic
libraries in addition to the static libraries.

I would appreciate any help on this, as my previous pleas have gone
unanswered.  If there's any other information I can provide, just let
me know.

Many thanks,
Nick


From robert.kern at gmail.com  Fri Mar 16 16:31:35 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 16 Mar 2007 15:31:35 -0500
Subject: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9: _fftpack.so
 problems?
In-Reply-To: <f687be130703161316i9933da8t17cf6b0c7a992a80@mail.gmail.com>
References: <f687be130703161316i9933da8t17cf6b0c7a992a80@mail.gmail.com>
Message-ID: <45FAFEA7.2010403@gmail.com>

Nick Fotopoulos wrote:
>> Message: 5
>> Date: Thu, 15 Mar 2007 15:07:26 -0400
>> From: Lawrence David <ldavid at MIT.EDU>
>> Subject: Re: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9:
>>         _fftpack.so     problems?
>> To: SciPy Users List <scipy-user at scipy.org>
>> Message-ID: <77114BB9-AB85-4BB2-85F5-DE324676EC3C at mit.edu>
>> Content-Type: text/plain; charset="us-ascii"
>>
>> Hot darn Batman - you nailed it!!
>>
>> Much thanks,
>> - lawrence
>>
>> http://www.stinkpot.org
>>
>>
>> On Mar 15, 2007, at 12:58 PM, Robert Kern wrote:
>>
>>> Lawrence David wrote:
>>>> Hi there,
>>>>
>>>> I'm pulling the little hair I have out trying to install SciPy on
>>>> my G4
>>>> PPC running OS 10.4.9.
>>>>
>>>> I've followed the instructions on the SciPy.org
>>>> website: http://www.scipy.org/Installing_SciPy/Mac_OS_X.  I've got
>>>> gcc
>>>> 4.0.1 and gfortran (gnu95 -> 4.3).
>>>>
>>>> Running build: "python setup.py build_src build_clib --
>>>> fcompiler=gnu95
>>>> build_ext --fcompiler=gnu95 build", yields a long error message that
>>>> terminates with a litany of "Undefined symbols" and this last
>>>> message:
>>>>
>>>> collect2: ld returned 1 exit status
>>>> error: Command "/usr/local/bin/gfortran -L/sw/lib
>>>> build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/Lib/
>>>> fftpack/_fftpackmodule.o
>>>> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfft.o
>>>> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/drfft.o
>>>> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zrfft.o
>>>> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfftnd.o
>>>> build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/
>>>> fortranobject.o
>>>> -L/usr/local/lib -L/usr/local/lib/gcc/powerpc-apple-darwin8.8.0/4.3.0
>>>> -Lbuild/temp.macosx-10.3-fat-2.5 -ldfftpack -lfftw3 -lgfortran -o
>>>> build/lib.macosx-10.3-fat-2.5/scipy/fftpack/_fftpack.so" failed with
>>>> exit status 1
>>>>
>>>> I've noticed several other folks have had similar error messages
>>>> on this
>>>> board, but none of the posted solutions worked for me.  If anyone
>>>> could
>>>> offer any suggestions on what to try, I'd love to hear them!
>>> It looks like you have LDFLAGS defined (thus, the -L/sw/lib). That
>>> *replaces*
>>> all of the Python-specific linker flags that distutils provides.
> 
> The symptoms sound exactly like mine, except that in my case LDFLAGS
> is definitely not defined, so the solution does not apply to me.  The
> big thread http://groups.google.com/group/comp.lang.python/browse_thread/thread/1fadaffda3a768a7/9a76e6cb3482d4a7?lnk=st&q=distutils+build+flags+problems&rnum=14&hl=en#9a76e6cb3482d4a7
> says more or less the same thing.

Except in that case, LDFLAGS was the problem.

> nvf at dirac:~$ echo $LD<tab><tab>
> $LDG_DIRECTORY          $LDG_LOCATION           $LD_LIBRARY_PATH
> $LDG_INSTALL_LOG        $LDG_SOFTWARE_LOCATION
> 
> Is there anything else that could cause missing flags?  Scipy builds for the
> Apple-supplied Python 2.3, but does not build for MacPython 2.4 or
> 2.5.
> 
> I eventually got it to build using "python setup.py build_src
> build_clib --fcompiler=gnu95 build_ext -lpython --fcompiler=gnu95
> build", but unsurprisingly, it doesn't work.  I think it's picking up
> the system python's libpython:

Yes, because you added -lpython which is not how you need to link against the
Python libraries for framework builds of Python.

> nvf at dirac:~$ python -c "import scipy; scipy.test(10,10)"
> Fatal Python error: Interpreter not initialized (version mismatch?)
> Abort trap
> nvf at dirac:~$ locate libpython
> /Developer/SDKs/MacOSX10.4u.sdk/usr/lib/libpython.dylib
> /Developer/SDKs/MacOSX10.4u.sdk/usr/lib/libpython2.3.dylib
> /Developer/SDKs/MacOSX10.4u.sdk/usr/lib/libpython2.dylib
> /Library/Frameworks/Python.framework/Versions/2.4/lib/python2.4/config/libpython2.4.a
> /Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/config/libpython2.5.a
> /usr/lib/libpython.dylib
> /usr/lib/libpython2.3.dylib
> /usr/lib/libpython2.dylib
> 
> Those libpython.dylibs just simlink to the libpython2.3.dylib.  I
> would have thought that the MacPython installer would provide dynamic
> libraries in addition to the static libraries.

No, everything should be contained in /Library/Frameworks/Python.framework . In
such a case, the "shared library" is actually the file
/Library/Frameworks/Python.framework/Python . It gets linked when "-framework
Python" is part of the link command, as distutils does provided it is not
interfered with.

> I would appreciate any help on this, as my previous pleas have gone
> unanswered.  If there's any other information I can provide, just let
> me know.

Can you supply the output from an unsuccessful build again? Don't do -lpython
this time.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From bldrake at adaptcs.com  Fri Mar 16 16:40:28 2007
From: bldrake at adaptcs.com (Barry Drake)
Date: Fri, 16 Mar 2007 13:40:28 -0700 (PDT)
Subject: [SciPy-user] Structured matrices
In-Reply-To: <db6b5ecc0703161102x8c9a148sdc3fc5bf8963c7c1@mail.gmail.com>
Message-ID: <173515.38726.qm@web409.biz.mail.mud.yahoo.com>

It is symmetric.  I was relying on someone else looking at this
and didn't double check it first.  I'll be more careful next time.

Thanks,
Barry

Fernando Perez <fperez.net at gmail.com> wrote: On 3/16/07, Barry Drake  wrote:
> Anand,
> Since the matrix is not symmetric, how can it be a covariance matrix?

Why do you say it's not symmetric?  Using Anne's beautiful one-liner:

In [223]: ss = lambda s: minimum.outer(r_[s:-1:-2],r_[s:-1:-2])

Checking for s=9:

In [224]: ss(9)
Out[224]:
array([[9, 7, 5, 3, 1],
       [7, 7, 5, 3, 1],
       [5, 5, 5, 3, 1],
       [3, 3, 3, 3, 1],
       [1, 1, 1, 1, 1]])


Structurally it seems pretty obvious they must be symmetric always,
and it's also straightforward to do an explicit check over a number of
sizes:

In [225]: sszero = lambda s: abs(ss(s)-ss(s).T).max()

In [226]: [sszero(n) for n in range(3,45,2)]
Out[226]: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]


Am I missing something?

Cheers,

f
_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070316/5f6d8385/attachment.html>

From hakan.jakobsson at gmail.com  Fri Mar 16 22:05:51 2007
From: hakan.jakobsson at gmail.com (=?ISO-8859-1?Q?H=E5kan_Jakobsson?=)
Date: Sat, 17 Mar 2007 03:05:51 +0100
Subject: [SciPy-user] FEM
Message-ID: <45FB4CFF.5050901@gmail.com>

Hi list,
Python/scipy newbie here. I'm currently writing my master's thesis in 
computational mathematics - discontinuous galerkin methods to be 
precise. I'v been doing all the numerics for my thesis in Matlab but 
stumbled upon Python and scipy by chance. As a little experiment I 
translated one of my solvers into Python using numpy, scipy and the 
pysparse packages, and I was very impressed with the effort/performance 
ratio. Now, to the point..

If you're doing finite element analysis, what are you're experiences 
using Python? Do you use Python as your main development tool? In 
conjunction with C, Fortran? Is there any drawbacks with using Python 
for this type of work? Anything else?

I hope I'm not asking to much here, but it would be really interesting 
to know a bit about what the situation's like. I think Python and scipy 
is really great, and now I'm trying to convince everyone else...

Please, share with me your experiences.

/H?kan J


From wbaxter at gmail.com  Fri Mar 16 22:25:17 2007
From: wbaxter at gmail.com (Bill Baxter)
Date: Sat, 17 Mar 2007 11:25:17 +0900
Subject: [SciPy-user] FEM
In-Reply-To: <45FB4CFF.5050901@gmail.com>
References: <45FB4CFF.5050901@gmail.com>
Message-ID: <e86a5fd00703161925l251517f3y1ca82f2e876782cc@mail.gmail.com>

Howdy, I don't do FEA, but the things I do are not so different.
You might find the matlab comparison web page useful:
   http://www.scipy.org/NumPy_for_Matlab_Users
Since I wrote a lot that page, I won't repeat the pros and cons
arguments here in detail, but rather refer you to that page.

Here's a quick summary of what I see to be the highlights:
* Python/Numpy/Scipy are much better when it comes to the language
itself and software development aspects, integrating with native code
in C/C++/Fortran, and applying spot-optimizations.

* Matlab is better (for now) when it comes to graphical debugging.
* Matlab is better (for now) when it comes to finding free code to
solve standard mathematical problem X.
* Matlab is better when it comes to profiling.

But keep on evangelizing.  As far as I see it, the advantages of
Python are intrinsic (it's based on a good language with broad
community), whereas Matlab's are temporary and derive primarily from
having a large head start.

Finally, especially for academia I think the licensing terms are a
very important consideration.

--bb

On 3/17/07, H?kan Jakobsson <hakan.jakobsson at gmail.com> wrote:
> Hi list,
> Python/scipy newbie here. I'm currently writing my master's thesis in
> computational mathematics - discontinuous galerkin methods to be
> precise. I'v been doing all the numerics for my thesis in Matlab but
> stumbled upon Python and scipy by chance. As a little experiment I
> translated one of my solvers into Python using numpy, scipy and the
> pysparse packages, and I was very impressed with the effort/performance
> ratio. Now, to the point..
>
> If you're doing finite element analysis, what are you're experiences
> using Python? Do you use Python as your main development tool? In
> conjunction with C, Fortran? Is there any drawbacks with using Python
> for this type of work? Anything else?
>
> I hope I'm not asking to much here, but it would be really interesting
> to know a bit about what the situation's like. I think Python and scipy
> is really great, and now I'm trying to convince everyone else...
>
> Please, share with me your experiences.
>
> /H?kan J
>


From josh8912 at yahoo.com  Sat Mar 17 01:10:30 2007
From: josh8912 at yahoo.com (JJ)
Date: Fri, 16 Mar 2007 22:10:30 -0700 (PDT)
Subject: [SciPy-user] use of Jacobian function in leastsq
Message-ID: <508387.65137.qm@web54007.mail.yahoo.com>

Hello:
This is a question based on previous discussions by
Brendan Simons and Travis (RE: optimize.leastsq ->
fjac and the covariance redux).  I am wondering why
use of the Jacobian function in the leastsq function
does not seem to work.  There is probably a very
simple answer.  I am trying to use a Jacobian function
for a different problem and am using this as a test. 
The code is as follows, and you can see that C is not
equal to C2:

from scipy import *
from scipy.optimize import leastsq

x = array([0., 1., 2., 3., 4., 5.]).astype('d')
coeffs =  [4., 3., 5., 2.]
yErrs = array([0.1, 0.2, -0.1, 0.05, 0,
-.02]).astype('d')
m = len(x)
n = len(coeffs)

def evalY(p, x):
    """a simple cubic"""
    return x**3 * p[3] + x**2 * p[2] + x * p[1] + p[0]

def J(p, y, x):
    """the jacobian of a cubic (not actually
    a function of y, p since resid is linear in p)
    """
    print '\n in J'
    result = zeros([m,n], 'd')
    result[:, 0] = ones(m, 'd')
    result[:, 1] = x
    result[:, 2] = x**2
    result[:, 3] = x**3
    print result
    return result

def resid(p, y, x):
    return y - evalY(p, x)

y_true = evalY(coeffs, x)
y_meas = y_true + yErrs

p0 = [3., 2., 4., 1.]

pMin, C, infodict, ier, mesg = leastsq(resid, p0,
        args=(y_meas, x),  full_output=True)

pMin2, C2, infodict2, ier2, mesg2 = leastsq(resid, p0,
Dfun=J,
        args=(y_meas, x),  full_output=True)

#check against know algorithms for computing C
JAtPMin = J(pMin, None, x)
Q_true, R_true = linalg.qr(JAtPMin)
C_true = linalg.inv(dot(transpose(JAtPMin), JAtPMin))
C_true2 = linalg.inv(dot(transpose(R_true), R_true))

print 'pMin'
print pMin
print 'pMin2'
print pMin2
print '\nC'
print C
print '\nC2'
print C2
print '\nC_true'
print C_true
print '\nC_true2'
print C_true2

Thanks, JJ


 
____________________________________________________________________________________
Looking for earth-friendly autos? 
Browse Top Cars by "Green Rating" at Yahoo! Autos' Green Center.
http://autos.yahoo.com/green_center/


From ryanlists at gmail.com  Sat Mar 17 11:08:41 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Sat, 17 Mar 2007 10:08:41 -0500
Subject: [SciPy-user] FEM
In-Reply-To: <e86a5fd00703161925l251517f3y1ca82f2e876782cc@mail.gmail.com>
References: <45FB4CFF.5050901@gmail.com>
	<e86a5fd00703161925l251517f3y1ca82f2e876782cc@mail.gmail.com>
Message-ID: <c5b438120703170808w5314fed6t305468239e9c5874@mail.gmail.com>

I think the question is a good one and very appropriate when you
consider the potential risk you are taking and the amount of time it
could take to really learn all you need to know to make a truly
informed decision.  I switched from Matlab to
Scipy/Numpy/Matplotlib/Mayavi/IPython half way through my Ph.D. work
and it was a great decision.  But, I didn't do any FEA in Python.
F2py makes using native FORTRAN code very easy.  Google for Python
FEA/FEM.  I know I have run across some packages, but I have never
looked at it in detail.

I have never been able to quantify it (others have had some success),
but in my experience, Python code is much easier to write and debug
than Matlab code.  As a simple example, if I need to operate on every
item in a vector (and can't come up with a good vectorized solution),
in Matlab I need to write

for i in length(vector),
    item = vector(i);
    {do something with item}
end

where in Python it is just

for item in vector:
    {do something with item}

That is a really simple example, but it shows to areas where Matlab
enables me to introduce bugs:
1. If I have nested for loops, I might screw up the indices (i,j,k,...)
2. It lets me overwrite the imaginary number i

I lost countless hours of productivity chasing down those inds of
bugs.  (You only make the i as index mistake once or twice).

And don't get me started about making sure you remember the stinking
semicolons....

FWIW,

Ryan

On 3/16/07, Bill Baxter <wbaxter at gmail.com> wrote:
> Howdy, I don't do FEA, but the things I do are not so different.
> You might find the matlab comparison web page useful:
>    http://www.scipy.org/NumPy_for_Matlab_Users
> Since I wrote a lot that page, I won't repeat the pros and cons
> arguments here in detail, but rather refer you to that page.
>
> Here's a quick summary of what I see to be the highlights:
> * Python/Numpy/Scipy are much better when it comes to the language
> itself and software development aspects, integrating with native code
> in C/C++/Fortran, and applying spot-optimizations.
>
> * Matlab is better (for now) when it comes to graphical debugging.
> * Matlab is better (for now) when it comes to finding free code to
> solve standard mathematical problem X.
> * Matlab is better when it comes to profiling.
>
> But keep on evangelizing.  As far as I see it, the advantages of
> Python are intrinsic (it's based on a good language with broad
> community), whereas Matlab's are temporary and derive primarily from
> having a large head start.
>
> Finally, especially for academia I think the licensing terms are a
> very important consideration.
>
> --bb
>
> On 3/17/07, H?kan Jakobsson <hakan.jakobsson at gmail.com> wrote:
> > Hi list,
> > Python/scipy newbie here. I'm currently writing my master's thesis in
> > computational mathematics - discontinuous galerkin methods to be
> > precise. I'v been doing all the numerics for my thesis in Matlab but
> > stumbled upon Python and scipy by chance. As a little experiment I
> > translated one of my solvers into Python using numpy, scipy and the
> > pysparse packages, and I was very impressed with the effort/performance
> > ratio. Now, to the point..
> >
> > If you're doing finite element analysis, what are you're experiences
> > using Python? Do you use Python as your main development tool? In
> > conjunction with C, Fortran? Is there any drawbacks with using Python
> > for this type of work? Anything else?
> >
> > I hope I'm not asking to much here, but it would be really interesting
> > to know a bit about what the situation's like. I think Python and scipy
> > is really great, and now I'm trying to convince everyone else...
> >
> > Please, share with me your experiences.
> >
> > /H?kan J
> >
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From nvf at uwm.edu  Sat Mar 17 11:35:02 2007
From: nvf at uwm.edu (Nick Fotopoulos)
Date: Sat, 17 Mar 2007 10:35:02 -0500
Subject: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9: _fftpack.so
	problems?
Message-ID: <f687be130703170835v65a7876etfc5012fd34ce10c8@mail.gmail.com>

On 3/17/07, scipy-user-request at scipy.org <scipy-user-request at scipy.org> wrote:
> Message: 3
> Date: Fri, 16 Mar 2007 15:31:35 -0500
> From: Robert Kern <robert.kern at gmail.com>
> Subject: Re: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9:
>         _fftpack.so problems?
> To: SciPy Users List <scipy-user at scipy.org>
> Message-ID: <45FAFEA7.2010403 at gmail.com>
> Content-Type: text/plain; charset=UTF-8

<snip>

> Nick Fotopoulos wrote:
> > I eventually got it to build using "python setup.py build_src
> > build_clib --fcompiler=gnu95 build_ext -lpython --fcompiler=gnu95
> > build", but unsurprisingly, it doesn't work.  I think it's picking up
> > the system python's libpython:
>
> Yes, because you added -lpython which is not how you need to link against the
> Python libraries for framework builds of Python.

I was guessing.  Thanks for the correction.

> > nvf at dirac:~$ python -c "import scipy; scipy.test(10,10)"
> > Fatal Python error: Interpreter not initialized (version mismatch?)
> > Abort trap
> > nvf at dirac:~$ locate libpython
> > /Developer/SDKs/MacOSX10.4u.sdk/usr/lib/libpython.dylib
> > /Developer/SDKs/MacOSX10.4u.sdk/usr/lib/libpython2.3.dylib
> > /Developer/SDKs/MacOSX10.4u.sdk/usr/lib/libpython2.dylib
> > /Library/Frameworks/Python.framework/Versions/2.4/lib/python2.4/config/libpython2.4.a
> > /Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/config/libpython2.5.a
> > /usr/lib/libpython.dylib
> > /usr/lib/libpython2.3.dylib
> > /usr/lib/libpython2.dylib
> >
> > Those libpython.dylibs just simlink to the libpython2.3.dylib.  I
> > would have thought that the MacPython installer would provide dynamic
> > libraries in addition to the static libraries.
>
> No, everything should be contained in /Library/Frameworks/Python.framework . In
> such a case, the "shared library" is actually the file
> /Library/Frameworks/Python.framework/Python . It gets linked when "-framework
> Python" is part of the link command, as distutils does provided it is not
> interfered with.

Huh, I wouldn't have guessed this on my own.

> Can you supply the output from an unsuccessful build again? Don't do -lpython
> this time.

<all the missing _Py symbols>
collect2: ld returned 1 exit status
error: Command "/usr/local/bin/gfortran -Wall -bundle
build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/Lib/fftpack/_fftpackmodule.o
build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfft.o
build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/drfft.o
build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zrfft.o
build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfftnd.o
build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/fortranobject.o
-L/opt/lscsoft/non-lsc/lib
-L/usr/local/lib/gcc/i386-apple-darwin8.8.1/4.3.0
-Lbuild/temp.macosx-10.3-fat-2.5 -ldfftpack -lfftw3 -lgfortran -o
build/lib.macosx-10.3-fat-2.5/scipy/fftpack/_fftpack.so" failed with
exit status 1

Many thanks for the information and for taking a shot at this.

Take care,
Nick


From hasslerjc at comcast.net  Sat Mar 17 11:43:05 2007
From: hasslerjc at comcast.net (John Hassler)
Date: Sat, 17 Mar 2007 11:43:05 -0400
Subject: [SciPy-user] FEM
In-Reply-To: <c5b438120703170808w5314fed6t305468239e9c5874@mail.gmail.com>
References: <45FB4CFF.5050901@gmail.com>	<e86a5fd00703161925l251517f3y1ca82f2e876782cc@mail.gmail.com>
	<c5b438120703170808w5314fed6t305468239e9c5874@mail.gmail.com>
Message-ID: <45FC0C89.8010104@comcast.net>

An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070317/3b8264d9/attachment.html>

From dschult at colgate.edu  Sat Mar 17 13:00:44 2007
From: dschult at colgate.edu (Dan Schult)
Date: Sat, 17 Mar 2007 13:00:44 -0400
Subject: [SciPy-user] ld unknown flag macosx_version_min
Message-ID: <FE1EFC88-5265-4BF8-8BF9-EAB1362CAEC5@colgate.edu>

I've got a Mac OSx 10.4.8 machine and am compiling
scipy according to the instructions on the webpage.  I've
got gcc 4.0.0 gfortran 4.3.0 fftw3.0 and svn versions of numpy
and scipy.   My python is version 2.5.
Building numpy goes smoothly, but when I try scipy I have an ld error.

It seems to be using a linker option I am not familiar with and I can't
find it in the package anywhere.  Could it be coming from the
python 2.5 binary I downloaded?  Anyway, how do I turn it off or get
an ld that accepts this option?
The option is  macosx_version_min
Thanks for your help!
Dan Schult

python setup.py build_src build_clib --fcompiler=gnu95 build_ext -- 
fcompiler=gnu95 build
...
Lots of stuff as usual, ending with
...
running build_ext
customize UnixCCompiler
customize UnixCCompiler using build_ext
customize Gnu95FCompiler
customize Gnu95FCompiler using build_ext
building 'scipy.fftpack._fftpack' extension
compiling C sources
C compiler: gcc -fno-strict-aliasing -Wno-long-double -no-cpp-precomp  
-mno-fused-madd -fno-common -dynamic -DNDEBUG -g -O3 -Wall

compile options: '-DSCIPY_FFTW3_H -I/usr/local/include -Ibuild/ 
src.macosx-10.3-ppc-2.5 -I/Library/Frameworks/Python.framework/ 
Versions/2.5/lib/python2.5/site-packages/numpy/core/include -I/ 
Library/Frameworks/Python.framework/Versions/2.5/include/python2.5 -c'
/usr/local/bin/gfortran -Wall -undefined dynamic_lookup -bundle build/ 
temp.macosx-10.3-ppc-2.5/build/src.macosx-10.3-ppc-2.5/Lib/fftpack/ 
_fftpackmodule.o build/temp.macosx-10.3-ppc-2.5/Lib/fftpack/src/ 
zfft.o build/temp.macosx-10.3-ppc-2.5/Lib/fftpack/src/drfft.o build/ 
temp.macosx-10.3-ppc-2.5/Lib/fftpack/src/zrfft.o build/ 
temp.macosx-10.3-ppc-2.5/Lib/fftpack/src/zfftnd.o build/ 
temp.macosx-10.3-ppc-2.5/build/src.macosx-10.3-ppc-2.5/ 
fortranobject.o -L/usr/local/lib -L/usr/local/lib/gcc/powerpc-apple- 
darwin8.8.0/4.3.0 -Lbuild/temp.macosx-10.3-ppc-2.5 -ldfftpack -lfftw3  
-lgfortran -o build/lib.macosx-10.3-ppc-2.5/scipy/fftpack/_fftpack.so
/usr/bin/ld: unknown flag: -macosx_version_min
collect2: ld returned 1 exit status
/usr/bin/ld: unknown flag: -macosx_version_min
collect2: ld returned 1 exit status
error: Command "/usr/local/bin/gfortran -Wall -undefined  
dynamic_lookup -bundle build/temp.macosx-10.3-ppc-2.5/build/ 
src.macosx-10.3-ppc-2.5/Lib/fftpack/_fftpackmodule.o build/ 
temp.macosx-10.3-ppc-2.5/Lib/fftpack/src/zfft.o build/ 
temp.macosx-10.3-ppc-2.5/Lib/fftpack/src/drfft.o build/ 
temp.macosx-10.3-ppc-2.5/Lib/fftpack/src/zrfft.o build/ 
temp.macosx-10.3-ppc-2.5/Lib/fftpack/src/zfftnd.o build/ 
temp.macosx-10.3-ppc-2.5/build/src.macosx-10.3-ppc-2.5/ 
fortranobject.o -L/usr/local/lib -L/usr/local/lib/gcc/powerpc-apple- 
darwin8.8.0/4.3.0 -Lbuild/temp.macosx-10.3-ppc-2.5 -ldfftpack -lfftw3  
-lgfortran -o build/lib.macosx-10.3-ppc-2.5/scipy/fftpack/ 
_fftpack.so" failed with exit status 1
  
  


From robert.kern at gmail.com  Sat Mar 17 15:41:33 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Sat, 17 Mar 2007 14:41:33 -0500
Subject: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9: _fftpack.so
 problems?
In-Reply-To: <f687be130703170835v65a7876etfc5012fd34ce10c8@mail.gmail.com>
References: <f687be130703170835v65a7876etfc5012fd34ce10c8@mail.gmail.com>
Message-ID: <45FC446D.7000409@gmail.com>

Nick Fotopoulos wrote:

> <all the missing _Py symbols>
> collect2: ld returned 1 exit status
> error: Command "/usr/local/bin/gfortran -Wall -bundle
> build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/Lib/fftpack/_fftpackmodule.o
> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfft.o
> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/drfft.o
> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zrfft.o
> build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfftnd.o
> build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/fortranobject.o
> -L/opt/lscsoft/non-lsc/lib
> -L/usr/local/lib/gcc/i386-apple-darwin8.8.1/4.3.0
> -Lbuild/temp.macosx-10.3-fat-2.5 -ldfftpack -lfftw3 -lgfortran -o
> build/lib.macosx-10.3-fat-2.5/scipy/fftpack/_fftpack.so" failed with
> exit status 1
> 
> Many thanks for the information and for taking a shot at this.

Where is this coming from if not LDFLAGS?

 -L/opt/lscsoft/non-lsc/lib

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From robert.kern at gmail.com  Sat Mar 17 16:27:39 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Sat, 17 Mar 2007 15:27:39 -0500
Subject: [SciPy-user] ld unknown flag macosx_version_min
In-Reply-To: <FE1EFC88-5265-4BF8-8BF9-EAB1362CAEC5@colgate.edu>
References: <FE1EFC88-5265-4BF8-8BF9-EAB1362CAEC5@colgate.edu>
Message-ID: <45FC4F3B.6050302@gmail.com>

Dan Schult wrote:
> I've got a Mac OSx 10.4.8 machine and am compiling
> scipy according to the instructions on the webpage.  I've
> got gcc 4.0.0 gfortran 4.3.0 fftw3.0 and svn versions of numpy
> and scipy.   My python is version 2.5.
> Building numpy goes smoothly, but when I try scipy I have an ld error.
> 
> It seems to be using a linker option I am not familiar with and I can't
> find it in the package anywhere.  Could it be coming from the
> python 2.5 binary I downloaded?  Anyway, how do I turn it off or get
> an ld that accepts this option?
> The option is  macosx_version_min

I have gcc 4.0.1 and gfortran 4.3.0 installed on my system, and I do not see
this problem. Can you try upgrading to the latest version of Xcode (which should
have gcc 4.0.1)? It's not coming from Python but, I suspect, gfortran.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From lxander.m at gmail.com  Sun Mar 18 08:18:19 2007
From: lxander.m at gmail.com (Alexander Michael)
Date: Sun, 18 Mar 2007 08:18:19 -0400
Subject: [SciPy-user] FEM
In-Reply-To: <45FB4CFF.5050901@gmail.com>
References: <45FB4CFF.5050901@gmail.com>
Message-ID: <525f23e80703180518q75557644ta39da20e32828100@mail.gmail.com>

On 3/16/07, H?kan Jakobsson <hakan.jakobsson at gmail.com> wrote:
> Hi list,
> Python/scipy newbie here. I'm currently writing my master's thesis in
> computational mathematics - discontinuous galerkin methods to be
> precise. I'v been doing all the numerics for my thesis in Matlab but
> stumbled upon Python and scipy by chance. As a little experiment I
> translated one of my solvers into Python using numpy, scipy and the
> pysparse packages, and I was very impressed with the effort/performance
> ratio. Now, to the point..
>
> If you're doing finite element analysis, what are you're experiences
> using Python? Do you use Python as your main development tool? In
> conjunction with C, Fortran? Is there any drawbacks with using Python
> for this type of work? Anything else?
>
> I hope I'm not asking to much here, but it would be really interesting
> to know a bit about what the situation's like. I think Python and scipy
> is really great, and now I'm trying to convince everyone else...
>
> Please, share with me your experiences.
>
> /H?kan J

I haven't built any FEM work in Python, but it looks like others have
given it a go (I remember seeing [1] and [2], for instance). If you
are building matrix-less solvers (i.e. you're not explicitly forming
the large sparse matrix for the whole system), then I think Python and
NumPy could potentially take you a long way. If you're solver requires
forming the complete sparse system matrix, then you'll need something
more than NumPy for the sparse matrix representation. PySparse looks
like a reasonable (but I haven't used it) candidate
(<http://pysparse.sourceforge.net/>). It provides iterative solvers
and access to SuperLU (I personally like UMFPACK as well, but the
matrix solver is secondary if this is not the focus of your research).
I've used PETSc and C++ in my pre-Python days for building FEM, but it
looks like some people have shoe-horned it into a Python-centric
environment (<http://www.cimec.org.ar/python/>). They also provide
access to ParMETIS, which may be helpful depending on your goals.

Honestly, though, if you are being productive in your current
computational environment and Matlab meets your peformance and
architectural requirements, then my *guess* would be that a switch to
Python will cost you a month of progress. If Matlab won't scale to the
size of problems you wish to solve, then you will need to move closer
to the metal and Python over something like SuperLU or PETSc (if you
need parallelization) will save you from working in C/C++ (and provide
a better development environment). If your goal is to build flexible
"end-user software," then I don't think Matlab will cut it
architecturally and you will need to switch to something else anyway.


[1] <http://www.python.org/pycon/papers/pysparse.html>
[2] <http://www.fenics.org/syfi/>


From acorriga at gmu.edu  Sun Mar 18 09:38:02 2007
From: acorriga at gmu.edu (Andrew Corrigan)
Date: Sun, 18 Mar 2007 09:38:02 -0400
Subject: [SciPy-user] spsolve not working
Message-ID: <45FD40BA.1020107@gmu.edu>

I have the same problem mentioned below with scipy 0.5.2.  I try to call 
spsolve and get an error message.  This only happens when running in 
64-bit Ubuntu, but doesn't happen in Windows XP with Enthought's 
Python.  Can someone please help?

Thanks,
Andrew

Traceback (most recent call last):

  File 
"/home/acorrigan/workspace/meshless/src/physics/potential/test1.py", 
line 75, in ?
    main()
  File 
"/home/acorrigan/workspace/meshless/src/physics/potential/test1.py", 
line 69, in main
    s_f = GeneralizedRBFInterpolant(points, values, operators, 
RadialKernel(ArbitrarySupport(Gaussian(1.0),1.0)), FixedGrid)
  File 
"/home/acorrigan/workspace/meshless/src/interpolation/interpolant/generalized_rbf.py", 
line 55, in __init__
    self.resolve(values)
  File 
"/home/acorrigan/workspace/meshless/src/interpolation/interpolant/generalized_rbf.py", 
line 64, in resolve
    self.coefficients = linsolve.spsolve(self.A, 
values.swapaxes(0,1)).swapaxes(0,1)     #.tocsr()
  File "/usr/lib/python2.4/site-packages/scipy/linsolve/linsolve.py", 
line 77, in spsolve
    mat, csc = _toCS_superLU( A )
  File "/usr/lib/python2.4/site-packages/scipy/linsolve/linsolve.py", 
line 24, in _toCS_superLU
    mat = A.tocsc()
  File "/usr/lib/python2.4/site-packages/scipy/sparse/sparse.py", line 
2684, in tocsc
    return self.tocsr(nzmax).tocsc()
  File "/usr/lib/python2.4/site-packages/scipy/sparse/sparse.py", line 
2673, in tocsr
    data[k : k+len(row)] = self.data[i]
ValueError: setting an array element with a sequence.









Nils Wagner <nwagner <at> iam.uni-stuttgart.de> writes:

 >
 > Hi all,
 >
 > I would like to solve
 >
 > K_dyn x = f,
 >
 > where K_dyn is a sparse matrix.  UMFPACK is not installed and I am using
 > the latest svn version.
 >
 > >>> K_dyn
 > <71987x71987 sparse matrix of type '<type 'numpy.float64'>'
 >         with 3083884 stored elements (space for 3083884)
 >         in Compressed Sparse Column format>
 > >>> f
 > <71987x1 sparse matrix of type '<type 'numpy.float64'>'
 >         with 52 stored elements (space for 52)
 >         in Compressed Sparse Column format>
 > >>> x = spsolve(K_dyn, f)
 > Traceback (most recent call last):
 >   File "<stdin>", line 1, in ?
 >   File
 > "/usr/local/lib64/python2.4/site-packages/scipy/linsolve/linsolve.py",
 > line 75, in spsolve
 >     b = asarray(b, dtype=data.dtype)
 >   File "/usr/local/lib64/python2.4/site-packages/numpy/core/numeric.py",
 > line 132, in asarray
 >     return array(a, dtype, copy=False, order=order)
 > ValueError: setting an array element with a sequence.
 >
 > Is this a bug ?
 >
 > Nils
 >




From jelle.feringa at ezct.net  Sun Mar 18 10:49:18 2007
From: jelle.feringa at ezct.net (Jelle Feringa / EZCT Architecture & Design Research)
Date: Sun, 18 Mar 2007 15:49:18 +0100
Subject: [SciPy-user] vector field integration
Message-ID: <001401c7696c$9cf0a6f0$c000a8c0@JELLE>


Has anyone worked on vector field integration in a numpy/scipy context?
I'm evaluating my options to do so, but I have little experience in this, so
that makes it hard to decipher which libraries might be of good use, from
those that won't. Any pointers would be really appreciated!

Cheers,

-jelle



From nwagner at iam.uni-stuttgart.de  Sun Mar 18 12:02:29 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Sun, 18 Mar 2007 17:02:29 +0100
Subject: [SciPy-user] spsolve not working
In-Reply-To: <45FD40BA.1020107@gmu.edu>
References: <45FD40BA.1020107@gmu.edu>
Message-ID: <web-80528738@uni-stuttgart.de>

On Sun, 18 Mar 2007 09:38:02 -0400
  Andrew Corrigan <acorriga at gmu.edu> wrote:
> I have the same problem mentioned below with scipy 
>0.5.2.  I try to call 
> spsolve and get an error message.  This only happens 
>when running in 
> 64-bit Ubuntu, but doesn't happen in Windows XP with 
>Enthought's 
> Python.  Can someone please help?

Fixed in svn (IIRC).

Nils
  


From lechtlr at yahoo.com  Sun Mar 18 12:13:36 2007
From: lechtlr at yahoo.com (lechtlr)
Date: Sun, 18 Mar 2007 09:13:36 -0700 (PDT)
Subject: [SciPy-user] use of Jacobian function in leastsq
Message-ID: <28602.56536.qm@web57911.mail.re3.yahoo.com>

JJ:

If you define the derivatives of the Jacobian function in columns instead of rows, it'll work better. Try the attached script.

-Lex

from scipy import *
from scipy.optimize import leastsq

x = array([0., 1., 2., 3., 4., 5.]).astype('d')
coeffs =  [4., 3., 5., 2.]
yErrs = array([0.1, 0.2, -0.1, 0.05, 0,-.02]).astype('d')
m = len(x)
n = len(coeffs)

def evalY(p, x):
    """a simple cubic"""
    return x**3 * p[3] + x**2 * p[2] + x * p[1] + p[0]

def J(p, y, x):
    """the jacobian of a cubic (not actually
    a function of y, p since resid is linear in p)
    """
    print '\n in J'
    result = zeros([n,m], 'd')
    result[0, :] = -ones(m, 'd')
    result[1, :] = -x
    result[2, :] = -x**2
    result[3, :] = -x**3
    print result
    return result

def resid(p, y, x):
    return y - evalY(p, x)

y_true = evalY(coeffs, x)
y_meas = y_true + yErrs

p0 = [1., 1., 1., 1.]

pMin, C, infodict, ier, mesg = leastsq(resid, p0, args=(y_meas, x),  full_output=True)

pMin2, C2, infodict2, ier2, mesg2 = leastsq(resid, p0, Dfun=J, col_deriv=1, args=(y_meas, x),  full_output=True)

#check against known algorithms for computing C
JAtPMin = J(pMin, None, x)
Q_true, R_true = linalg.qr(JAtPMin)
C_true = linalg.inv(dot(JAtPMin, transpose(JAtPMin)))
C_true2 = linalg.inv(dot(R_true, transpose(R_true)))

print 'pMin'
print pMin
print 'pMin2'
print pMin2
print '\nC'
print C
print '\nC2'
print C2
print '\nC_true'
print C_true
print '\nC_true2'
print C_true2



 
---------------------------------
Don't be flakey. Get Yahoo! Mail for Mobile and 
always stay connected to friends.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070318/82ee5416/attachment.html>

From nvf at uwm.edu  Sun Mar 18 14:06:24 2007
From: nvf at uwm.edu (Nick Fotopoulos)
Date: Sun, 18 Mar 2007 13:06:24 -0500
Subject: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9: _fftpack.so
	problems?
Message-ID: <f687be130703181106h69508c93had66c2090356f4fc@mail.gmail.com>

On 3/18/07, scipy-user-request at scipy.org <scipy-user-request at scipy.org> wrote:
> Message: 2
> Date: Sat, 17 Mar 2007 14:41:33 -0500
> From: Robert Kern <robert.kern at gmail.com>
> Subject: Re: [SciPy-user] Plea: Installing SciPy on Mac 10.4.9:
>         _fftpack.so problems?
> To: SciPy Users List <scipy-user at scipy.org>
> Message-ID: <45FC446D.7000409 at gmail.com>
> Content-Type: text/plain; charset=UTF-8
>
> Nick Fotopoulos wrote:
>
> > <all the missing _Py symbols>
> > collect2: ld returned 1 exit status
> > error: Command "/usr/local/bin/gfortran -Wall -bundle
> > build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/Lib/fftpack/_fftpackmodule.o
> > build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfft.o
> > build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/drfft.o
> > build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zrfft.o
> > build/temp.macosx-10.3-fat-2.5/Lib/fftpack/src/zfftnd.o
> > build/temp.macosx-10.3-fat-2.5/build/src.macosx-10.3-fat-2.5/fortranobject.o
> > -L/opt/lscsoft/non-lsc/lib
> > -L/usr/local/lib/gcc/i386-apple-darwin8.8.1/4.3.0
> > -Lbuild/temp.macosx-10.3-fat-2.5 -ldfftpack -lfftw3 -lgfortran -o
> > build/lib.macosx-10.3-fat-2.5/scipy/fftpack/_fftpack.so" failed with
> > exit status 1
> >
> > Many thanks for the information and for taking a shot at this.
>
> Where is this coming from if not LDFLAGS?
>
>  -L/opt/lscsoft/non-lsc/lib

nvf at dirac:~/temp/scipy-svn$ cat site.cfg
[fftw3]
library_dirs = /opt/lscsoft/non-lsc/lib
fftw3_libs = fftw3
include_dirs = /opt/lscsoft/non-lsc/include


From acorriga at gmu.edu  Sun Mar 18 15:27:04 2007
From: acorriga at gmu.edu (Andrew Corrigan)
Date: Sun, 18 Mar 2007 19:27:04 +0000 (UTC)
Subject: [SciPy-user] spsolve not working
References: <45FD40BA.1020107@gmu.edu> <web-80528738@uni-stuttgart.de>
Message-ID: <loom.20070318T202627-760@post.gmane.org>

Thanks.  I'll give that a try.

Nils Wagner <nwagner <at> iam.uni-stuttgart.de> writes:
> Fixed in svn (IIRC).
> 
> Nils
> 






From sgarcia at olfac.univ-lyon1.fr  Mon Mar 19 05:00:28 2007
From: sgarcia at olfac.univ-lyon1.fr (Samuel GARCIA)
Date: Mon, 19 Mar 2007 10:00:28 +0100
Subject: [SciPy-user] Problem with weave blitz win32
In-Reply-To: <45FA661F.9020004@olfac.univ-lyon1.fr>
References: <45FA661F.9020004@olfac.univ-lyon1.fr>
Message-ID: <45FE512C.1020208@olfac.univ-lyon1.fr>

Hi list,

does anyone have experience a scipy.weave with blitz on winXP ?
My still have the problem running the example array3d.py ...

And I have few knowledge about compilation ...

Any help would be wonderfull...

Sam

Samuel GARCIA wrote:
> Hi list,
>
> I have problem running weave on windowsXP.
>
> Version :
> Python 2.5
> MiGW 3.4.2
> Scipy 0.5.2
> Numpy 1.0.1
>
>
> array3d.py give me :
>
> numpy:
> [[[ 0  1  2  3]
>   [ 4  5  6  7]
>   [ 8  9 10 11]]
>
>  [[12 13 14 15]
>   [16 17 18 19]
>   [20 21 22 23]]]
> Pure Inline:
> <weave: compiling>
> img[  0][  0]=   0   1   2   3
> img[  0][  1]=   4   5   6   7
> img[  0][  2]=   8   9  10  11
> img[  1][  0]=  12  13  14  15
> img[  1][  1]=  16  17  18  19
> img[  1][  2]=  20  21  22  23
> Blitz Inline:
> <weave: compiling>
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:5:
> warning: ignoring #pragma warning
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:6:
> warning: ignoring #pragma warning
> In file included from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/applics.h:400,
>                  from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/vecexpr.h:32,
>                  from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/vecpick.cc:16,
>                  from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/vecpick.h:293,
>                  from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/vector.h:449,
>                  from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/tinyvec.h:430,
>                  from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array-impl.h:44,
>                  from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array.h:32,
>                  from
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:9:
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/mathfunc.h: In
> static member function `static double
> blitz::_bz_expm1<P_numtype1>::apply(P_numtype1)':
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/mathfunc.h:1353:
> error: `::expm1' has not been declared
> In file included from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array/funcs.h:29,
>                  from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array/newet.h:29,
>                  from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array/et.h:27,
>                  from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array-impl.h:2515,
>                  from
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/array.h:32,
>                  from
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:9:
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/funcs.h: In
> static member function `static T_numtype1
> blitz::Fn_expm1<T_numtype1>::apply(T_numtype1)':
> G:/Python25/lib/site-packages/scipy/weave/blitz/blitz/funcs.h:113:
> error: `::expm1' has not been declared
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:
> In function `PyObject* file_to_py(FILE*, char*, char*)':
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:404:
> warning: unused variable 'py_obj'
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:
> In function `blitz::Array<P_numtype, N_rank>
> convert_to_blitz(PyArrayObject*, const char*) [with T = int, int N = 3]':
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:708:  
> instantiated from here
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp:657:
> warning: unused variable 'stride_acc'
> Traceback (most recent call last):
>   File "array3d.py", line 105, in <module>
>     main()
>   File "array3d.py", line 101, in main
>     blitz_inline(arr)
>   File "array3d.py", line 89, in blitz_inline
>     weave.inline(code, ['arr'], type_converters=converters.blitz)
>   File "G:\Python25\Lib\site-packages\scipy\weave\inline_tools.py",
> line 339, in inline
>     **kw)
>   File "G:\Python25\Lib\site-packages\scipy\weave\inline_tools.py",
> line 447, in compile_function
>     verbose=verbose, **kw)
>   File "G:\Python25\Lib\site-packages\scipy\weave\ext_tools.py", line
> 365, in compile
>     verbose = verbose, **kw)
>   File "G:\Python25\Lib\site-packages\scipy\weave\build_tools.py",
> line 269, in build_extension
>     setup(name = module_name, ext_modules = [ext],verbose=verb)
>   File "G:\Python25\Lib\site-packages\numpy\distutils\core.py", line
> 174, in setup
>     return old_setup(**new_attr)
>   File "G:\Python25\lib\distutils\core.py", line 168, in setup
>     raise SystemExit, "error: " + str(msg)
> distutils.errors.CompileError: error: Command "g++ -mno-cygwin -O2
> -Wall -IG:\Python25\lib\site-packages\scipy\weave
> -IG:\Python25\lib\site-packages\scipy\weave\scxx
> -IG:\Python25\lib\site-packages\scipy\weave\blitz
> -IG:\Python25\lib\site-packages\numpy\core\include
> -IG:\Python25\include -IG:\Python25\PC -c
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.cpp
> -o
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_intermediate\compiler_894ad5ed761bb51736c6d2b7872dc212\Release\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\sc_49e94d1bdd1ad16917064c910093194f0.o"
> failed with exit status 1
>
>
> And weave.test() give me :
>
>
> Found 0 tests for scipy.weave.c_spec
>   Found 2 tests for scipy.weave.blitz_tools
> building extensions here:
> g:\docume~1\sgarcia\locals~1\temp\sgarcia\python25_compiled\m7
>   Found 1 tests for scipy.weave.ext_tools
>   Found 9 tests for scipy.weave.build_tools
>   Found 0 tests for scipy.weave.inline_tools
>   Found 1 tests for scipy.weave.ast_tools
> Warning: FAILURE importing tests for <module 'scipy.weave.wx_spec'
> from '...ckages\\scipy\\weave\\wx_spec.pyc'>
> G:\Python25\Lib\site-packages\scipy\weave\tests\test_wx_spec.py:16:
> ImportError: No module named wxPython (in <module>)
>   Found 3 tests for scipy.weave.standard_array_spec
>   Found 74 tests for scipy.weave.size_check
>   Found 26 tests for scipy.weave.catalog
>   Found 16 tests for scipy.weave.slice_handler
>   Found 0 tests for __main__
> ...warning: specified build_dir '_bad_path_' does not exist or is not
> writable. Trying default locations
> .....warning: specified build_dir '_bad_path_' does not exist or is
> not writable. Trying default locations
> ....................................F..F..................................................................removing
> 'g:\docume~1\sgarcia\locals~1\temp\tmpga71mmcat_test' (and everything
> under it)
> error removing g:\docume~1\sgarcia\locals~1\temp\tmpga71mmcat_test:
> g:\docume~1\sgarcia\locals~1\temp\tmpga71mmcat_test: Le r?pertoire
> n'est pas vide
> .removing 'g:\docume~1\sgarcia\locals~1\temp\tmpimmvs6cat_test' (and
> everything under it)
> .................
> ======================================================================
> FAIL: check_1d_3
> (scipy.weave.tests.test_size_check.test_dummy_array_indexing)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File
> "G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py",
> line 168, in check_1d_3
>     self.generic_1d('a[-11:]')
>   File
> "G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py",
> line 135, in generic_1d
>     self.generic_wrap(a,expr)
>   File
> "G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py",
> line 127, in generic_wrap
>     self.generic_test(a,expr,desired)
>   File
> "G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py",
> line 123, in generic_test
>     assert_array_equal(actual,desired, expr)
>   File "G:\Python25\lib\site-packages\numpy\testing\utils.py", line
> 223, in assert_array_equal
>     verbose=verbose, header='Arrays are not equal')
>   File "G:\Python25\lib\site-packages\numpy\testing\utils.py", line
> 215, in assert_array_compare
>     assert cond, msg
> AssertionError:
> Arrays are not equal
> a[-11:]
> (mismatch 100.0%)
>  x: array([1])
>  y: array([10])
>
> ======================================================================
> FAIL: check_1d_6
> (scipy.weave.tests.test_size_check.test_dummy_array_indexing)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File
> "G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py",
> line 174, in check_1d_6
>     self.generic_1d('a[:-11]')
>   File
> "G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py",
> line 135, in generic_1d
>     self.generic_wrap(a,expr)
>   File
> "G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py",
> line 127, in generic_wrap
>     self.generic_test(a,expr,desired)
>   File
> "G:\Python25\Lib\site-packages\scipy\weave\tests\test_size_check.py",
> line 123, in generic_test
>     assert_array_equal(actual,desired, expr)
>   File "G:\Python25\lib\site-packages\numpy\testing\utils.py", line
> 223, in assert_array_equal
>     verbose=verbose, header='Arrays are not equal')
>   File "G:\Python25\lib\site-packages\numpy\testing\utils.py", line
> 215, in assert_array_compare
>     assert cond, msg
> AssertionError:
> Arrays are not equal
> a[:-11]
> (mismatch 100.0%)
>  x: array([9])
>  y: array([0])
>
> ----------------------------------------------------------------------
> Ran 132 tests in 2.266s
>
> FAILED (failures=2)
> >Exit code: 0
>
>
> And of course my main problem is that some of my code that work on
> linux debian does'nt work on win32.
> (But it used to work on the old enthon distribution version!)
>
> Any idea ?
>
> Thank
>
> Sam
>
>
>
> -- 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Samuel Garcia
> Laboratoire de Neurosciences Sensorielles, Comportement, Cognition.
> CNRS - UMR5020 - Universite Claude Bernard LYON 1
> Equipe logisique et technique
> 50, avenue Tony Garnier
> 69366 LYON Cedex 07
> FRANCE
> T?l : 04 37 28 74 64
> Fax : 04 37 28 76 01
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   
> ------------------------------------------------------------------------
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Samuel Garcia
Laboratoire de Neurosciences Sensorielles, Comportement, Cognition.
CNRS - UMR5020 - Universite Claude Bernard LYON 1
Equipe logisique et technique
50, avenue Tony Garnier
69366 LYON Cedex 07
FRANCE
T?l : 04 37 28 74 64
Fax : 04 37 28 76 01
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070319/40a0a822/attachment.html>

From nogradi at gmail.com  Mon Mar 19 06:28:13 2007
From: nogradi at gmail.com (Daniel Nogradi)
Date: Mon, 19 Mar 2007 11:28:13 +0100
Subject: [SciPy-user] from maple to scipy/numpy
Message-ID: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>

[I sent this message to c.l.p and was told to come here]

I'm just getting started with numpy/scipy and first would like to get
a view of what it can do and what it can't. The main idea is to move
from maple to python and the first thing that poped up is the fact
that maple is very convenient for both formal manipulations and exact
integer calculations. For instance if a matrix has integer entries and
the eigenvalues are integers maple can find these exactly.

Can numpy/scipy do this? Or the eigenvalues will always be floating
point numbers?

A more general question, to what extent can numpy/scipy do symbolic
(formal) manipulations? (I know this is rather vague, but for those of
you who are familiar with maple/mathematica I guess it's clear.)


From bruno.chazelas at ias.u-psud.fr  Mon Mar 19 06:37:58 2007
From: bruno.chazelas at ias.u-psud.fr (bruno)
Date: Mon, 19 Mar 2007 11:37:58 +0100
Subject: [SciPy-user] from maple to scipy/numpy
In-Reply-To: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>
Message-ID: <45FE6806.4040003@ias.u-psud.fr>

Hello,

though I am rather new to python/scipy I think it is more focalized on 
numerical calculation than on symbolical calculation. A more straight 
forward replacement for Maple would be Maxima I think.

http://maxima.sourceforge.net/

Bruno


Daniel Nogradi a ?crit :
> [I sent this message to c.l.p and was told to come here]
>
> I'm just getting started with numpy/scipy and first would like to get
> a view of what it can do and what it can't. The main idea is to move
> from maple to python and the first thing that poped up is the fact
> that maple is very convenient for both formal manipulations and exact
> integer calculations. For instance if a matrix has integer entries and
> the eigenvalues are integers maple can find these exactly.
>
> Can numpy/scipy do this? Or the eigenvalues will always be floating
> point numbers?
>
> A more general question, to what extent can numpy/scipy do symbolic
> (formal) manipulations? (I know this is rather vague, but for those of
> you who are familiar with maple/mathematica I guess it's clear.)
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   



From jan at aims.ac.za  Mon Mar 19 06:42:38 2007
From: jan at aims.ac.za (Jan Groenewald)
Date: Mon, 19 Mar 2007 12:42:38 +0200
Subject: [SciPy-user] from maple to scipy/numpy
In-Reply-To: <45FE6806.4040003@ias.u-psud.fr>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>
	<45FE6806.4040003@ias.u-psud.fr>
Message-ID: <20070319104238.GK8808@aims.ac.za>

Hi

On Mon, Mar 19, 2007 at 11:37:58AM +0100, bruno wrote:
> though I am rather new to python/scipy I think it is more focalized on 
> numerical calculation than on symbolical calculation. A more straight 
> forward replacement for Maple would be Maxima I think.
> 
> http://maxima.sourceforge.net/

SAGE is probably what you need. It integrates maxima, Gap, NLT, 
Pari-GP, with python and scipy.

http://modular.math.washington.edu/sage/

To install on debian-based systems (binary):
http://modular.math.washington.edu/SAGEbin/linux_32bit/sage-2.3-debian-i686-Linux.tar.gz

Or elsewhere:
http://modular.math.washington.edu/sage/download.html

Their blurb:
SAGE is free open source math software that supports research and teaching in algebra, geometry, number theory, cryptography, numerical computation, and related areas. Both the SAGE development model and the technology in SAGE itself is distinguished by an extremely strong emphasis on openness, community, cooperation, and collaboration: we are building the car, not reinventing the wheel. Our overall goal is to create a viable free open source alternative to commercial mathematics software.

SAGE includes the following core software:

Group theory and combinatorics 	GAP, NetworkX
Symbolic computation and Calculus 	Maxima
Commutative algebra 	Singular
Number theory 	PARI, MWRANK, NTL
Graphics 	Matplotlib
Numerical methods 	GSL, Numpy
Mainstream programming language 	Python
Interactive shell 	IPython
Graphical User Interface 	The SAGE Notebook
Versioned Source Tracking 	Mercurial HG

cheers,
Jan
-- 
   .~.
   /V\     Jan Groenewald
  /( )\    www.aims.ac.za
  ^^-^^


From cimrman3 at ntc.zcu.cz  Mon Mar 19 06:44:15 2007
From: cimrman3 at ntc.zcu.cz (Robert Cimrman)
Date: Mon, 19 Mar 2007 11:44:15 +0100
Subject: [SciPy-user] FEM
In-Reply-To: <45FB4CFF.5050901@gmail.com>
References: <45FB4CFF.5050901@gmail.com>
Message-ID: <45FE697F.8090308@ntc.zcu.cz>

H?kan Jakobsson wrote:
> Hi list,
> Python/scipy newbie here. I'm currently writing my master's thesis in 
> computational mathematics - discontinuous galerkin methods to be 
> precise. I'v been doing all the numerics for my thesis in Matlab but 
> stumbled upon Python and scipy by chance. As a little experiment I 
> translated one of my solvers into Python using numpy, scipy and the 
> pysparse packages, and I was very impressed with the effort/performance 
> ratio. Now, to the point..
> 
> If you're doing finite element analysis, what are you're experiences 
> using Python? Do you use Python as your main development tool? In 
> conjunction with C, Fortran? Is there any drawbacks with using Python 
> for this type of work? Anything else?
> 
> I hope I'm not asking to much here, but it would be really interesting 
> to know a bit about what the situation's like. I think Python and scipy 
> is really great, and now I'm trying to convince everyone else...
> 
> Please, share with me your experiences.

I have switched to Python + scipy + matplotlib + ... combo two years ago
and never regretted that decision. You can see the result at
http://ui505p06-mbs.ntc.zcu.cz/sfe/SFE.

Before I coded a lot in matlab (a heart simulation FE code, with inverse
problem solution capability), writing critical parts as C mex-files. It
was rather painful comparing to ease of SWIG or f2py or ctypes or (plug
your favourite wrapper generator here). The pass-by-value (all function
arguments are immutable) concept of matlab was one of the main obstacles
to get a decent speed - we had to use dirty tricks to override this.

The combination Python + C (or fortran or...) is a really powerful
concept. Moreover numpy and scipy advances in big paces, so if you miss
something now, there is a great chance you will not miss it in a few
weeks...

cheers,
r.


From david at ar.media.kyoto-u.ac.jp  Mon Mar 19 06:47:42 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Mon, 19 Mar 2007 19:47:42 +0900
Subject: [SciPy-user] FEM
In-Reply-To: <45FE697F.8090308@ntc.zcu.cz>
References: <45FB4CFF.5050901@gmail.com> <45FE697F.8090308@ntc.zcu.cz>
Message-ID: <45FE6A4E.1080206@ar.media.kyoto-u.ac.jp>

Robert Cimrman wrote:
>
> Before I coded a lot in matlab (a heart simulation FE code, with inverse
> problem solution capability), writing critical parts as C mex-files. It
> was rather painful comparing to ease of SWIG or f2py or ctypes or (plug
> your favourite wrapper generator here). The pass-by-value (all function
> arguments are immutable) concept of matlab was one of the main obstacles
> to get a decent speed - we had to use dirty tricks to override this.
To be exact, matlab uses COW to pass arguments efficiently. Otherwise, I 
agree entirely, interfacing matlab with C is really a painful 
experience: there is no way to do it efficiently and nicely at the same 
time, the C-API of matlab being far from complete for anything non 
trivial. That was one of my main reason to try (and stay with 
afterwards) python.

David


From nauss at lcrs.de  Mon Mar 19 06:52:51 2007
From: nauss at lcrs.de (Thomas Nauss)
Date: Mon, 19 Mar 2007 11:52:51 +0100
Subject: [SciPy-user] Next neighbor interpolation inside regular gridded
	arrays
Message-ID: <45FE6B83.10902@lcrs.de>

Hi everyone,
I have projected raw swath satellite data to map coordinates and stored 
the data values in a 2D target array that has the geometry of the output 
dataset already. I have initialized the target array using Numeric.zeros 
and since not all fields of the target array had corresponding fields in 
the input dataset, some fields of the target array have still the value 0.
Is there a function which performs a next neighbor interpolation for 
each array field with a specified value?
Cheers,
Thomas


From nogradi at gmail.com  Mon Mar 19 06:54:13 2007
From: nogradi at gmail.com (Daniel Nogradi)
Date: Mon, 19 Mar 2007 11:54:13 +0100
Subject: [SciPy-user] from maple to scipy/numpy
In-Reply-To: <20070319104238.GK8808@aims.ac.za>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>
	<45FE6806.4040003@ias.u-psud.fr> <20070319104238.GK8808@aims.ac.za>
Message-ID: <5f56302b0703190354ub83a380u319c38ec31dd4c9d@mail.gmail.com>

> > though I am rather new to python/scipy I think it is more focalized on
> > numerical calculation than on symbolical calculation. A more straight
> > forward replacement for Maple would be Maxima I think.
> >
> > http://maxima.sourceforge.net/
>
> SAGE is probably what you need. It integrates maxima, Gap, NLT,
> Pari-GP, with python and scipy.
>
> http://modular.math.washington.edu/sage/
>
> To install on debian-based systems (binary):
> http://modular.math.washington.edu/SAGEbin/linux_32bit/sage-2.3-debian-i686-Linux.tar.gz
>
> Or elsewhere:
> http://modular.math.washington.edu/sage/download.html


Thanks very much for the advice. These two packages (sage and maxima)
seem all right for symbolic manipulation indeed.

Concerning integer arithmetics for example for matrix eigenvalues
aren't they a bit of an overkill? Can't numpy/scipy do that by itself?


From nielsellegaard at gmail.com  Mon Mar 19 05:32:13 2007
From: nielsellegaard at gmail.com (Niels L. Ellegaard)
Date: Mon, 19 Mar 2007 10:32:13 +0100
Subject: [SciPy-user] FEM
References: <45FB4CFF.5050901@gmail.com>
Message-ID: <87aby9lgte.fsf@gmail.com>

H?kan Jakobsson <hakan.jakobsson at gmail.com> writes:
> If you're doing finite element analysis, what are you're experiences 
> using Python? Do you use Python as your main development tool? In 
> conjunction with C, Fortran? Is there any drawbacks with using Python 
> for this type of work? Anything else?

You may wish to have a look at fenics. It's  a FEM engine with a python
frontend which I believe is based on numpy.

I must admit that I haven't tried the program, but judging from the
webpage it is pretty cool :)

http://www.fenics.org/

     Niels





From bryanv at enthought.com  Mon Mar 19 09:23:10 2007
From: bryanv at enthought.com (Bryan Van de Ven)
Date: Mon, 19 Mar 2007 08:23:10 -0500
Subject: [SciPy-user] from maple to scipy/numpy
In-Reply-To: <20070319104238.GK8808@aims.ac.za>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>	<45FE6806.4040003@ias.u-psud.fr>
	<20070319104238.GK8808@aims.ac.za>
Message-ID: <45FE8EBE.1040002@enthought.com>


SAGE looks pretty cool, I would just add PyDX

http://gr.anu.edu.au/svn/people/sdburton/pydx/doc/user-guide.html

as a package for differential geometry.


Jan Groenewald wrote:
> SAGE includes the following core software:
> 
> Group theory and combinatorics 	GAP, NetworkX
> Symbolic computation and Calculus 	Maxima
> Commutative algebra 	Singular
> Number theory 	PARI, MWRANK, NTL
> Graphics 	Matplotlib
> Numerical methods 	GSL, Numpy
> Mainstream programming language 	Python
> Interactive shell 	IPython
> Graphical User Interface 	The SAGE Notebook
> Versioned Source Tracking 	Mercurial HG
> 
> cheers,
> Jan



From hakan.jakobsson at gmail.com  Mon Mar 19 10:51:24 2007
From: hakan.jakobsson at gmail.com (=?ISO-8859-1?Q?H=E5kan_Jakobsson?=)
Date: Mon, 19 Mar 2007 15:51:24 +0100
Subject: [SciPy-user] FEM
In-Reply-To: <45FE697F.8090308@ntc.zcu.cz>
References: <45FB4CFF.5050901@gmail.com> <45FE697F.8090308@ntc.zcu.cz>
Message-ID: <687bb3e80703190751l582e407epaf31d0d685fff50d@mail.gmail.com>

Thanks everyone for your replies, they have been most helpful. I've studied
the pros and cons for python, and I'm fairly sure I'll be using python more
and more.

Your codes, Robert, seems very interesting. I've downloaded the slides from
your website and I'll be looking through them. They seem to adresss many of
the questions I'd like answered. Furthermore, It's comforting to see that
people out there are actually writing research codes using python.

I'm sure I'll be posting more questions here in the future.

Again thanks, and I certainly will keep on evangelizing.
/Hakan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070319/148dedcf/attachment.html>

From Andrew.Reid at nist.gov  Mon Mar 19 11:03:08 2007
From: Andrew.Reid at nist.gov (Andrew Reid)
Date: Mon, 19 Mar 2007 11:03:08 -0400
Subject: [SciPy-user] FEM
In-Reply-To: <687bb3e80703190751l582e407epaf31d0d685fff50d@mail.gmail.com>
References: <45FB4CFF.5050901@gmail.com> <45FE697F.8090308@ntc.zcu.cz>
	<687bb3e80703190751l582e407epaf31d0d685fff50d@mail.gmail.com>
Message-ID: <20070319150308.GD14247@smithers.nist.gov>

On Mon, Mar 19, 2007 at 03:51:24PM +0100, H?kan Jakobsson wrote:
> Thanks everyone for your replies, they have been most helpful. I've studied
> the pros and cons for python, and I'm fairly sure I'll be using python more
> and more.

  But wait!  It's not over yet!

  I'm part of a team working on an FEM project at NIST which 
uses Python quite heavily.  Our brief is to address Materials
Science users, so we have a lot of custom tools for adapting
our mesh to user-provided microstructural image, and we also
have an extensible interface for adding new types of 
user-provided custom constitutive rules.  The actual FEM 
part is pretty standard.

  Our numerical work and solvers are all in C++, 
and we don't actually use SciPy/NumPy, we just use SWIG
to get to the numerical code.  

  The flexibility of Python has been crucial in working through 
our API and mesh-construction tools -- it has the right level of 
easy scriptability, and the right sort of object-orientation,
for what we need.

  The project page is at <http://www.ctcms.nist.gov/oof>, 
and there's a relatively recent talk (in PDF form) at 
<http://www.cacr.caltech.edu/projects/danse/talks/kickoff/22-Reid/danse_oof.pdf>, 
which very strongly resembles the talk I gave at SciPy2006.

				-- A.
-- 
Dr. Andrew C. E. Reid, Guest Researcher 
Center for Theoretical and Computational Materials Science
National Institute of Standards and Technology, Mail Stop 8910
Gaithersburg MD 20899 USA
andrew.reid at nist.gov


From dschult at colgate.edu  Mon Mar 19 12:19:25 2007
From: dschult at colgate.edu (Dan Schult)
Date: Mon, 19 Mar 2007 12:19:25 -0400
Subject: [SciPy-user] ld unknown flag macosx_version_min
In-Reply-To: <45FC4F3B.6050302@gmail.com>
References: <FE1EFC88-5265-4BF8-8BF9-EAB1362CAEC5@colgate.edu>
	<45FC4F3B.6050302@gmail.com>
Message-ID: <3B3E8AF7-A043-4FD6-8EAC-673B200DF199@colgate.edu>


On Mar 17, 2007, at 4:27 PM, Robert Kern wrote:
> Dan Schult wrote:
>> I've got a Mac OSx 10.4.8 machine and am compiling
>> scipy according to the instructions on the webpage.  I've
>> got gcc 4.0.0 gfortran 4.3.0 fftw3.0 and svn versions of numpy
>> and scipy.   My python is version 2.5.
>> Building numpy goes smoothly, but when I try scipy I have an ld  
>> error.
>
> I have gcc 4.0.1 and gfortran 4.3.0 installed on my system, and I  
> do not see
> this problem. Can you try upgrading to the latest version of Xcode  
> (which should
> have gcc 4.0.1)? It's not coming from Python but, I suspect, gfortran.
> Robert Kern

Indeed-- after an almost 1Gig download of XCode 2.4.1 to get gcc 4.0.1
it compiled and runs. :)  The tests give me 4 failed, but these  
already have
been noted on the email list.

I assume if I had compiled gfortran with gcc4.0 it would have worked the
first time.  I always seem to have trouble when using precompiled  
binaries.
Thanks very much Robert for the helpful tip!


From robert.kern at gmail.com  Mon Mar 19 12:43:14 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Mon, 19 Mar 2007 11:43:14 -0500
Subject: [SciPy-user] from maple to scipy/numpy
In-Reply-To: <5f56302b0703190354ub83a380u319c38ec31dd4c9d@mail.gmail.com>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>	<45FE6806.4040003@ias.u-psud.fr>
	<20070319104238.GK8808@aims.ac.za>
	<5f56302b0703190354ub83a380u319c38ec31dd4c9d@mail.gmail.com>
Message-ID: <45FEBDA2.9070902@gmail.com>

Daniel Nogradi wrote:

> Concerning integer arithmetics for example for matrix eigenvalues
> aren't they a bit of an overkill? Can't numpy/scipy do that by itself?

I guess they could if someone were to implement such algorithms for us. The
focus of contributions has largely been for floating point algorithms, so we use
LAPACK to do most of our linear algebra.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From ggellner at uoguelph.ca  Mon Mar 19 12:55:34 2007
From: ggellner at uoguelph.ca (Gabriel Gellner)
Date: Mon, 19 Mar 2007 12:55:34 -0400
Subject: [SciPy-user] from maple to scipy/numpy
In-Reply-To: <45FE8EBE.1040002@enthought.com>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>
	<45FE6806.4040003@ias.u-psud.fr> <20070319104238.GK8808@aims.ac.za>
	<45FE8EBE.1040002@enthought.com>
Message-ID: <20070319165534.GG31245@encolpuis>

I second PyDX it is great, I use it mostly for the automatic
derivatives, very nice if you don't need compiled speed (it is still
quite fast anyway).

Gabriel

On Mon, Mar 19, 2007 at 08:23:10AM -0500, Bryan Van de Ven wrote:
> 
> SAGE looks pretty cool, I would just add PyDX
> 
> http://gr.anu.edu.au/svn/people/sdburton/pydx/doc/user-guide.html
> 
> as a package for differential geometry.
> 
> 
> Jan Groenewald wrote:
> > SAGE includes the following core software:
> > 
> > Group theory and combinatorics 	GAP, NetworkX
> > Symbolic computation and Calculus 	Maxima
> > Commutative algebra 	Singular
> > Number theory 	PARI, MWRANK, NTL
> > Graphics 	Matplotlib
> > Numerical methods 	GSL, Numpy
> > Mainstream programming language 	Python
> > Interactive shell 	IPython
> > Graphical User Interface 	The SAGE Notebook
> > Versioned Source Tracking 	Mercurial HG
> > 
> > cheers,
> > Jan
> 
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user


From simon at arrowtheory.com  Mon Mar 19 13:31:34 2007
From: simon at arrowtheory.com (Simon Burton)
Date: Mon, 19 Mar 2007 10:31:34 -0700
Subject: [SciPy-user] from maple to scipy/numpy
In-Reply-To: <20070319165534.GG31245@encolpuis>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>
	<45FE6806.4040003@ias.u-psud.fr> <20070319104238.GK8808@aims.ac.za>
	<45FE8EBE.1040002@enthought.com> <20070319165534.GG31245@encolpuis>
Message-ID: <20070319103134.f71de165.simon@arrowtheory.com>

On Mon, 19 Mar 2007 12:55:34 -0400
Gabriel Gellner <ggellner at uoguelph.ca> wrote:

> 
> I second PyDX it is great, I use it mostly for the automatic
> derivatives, very nice if you don't need compiled speed (it is still
> quite fast anyway).
> 
> Gabriel

wow, a user! (*blushing*).

Simon.


From stefan at sun.ac.za  Mon Mar 19 16:24:43 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Mon, 19 Mar 2007 20:24:43 +0000
Subject: [SciPy-user] Next neighbor interpolation inside regular
	griddedarrays
In-Reply-To: <45FE6B83.10902@lcrs.de>
References: <45FE6B83.10902@lcrs.de>
Message-ID: <a5f999804aa93f6ef178fb91facabe7d@localhost>


Hi Thomas

On Mon, 19 Mar 2007 11:52:51 +0100, Thomas Nauss <nauss at lcrs.de> wrote:
> Hi everyone,
> I have projected raw swath satellite data to map coordinates and stored
> the data values in a 2D target array that has the geometry of the output
> dataset already. I have initialized the target array using Numeric.zeros
> and since not all fields of the target array had corresponding fields in
> the input dataset, some fields of the target array have still the value 0.
> Is there a function which performs a next neighbor interpolation for
> each array field with a specified value?

Normally, the easiest way to work around this problem is to perform the
process in reverse.  I.e. for every coordinate in the output frame, do the
inverse mapping to find the coordinate in the input frame, and, using
interpolation, calculate the value.

If you are interested, I have code available that does all sorts of warping,
or you can take a look at scipy's ndimage.map_coordinates.

Cheers
St?fan




From joishi at amnh.org  Mon Mar 19 18:09:11 2007
From: joishi at amnh.org (J Oishi)
Date: Mon, 19 Mar 2007 18:09:11 -0400
Subject: [SciPy-user] wavevector arrays
In-Reply-To: <mailman.9.1174323608.22841.scipy-user@scipy.org>
References: <mailman.9.1174323608.22841.scipy-user@scipy.org>
Message-ID: <9F48FF87-3B7A-4763-9186-68AC32AF577A@amnh.org>

Hi,

I am performing a number of 3D FFTs using scipy, and I had a question  
as to how to setup 3 3D arrays of wavevectors, much like one might do  
with mgrid. However, because the FFT returns wavevectors running from  
0,...,k_c, -k_c, ..., -1 (where k_c is the nyquist wavenumber) I'm  
not sure how to get mgrid to do this.

In 2D, I could create two 1D kx and ky arrays and use meshgrid (for  
an even number of points):

import numpy as N

kx = 2*N.pi/Lx * N.concatenate(N.arange(0,nx/2-1),N.arange(-nx/2, 0))
ky = 2*N.pi/Ly * N.concatenate(N.arange(0,ny/2-1),N.arange(-ny/2, 0))

kk = meshgrid(kx,ky)

where kk would have a shape (2,ny,nx). However, this (to my  
knowledge) doesn't generalize to 3D.

I'm a total python/scipy neophyte, so any help you could provide  
would be quite appreciated.

thanks,

jeff


From steve at shrogers.com  Mon Mar 19 23:16:07 2007
From: steve at shrogers.com (Steven H. Rogers)
Date: Mon, 19 Mar 2007 21:16:07 -0600
Subject: [SciPy-user] John Backus R.I.P.
Message-ID: <45FF51F7.8090009@shrogers.com>

John Backus, the father of Fortran, `has died 
<http://www.nytimes.com/2007/03/19/obituaries/20cnd-backus.html?_r=2&oref=slogin&oref=slogin>`_. 




From topengineer at gmail.com  Tue Mar 20 03:20:45 2007
From: topengineer at gmail.com (Hui Chang Moon)
Date: Tue, 20 Mar 2007 16:20:45 +0900
Subject: [SciPy-user] I can't find unit step function.
Message-ID: <296323b50703200020m2a33d3d8kc0bd0f0c8a9cac3e@mail.gmail.com>

Hello, users,

Scipy has many convenient functions, so I can't find the function that I am
willing to use.
So, anybody help me find the unit step function.

Thank you~.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070320/6463dbd0/attachment.html>

From david.warde.farley at utoronto.ca  Tue Mar 20 03:38:05 2007
From: david.warde.farley at utoronto.ca (David Warde-Farley)
Date: Tue, 20 Mar 2007 03:38:05 -0400
Subject: [SciPy-user] I can't find unit step function.
In-Reply-To: <296323b50703200020m2a33d3d8kc0bd0f0c8a9cac3e@mail.gmail.com>
References: <296323b50703200020m2a33d3d8kc0bd0f0c8a9cac3e@mail.gmail.com>
Message-ID: <1174376285.28167.6.camel@rodimus>

On Tue, 2007-03-20 at 16:20 +0900, Hui Chang Moon wrote:
> Hello, users,
>  
> Scipy has many convenient functions, so I can't find the function that
> I am willing to use.
> So, anybody help me find the unit step function.

That might be just a little too basic for a built-in. But you can easily
define it.

def step(x):
	if x == 0: return 0.5
	return x > 0 and 1 or 0

(Your idea of what constitutes 'the' unit step function may vary)

Cheers,

David




From peridot.faceted at gmail.com  Tue Mar 20 03:45:12 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Tue, 20 Mar 2007 03:45:12 -0400
Subject: [SciPy-user] I can't find unit step function.
In-Reply-To: <1174376285.28167.6.camel@rodimus>
References: <296323b50703200020m2a33d3d8kc0bd0f0c8a9cac3e@mail.gmail.com>
	<1174376285.28167.6.camel@rodimus>
Message-ID: <ce557a360703200045t4d2a6a2fnca1197a4df76153b@mail.gmail.com>

On 20/03/07, David Warde-Farley <david.warde.farley at utoronto.ca> wrote:
> On Tue, 2007-03-20 at 16:20 +0900, Hui Chang Moon wrote:
> > Hello, users,
> >
> > Scipy has many convenient functions, so I can't find the function that
> > I am willing to use.
> > So, anybody help me find the unit step function.
>
> That might be just a little too basic for a built-in. But you can easily
> define it.
>
> def step(x):
>         if x == 0: return 0.5
>         return x > 0 and 1 or 0

You might prefer that it be vectorized:

In [7]: def step(x):
   ...:     return asarray(x>=0,dtype=float)
   ...:

In [8]: step(linspace(-1,1,9))
Out[8]: array([ 0.,  0.,  0.,  0.,  1.,  1.,  1.,  1.,  1.])

Though if you prefer step(0)==0.5 it's a bit of a pain.

Anne


From topengineer at gmail.com  Tue Mar 20 03:54:35 2007
From: topengineer at gmail.com (Hui Chang Moon)
Date: Tue, 20 Mar 2007 16:54:35 +0900
Subject: [SciPy-user] I can't find unit step function.
In-Reply-To: <ce557a360703200045t4d2a6a2fnca1197a4df76153b@mail.gmail.com>
References: <296323b50703200020m2a33d3d8kc0bd0f0c8a9cac3e@mail.gmail.com>
	<1174376285.28167.6.camel@rodimus>
	<ce557a360703200045t4d2a6a2fnca1197a4df76153b@mail.gmail.com>
Message-ID: <296323b50703200054k4398ab0dje2d11012384bc242@mail.gmail.com>

Thank you for your kindness~!
Have a nice day~!!


2007/3/20, Anne Archibald <peridot.faceted at gmail.com>:
>
> On 20/03/07, David Warde-Farley <david.warde.farley at utoronto.ca> wrote:
> > On Tue, 2007-03-20 at 16:20 +0900, Hui Chang Moon wrote:
> > > Hello, users,
> > >
> > > Scipy has many convenient functions, so I can't find the function that
> > > I am willing to use.
> > > So, anybody help me find the unit step function.
> >
> > That might be just a little too basic for a built-in. But you can easily
> > define it.
> >
> > def step(x):
> >         if x == 0: return 0.5
> >         return x > 0 and 1 or 0
>
> You might prefer that it be vectorized:
>
> In [7]: def step(x):
>   ...:     return asarray(x>=0,dtype=float)
>   ...:
>
> In [8]: step(linspace(-1,1,9))
> Out[8]: array([ 0.,  0.,  0.,  0.,  1.,  1.,  1.,  1.,  1.])
>
> Though if you prefer step(0)==0.5 it's a bit of a pain.
>
> Anne
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070320/34416464/attachment.html>

From nauss at lcrs.de  Tue Mar 20 04:19:01 2007
From: nauss at lcrs.de (Thomas Nauss)
Date: Tue, 20 Mar 2007 09:19:01 +0100
Subject: [SciPy-user] Next neighbor interpolation inside
	regular	griddedarrays
In-Reply-To: <a5f999804aa93f6ef178fb91facabe7d@localhost>
References: <45FE6B83.10902@lcrs.de>
	<a5f999804aa93f6ef178fb91facabe7d@localhost>
Message-ID: <45FF98F5.8050401@lcrs.de>

Hi St?fan,
thank's for your response.
I have checked map_coordinates but it does not feature next neighbor 
interpolation and as I understood ndimage requires equally-spaced data. 
I think the equally-spaced data is the main problem since the raw 
satellite data doesn't have it (neither lat nor long is continuously 
increasing along cols or rows). That's why I didn't use the inverse way 
because for that I had to go through the entire array for every output 
pixel to check the closest lat/long combination. I also tried 
PyNGL.natgrid to get an equally spaced input grid but it takes much to 
long (> 1e6 pixels) .
I could of course warp the raw data to an equally-spaced 
(non-interpolated) lat/long coordination system so that you can access 
the value of a specific lat/long coordinate directly through the row/col 
number and perform an inverse projection afterwards - this should have a 
slightly better accuracy in the end.
Do you have any other idea?
Cheers
Thomas


Stefan van der Walt wrote:
> Hi Thomas
> 
> On Mon, 19 Mar 2007 11:52:51 +0100, Thomas Nauss <nauss at lcrs.de> wrote:
>> Hi everyone,
>> I have projected raw swath satellite data to map coordinates and stored
>> the data values in a 2D target array that has the geometry of the output
>> dataset already. I have initialized the target array using Numeric.zeros
>> and since not all fields of the target array had corresponding fields in
>> the input dataset, some fields of the target array have still the value 0.
>> Is there a function which performs a next neighbor interpolation for
>> each array field with a specified value?
> 
> Normally, the easiest way to work around this problem is to perform the
> process in reverse.  I.e. for every coordinate in the output frame, do the
> inverse mapping to find the coordinate in the input frame, and, using
> interpolation, calculate the value.
> 
> If you are interested, I have code available that does all sorts of warping,
> or you can take a look at scipy's ndimage.map_coordinates.
> 
> Cheers
> St?fan
> 
> 
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user



From david.warde.farley at utoronto.ca  Tue Mar 20 04:29:34 2007
From: david.warde.farley at utoronto.ca (David Warde-Farley)
Date: Tue, 20 Mar 2007 04:29:34 -0400
Subject: [SciPy-user] I can't find unit step function.
In-Reply-To: <ce557a360703200045t4d2a6a2fnca1197a4df76153b@mail.gmail.com>
References: <296323b50703200020m2a33d3d8kc0bd0f0c8a9cac3e@mail.gmail.com>
	<1174376285.28167.6.camel@rodimus>
	<ce557a360703200045t4d2a6a2fnca1197a4df76153b@mail.gmail.com>
Message-ID: <1174379374.5075.5.camel@rodimus>

On Tue, 2007-03-20 at 03:45 -0400, Anne Archibald wrote:

> In [7]: def step(x):
>    ...:     return asarray(x>=0,dtype=float)
>    ...:
> 
> In [8]: step(linspace(-1,1,9))
> Out[8]: array([ 0.,  0.,  0.,  0.,  1.,  1.,  1.,  1.,  1.])
> 
> Though if you prefer step(0)==0.5 it's a bit of a pain.

For the sake of completeness and utter boredom, here's the vectorized
version with step(0) == 0.5, and it's still a one-liner ;)

def step(x):
    return asarray(x>0,dtype=float)+0.5*asarray(x==0,dtype=float)

In [9]: step(arange(-5,5))
Out[9]: array([ 0. ,  0. ,  0. ,  0. ,  0. ,  0.5,  1. ,  1. ,  1. ,
1. ])

Cheers,

David



From raphael.langella at steria.com  Tue Mar 20 07:08:14 2007
From: raphael.langella at steria.com (raphael langella)
Date: Tue, 20 Mar 2007 12:08:14 +0100
Subject: [SciPy-user] building numpy/scipy on Solaris
Message-ID: <8feab0f2.b0f28fea@steria.com>

I'm trying to build numpy and scipy on Solaris 8.
The BLAS FAQ on netlib.org suggests using optimized BLAS librairies
provided by computer vendor, like the SUN Performance Library. This
library is supposed to provide enhanced and optimized version of BLAS
and LAPACK. I happen to have Forte 7 installed, so I first tried to
build against this library (libsunperf.a).
I tried several versions and different compilation options, but I always
get undefined symbols. Is this supported? As anyone ever succeeded in
using this library to compile scipy and numpy?
I also did some unsuccessful tests with Atlas. But, I think I'm now very
close to have it working with the standard BLAS/LAPACK (from
netlib.org). I have to use GNU ld (unresolved symbols when using Sun's
ld). Numpy builds and tests without any error. I've got a few errors and
failures with scipy (see attachment for the complete test report). Are
these errors critical? I don't understand where they come from and how
to correct them.
Thanks for your help.

Rapha?l Langella
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: scipy.test.txt
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070320/7dbb4363/attachment.txt>

From hakan.jakobsson at gmail.com  Tue Mar 20 07:27:18 2007
From: hakan.jakobsson at gmail.com (=?ISO-8859-1?Q?H=E5kan_Jakobsson?=)
Date: Tue, 20 Mar 2007 12:27:18 +0100
Subject: [SciPy-user] FEM
In-Reply-To: <20070319150308.GD14247@smithers.nist.gov>
References: <45FB4CFF.5050901@gmail.com> <45FE697F.8090308@ntc.zcu.cz>
	<687bb3e80703190751l582e407epaf31d0d685fff50d@mail.gmail.com>
	<20070319150308.GD14247@smithers.nist.gov>
Message-ID: <687bb3e80703200427m2477eccdwa3dd8b6722c311f4@mail.gmail.com>

I see that now :-)

Thanks for the input and the links to your project. Part of my plan is to
look into SWIG and some of the other wrapper generators out there in the
future. This does help a lot.
/H?kan J

On 3/19/07, Andrew Reid <Andrew.Reid at nist.gov> wrote:
>
> On Mon, Mar 19, 2007 at 03:51:24PM +0100, H?kan Jakobsson wrote:
> > Thanks everyone for your replies, they have been most helpful. I've
> studied
> > the pros and cons for python, and I'm fairly sure I'll be using python
> more
> > and more.
>
>   But wait!  It's not over yet!
>
>   I'm part of a team working on an FEM project at NIST which
> uses Python quite heavily.  Our brief is to address Materials
> Science users, so we have a lot of custom tools for adapting
> our mesh to user-provided microstructural image, and we also
> have an extensible interface for adding new types of
> user-provided custom constitutive rules.  The actual FEM
> part is pretty standard.
>
>   Our numerical work and solvers are all in C++,
> and we don't actually use SciPy/NumPy, we just use SWIG
> to get to the numerical code.
>
>   The flexibility of Python has been crucial in working through
> our API and mesh-construction tools -- it has the right level of
> easy scriptability, and the right sort of object-orientation,
> for what we need.
>
>   The project page is at <http://www.ctcms.nist.gov/oof>,
> and there's a relatively recent talk (in PDF form) at
> <
> http://www.cacr.caltech.edu/projects/danse/talks/kickoff/22-Reid/danse_oof.pdf
> >,
> which very strongly resembles the talk I gave at SciPy2006.
>
>                                 -- A.
> --
> Dr. Andrew C. E. Reid, Guest Researcher
> Center for Theoretical and Computational Materials Science
> National Institute of Standards and Technology, Mail Stop 8910
> Gaithersburg MD 20899 USA
> andrew.reid at nist.gov
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070320/2fe3a54f/attachment.html>

From ericpellegrini at yahoo.fr  Tue Mar 20 08:13:23 2007
From: ericpellegrini at yahoo.fr (Pellegrini Eric)
Date: Tue, 20 Mar 2007 13:13:23 +0100 (CET)
Subject: [SciPy-user] Problem to understand kstest
Message-ID: <281487.627.qm@web23405.mail.ird.yahoo.com>

Hi everybody,

I have a set of points for which I would like to know if it follows a normal distribution.
To do so, I would like to perform the KS test. Here is an example:

x = range(1000)
mean = scipy.stats.mean(x)
std = scipy.stats.std(x)
stats.kstest(x,'norm',args=(mean,std))

is it right ?

If yes, how to set up the significance level and what is its default value ?

is it a double-sided test ?

Thank you ery much

Eric Pellegrini



 		
---------------------------------
 D?couvrez une nouvelle fa?on d'obtenir des r?ponses ? toutes vos questions ! Profitez des connaissances, des opinions et des exp?riences des internautes sur Yahoo! Questions/R?ponses.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070320/766c2a4c/attachment.html>

From franckm at aims.ac.za  Tue Mar 20 08:31:35 2007
From: franckm at aims.ac.za (Franck Kalala Mutombo)
Date: Tue, 20 Mar 2007 14:31:35 +0200
Subject: [SciPy-user] animate solitons
Message-ID: <45FFD427.3090900@aims.ac.za>

hi,

I have solved the sine-cordon equation(which is a partial differential
equation) and I found my solitons(the solution to that equation), it is
easy to plot it. but want I need is to animate it in a certain range. do
you know any python module which can allows me to animate my solitons?

ps: solitons a solution of a partial differential, it is a plane wave
moving a certain speed, so I want to animate it as is moving.


-- 
Franck
African Institute for Mathematical Sciences -- www.aims.ac.za



From ryanlists at gmail.com  Tue Mar 20 08:32:19 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Tue, 20 Mar 2007 07:32:19 -0500
Subject: [SciPy-user] I can't find unit step function.
In-Reply-To: <1174379374.5075.5.camel@rodimus>
References: <296323b50703200020m2a33d3d8kc0bd0f0c8a9cac3e@mail.gmail.com>
	<1174376285.28167.6.camel@rodimus>
	<ce557a360703200045t4d2a6a2fnca1197a4df76153b@mail.gmail.com>
	<1174379374.5075.5.camel@rodimus>
Message-ID: <c5b438120703200532n49fe0d71g74519593eddaeab4@mail.gmail.com>

Sorry to contribute to this discussion out of boredom and curiosity,
but what about:

y=where(t>0,1.0,0)

(assuming t already exists as some time vector i.e. t=arange(-1,5,0.01))

Since where is built in, you could argue that step is built in as well :0

Ryan

On 3/20/07, David Warde-Farley <david.warde.farley at utoronto.ca> wrote:
> On Tue, 2007-03-20 at 03:45 -0400, Anne Archibald wrote:
>
> > In [7]: def step(x):
> >    ...:     return asarray(x>=0,dtype=float)
> >    ...:
> >
> > In [8]: step(linspace(-1,1,9))
> > Out[8]: array([ 0.,  0.,  0.,  0.,  1.,  1.,  1.,  1.,  1.])
> >
> > Though if you prefer step(0)==0.5 it's a bit of a pain.
>
> For the sake of completeness and utter boredom, here's the vectorized
> version with step(0) == 0.5, and it's still a one-liner ;)
>
> def step(x):
>     return asarray(x>0,dtype=float)+0.5*asarray(x==0,dtype=float)
>
> In [9]: step(arange(-5,5))
> Out[9]: array([ 0. ,  0. ,  0. ,  0. ,  0. ,  0.5,  1. ,  1. ,  1. ,
> 1. ])
>
> Cheers,
>
> David
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From bruno.chazelas at ias.u-psud.fr  Tue Mar 20 08:52:10 2007
From: bruno.chazelas at ias.u-psud.fr (bruno)
Date: Tue, 20 Mar 2007 13:52:10 +0100
Subject: [SciPy-user] animate solitons
In-Reply-To: <45FFD427.3090900@aims.ac.za>
References: <45FFD427.3090900@aims.ac.za>
Message-ID: <45FFD8FA.2080301@ias.u-psud.fr>

Hello,

If you look for graphic animation you can do that with matplotlib

http://www.scipy.org/Cookbook/Matplotlib/Animations

Bruno

Franck Kalala Mutombo a ?crit :
> hi,
>
> I have solved the sine-cordon equation(which is a partial differential
> equation) and I found my solitons(the solution to that equation), it is
> easy to plot it. but want I need is to animate it in a certain range. do
> you know any python module which can allows me to animate my solitons?
>
> ps: solitons a solution of a partial differential, it is a plane wave
> moving a certain speed, so I want to animate it as is moving.
>
>
>   



From takanobu.amano at gmail.com  Tue Mar 20 09:06:15 2007
From: takanobu.amano at gmail.com (Takanobu Amano)
Date: Tue, 20 Mar 2007 22:06:15 +0900
Subject: [SciPy-user] wavevector arrays
In-Reply-To: <9F48FF87-3B7A-4763-9186-68AC32AF577A@amnh.org>
References: <mailman.9.1174323608.22841.scipy-user@scipy.org>
	<9F48FF87-3B7A-4763-9186-68AC32AF577A@amnh.org>
Message-ID: <d56209900703200606w14274ebbx9a2bbffb3e2edbd@mail.gmail.com>

fftpack.fftshift may be useful for you.

kx_, ky_, kz_ = mgrid[-nx/2:+nx/2+1,-ny/2:+ny/2+1,-nz/2:+nz/2+1]
kx = fftpack.fftshift(kx_)
ky = fftpack.fftshift(ky_)
kz = fftpack.fftshift(kz_)

I'm not sure if this is what you need.

Takanobu

2007/3/20, J Oishi <joishi at amnh.org>:
> Hi,
>
> I am performing a number of 3D FFTs using scipy, and I had a question
> as to how to setup 3 3D arrays of wavevectors, much like one might do
> with mgrid. However, because the FFT returns wavevectors running from
> 0,...,k_c, -k_c, ..., -1 (where k_c is the nyquist wavenumber) I'm
> not sure how to get mgrid to do this.
>
> In 2D, I could create two 1D kx and ky arrays and use meshgrid (for
> an even number of points):
>
> import numpy as N
>
> kx = 2*N.pi/Lx * N.concatenate(N.arange(0,nx/2-1),N.arange(-nx/2, 0))
> ky = 2*N.pi/Ly * N.concatenate(N.arange(0,ny/2-1),N.arange(-ny/2, 0))
>
> kk = meshgrid(kx,ky)
>
> where kk would have a shape (2,ny,nx). However, this (to my
> knowledge) doesn't generalize to 3D.
>
> I'm a total python/scipy neophyte, so any help you could provide
> would be quite appreciated.
>
> thanks,
>
> jeff
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From eike.welk at gmx.net  Tue Mar 20 12:54:45 2007
From: eike.welk at gmx.net (Eike Welk)
Date: Tue, 20 Mar 2007 17:54:45 +0100
Subject: [SciPy-user] Next neighbor interpolation inside regular gridded
 arrays
In-Reply-To: <45FE6B83.10902@lcrs.de>
References: <45FE6B83.10902@lcrs.de>
Message-ID: <200703201754.45322.eike.welk@gmx.net>

The Matplotlib cookbook mentions two libs that might be usefull for 
you:
http://www.scipy.org/Cookbook/Matplotlib/Gridding_irregularly_spaced_data

Regards Eike.



From cookedm at physics.mcmaster.ca  Tue Mar 20 13:56:38 2007
From: cookedm at physics.mcmaster.ca (David M. Cooke)
Date: Tue, 20 Mar 2007 13:56:38 -0400
Subject: [SciPy-user] building numpy/scipy on Solaris
In-Reply-To: <8feab0f2.b0f28fea@steria.com>
References: <8feab0f2.b0f28fea@steria.com>
Message-ID: <20070320175638.GA29831@arbutus.physics.mcmaster.ca>

On Tue, Mar 20, 2007 at 12:08:14PM +0100, raphael langella wrote:
> I'm trying to build numpy and scipy on Solaris 8.
> The BLAS FAQ on netlib.org suggests using optimized BLAS librairies
> provided by computer vendor, like the SUN Performance Library. This
> library is supposed to provide enhanced and optimized version of BLAS
> and LAPACK. I happen to have Forte 7 installed, so I first tried to
> build against this library (libsunperf.a).
> I tried several versions and different compilation options, but I always
> get undefined symbols. Is this supported? As anyone ever succeeded in
> using this library to compile scipy and numpy?

Haven't heard about anybody else trying it. What are the undefined
symbols? One thing I can think of that I've run into before on a Alpha
was that the Compaq Math Library only had LAPACK version 2, not version
3 (which had been already been out for several years when our group got
the Alpha...).

> I also did some unsuccessful tests with Atlas. But, I think I'm now very
> close to have it working with the standard BLAS/LAPACK (from
> netlib.org). I have to use GNU ld (unresolved symbols when using Sun's
> ld). Numpy builds and tests without any error. I've got a few errors and
> failures with scipy (see attachment for the complete test report). Are
> these errors critical? I don't understand where they come from and how
> to correct them.
> Thanks for your help.
> 
> Rapha?l Langella

> calc-gen3-ci:/home/user1/ctcils/poladmin/rla/root $ python -c "import scipy; scipy.test(level=1)"
>   Found 1 tests for scipy.cluster.vq
>   Found 128 tests for scipy.linalg.fblas
>   Found 397 tests for scipy.ndimage
>   Found 10 tests for scipy.integrate.quadpack
>   Found 98 tests for scipy.stats.stats
>   Found 53 tests for scipy.linalg.decomp
>   Found 3 tests for scipy.integrate.quadrature
>   Found 96 tests for scipy.sparse.sparse
>   Found 20 tests for scipy.fftpack.pseudo_diffs
>   Found 6 tests for scipy.optimize.optimize
>   Found 6 tests for scipy.interpolate.fitpack
>   Found 6 tests for scipy.interpolate
>   Found 70 tests for scipy.stats.distributions
>   Found 10 tests for scipy.stats.morestats
>   Found 4 tests for scipy.linalg.lapack
>   Found 18 tests for scipy.fftpack.basic
> Warning: FAILURE importing tests for <module 'scipy.linsolve.umfpack' from '.../linsolve/umfpack/__init__.pyc'>
> /home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/linsolve/umfpack/tests/test_umfpack.py:17: AttributeError: 'module' object has no attribute 'umfpack' (in ?)

UMFPACK not built?

> Warning: FAILURE importing tests for <module 'scipy.optimize.zeros' from '...kages/scipy/optimize/zeros.pyc'>
> /home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/optimize/lbfgsb.py:30: ImportError: ld.so.1: python: fatal: relocation error: file /home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/optimize/_lbfgsb.so: symbol etime_: referenced symbol not found (in ?)

This one looks like the Sun Fortran compiler doesn't have etime. Not
serious unless you use optimize.fmin_l_bfgs_b.

> Warning: FAILURE importing tests for <module 'scipy.io.mio' from '...site-packages/scipy/io/mio.pyc'>
> /home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/io/tests/test_mio.py:14: ImportError: cannot import name loadmat (in ?)

Odd. loadmat comes from scipy.io.mio, and it's a pure python routine.

Did you remove the build/ directory before trying again? Distutils can
get confused if it's interrupted (it doesn't do dependency tracking that
well).

> Warning: The maximum number of subdivisions (50) has been achieved.
>   If increasing the limit yields no improvement it is advised to analyze
>   the integrand in order to determine the difficulties.  If the position of a
>   local difficulty can be determined (singularity, discontinuity) one will
>   probably gain from splitting up the interval and calling the integrator
>   on the subranges.  Perhaps a special-purpose integrator should be used.
[clip]

I don't know why this is happening; it doesn't occur on other platforms.

> ****************************************************************
> WARNING: clapack module is empty
> -----------
> See scipy/INSTALL.txt for troubleshooting.
> Notes:
> * If atlas library is not found by numpy/distutils/system_info.py,
>   then scipy uses flapack instead of clapack.
> ****************************************************************

This is OK.

> ======================================================================
> ERROR: check_normalize (scipy.sparse.tests.test_sparse.test_coo)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/sparse/tests/test_sparse.py", line 658, in check_normalize
>     assert(zip(nrow,ncol,ndata) == sorted(zip(row,col,data))) #should sort by rows, then cols
> NameError: global name 'sorted' is not defined

Fixed in svn; a 2.4ism snuck in.

> ======================================================================
> FAIL: check_double_integral (scipy.integrate.tests.test_quadpack.test_quad)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/integrate/tests/test_quadpack.py", line 95, in check_double_integral
>     5/6.0 * (b**3.0-a**3.0))
>   File "/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/integrate/tests/test_quadpack.py", line 9, in assert_quad
>     assert abs(value-tabledValue) < err, (value, tabledValue, err)
> AssertionError: (4036.4928418995069, 5.8333333333333339, 1835.6069980919819)
> 
> ======================================================================
> FAIL: check_triple_integral (scipy.integrate.tests.test_quadpack.test_quad)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/integrate/tests/test_quadpack.py", line 106, in check_triple_integral
>     8/3.0 * (b**4.0 - a**4.0))
>   File "/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/integrate/tests/test_quadpack.py", line 9, in assert_quad
>     assert abs(value-tabledValue) < err, (value, tabledValue, err)
> AssertionError: (8.6265941452283596e-09, 40.0, 3.0164375235041808e-09)

These look like the same problem as the warnings before about
subdivisions. Something funny.

> ======================================================================
> FAIL: check_syevr (scipy.lib.tests.test_lapack.test_flapack_float)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/lib/lapack/tests/esv_tests.py", line 41, in check_syevr
>     assert_array_almost_equal(w,exact_w)
>   File "/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/numpy/testing/utils.py", line 230, in assert_array_almost_equal
>     header='Arrays are not almost equal')
>   File "/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/numpy/testing/utils.py", line 215, in assert_array_compare
>     assert cond, msg
> AssertionError:
> Arrays are not almost equal
> 
> (mismatch 33.3333333333%)
>  x: array([-0.66992444,  0.48769474,  9.18222618], dtype=float32)
>  y: array([-0.66992434,  0.48769389,  9.18223045])
> ======================================================================
> FAIL: check_syevr_irange (scipy.lib.tests.test_lapack.test_flapack_float)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/lib/lapack/tests/esv_tests.py", line 66, in check_syevr_irange
>     assert_array_almost_equal(w,exact_w[rslice])
>   File "/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/numpy/testing/utils.py", line 230, in assert_array_almost_equal
>     header='Arrays are not almost equal')
>   File "/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/numpy/testing/utils.py", line 215, in assert_array_compare
>     assert cond, msg
> AssertionError:
> Arrays are not almost equal
> 
> (mismatch 33.3333333333%)
>  x: array([-0.66992444,  0.48769474,  9.18222618], dtype=float32)
>  y: array([-0.66992434,  0.48769389,  9.18223045])

Huh, I'd say the Sun library isn't accurate enough. Sacrificing accuracy
for speed, no doubt.

> ----------------------------------------------------------------------
> Ran 1543 tests in 12.138s
> 
> FAILED (failures=4, errors=2)
> calc-gen3-ci:/home/user1/ctcils/poladmin/rla/root $

-- 
|>|\/|<
/--------------------------------------------------------------------------\
|David M. Cooke                      http://arbutus.physics.mcmaster.ca/dmc/
|cookedm at physics.mcmaster.ca


From emilia12 at mail.bg  Tue Mar 20 15:03:54 2007
From: emilia12 at mail.bg (emilia12 at mail.bg)
Date: Tue, 20 Mar 2007 21:03:54 +0200
Subject: [SciPy-user] problem with the determinant of matrix
Message-ID: <1174417434.e88e400550ce2@mail.bg>


Hi list,
i have a problem with the determinant of matrix. Following
the example ("http://www.scipy.org/SciPy_Tutorial"):
#code
from numpy import matrix
from scipy.linalg import inv, det, eig

A=matrix([[1,1,1],[4,4,3],[7,8,5]]) # 3 lines 3 rows
b = matrix([1,2,1]).transpose()     # 3 lines 1 rows.

print det(A)     # We can check, whether the matrix is
regular
#/code

the python crashes with "Exit code: -1073741795"

versions:
Python 2.4.4 (#71, Oct 18 2006, 08:34:43) [MSC v.1310 32 bit
(Intel)] on win32
NumPy 1.0.1 for Python 2.4
SciPy 0.5.2 for Python 2.4

so, any ideas why this happens?

e.





-----------------------------

SCENA - ???????????? ????????? ???????? ?? ??????? ??????????? ? ??????????.
http://www.bgscena.com/



From aisaac at american.edu  Tue Mar 20 15:44:25 2007
From: aisaac at american.edu (Alan G Isaac)
Date: Tue, 20 Mar 2007 15:44:25 -0400
Subject: [SciPy-user] problem with the determinant of matrix
In-Reply-To: <1174417434.e88e400550ce2@mail.bg>
References: <1174417434.e88e400550ce2@mail.bg>
Message-ID: <Mahogany-0.67.0-1624-20070320-154425.00@american.edu>

On Tue, 20 Mar 2007, emilia12 at mail.bg apparently wrote: 

> i have a problem with the determinant of matrix. 


Python 2.5 (r25:51908, Sep 19 2006, 09:52:17) [MSC v.1310 32 
bit (Intel)] on win 32
Type "help", "copyright", "credits" or "license" for more information.
>>> from numpy import matrix
>>> from numpy.linalg import det
>>> A=matrix([[1,1,1],[4,4,3],[7,8,5]])
>>> det(A)
0.99999999999999956
>>>

fwiw,
Alan Isaac





From peridot.faceted at gmail.com  Tue Mar 20 16:54:20 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Tue, 20 Mar 2007 16:54:20 -0400
Subject: [SciPy-user] I can't find unit step function.
In-Reply-To: <c5b438120703200532n49fe0d71g74519593eddaeab4@mail.gmail.com>
References: <296323b50703200020m2a33d3d8kc0bd0f0c8a9cac3e@mail.gmail.com>
	<1174376285.28167.6.camel@rodimus>
	<ce557a360703200045t4d2a6a2fnca1197a4df76153b@mail.gmail.com>
	<1174379374.5075.5.camel@rodimus>
	<c5b438120703200532n49fe0d71g74519593eddaeab4@mail.gmail.com>
Message-ID: <ce557a360703201354p164963b7v9d623d03033df8d2@mail.gmail.com>

On 20/03/07, Ryan Krauss <ryanlists at gmail.com> wrote:
> Sorry to contribute to this discussion out of boredom and curiosity,
> but what about:
>
> y=where(t>0,1.0,0)

Or
y=(1+sign(t))/2.
for conciseness and the 0.5 both...

Anne


From david.warde.farley at utoronto.ca  Tue Mar 20 16:57:14 2007
From: david.warde.farley at utoronto.ca (David Warde-Farley)
Date: Tue, 20 Mar 2007 16:57:14 -0400
Subject: [SciPy-user] I can't find unit step function.
In-Reply-To: <ce557a360703201354p164963b7v9d623d03033df8d2@mail.gmail.com>
References: <296323b50703200020m2a33d3d8kc0bd0f0c8a9cac3e@mail.gmail.com>
	<1174376285.28167.6.camel@rodimus>
	<ce557a360703200045t4d2a6a2fnca1197a4df76153b@mail.gmail.com>
	<1174379374.5075.5.camel@rodimus>
	<c5b438120703200532n49fe0d71g74519593eddaeab4@mail.gmail.com>
	<ce557a360703201354p164963b7v9d623d03033df8d2@mail.gmail.com>
Message-ID: <86BC2FBA-BEAC-4571-846D-B46A90A6A6C1@utoronto.ca>

On 20-Mar-07, at 4:54 PM, Anne Archibald wrote:

> On 20/03/07, Ryan Krauss <ryanlists at gmail.com> wrote:
>> Sorry to contribute to this discussion out of boredom and curiosity,
>> but what about:
>>
>> y=where(t>0,1.0,0)
>
> Or
> y=(1+sign(t))/2.
> for conciseness and the 0.5 both...
>
> Anne

Bravo! I think we can consider this dead horse adequately flogged ;)

David


From ckkart at hoc.net  Tue Mar 20 23:15:12 2007
From: ckkart at hoc.net (Christian)
Date: Wed, 21 Mar 2007 12:15:12 +0900
Subject: [SciPy-user] odr thread safe?
Message-ID: <etq7vp$ldc$1@sea.gmane.org>

Hi,
I noticed that I cannot run odr more than once at a time using python threads.
Is that true? Debugging yields:

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1294730336 (LWP 9312)]
0xb6fe0934 in fcn_callback (n=0xb2d3e9ec, m=0xb2d3e9e8, np=0xb2d3e9e4,
nq=0xb2d3e9e0,
    ldn=0xb2d3e9ec, ldm=0xb2d3e9e8, ldnp=0xb2d3e9e4, beta=0x90e0350,
xplusd=0xb19b4e58,
    ifixb=0x90e6580, ifixx=0x90e65d8, ldfix=0xb2d3e9c4, ideval=0xb701a098,
f=0xb19d76e8,
    fjacb=0xb19d8ec0, fjacd=0xb19b2008, istop=0xb2d3e288) at
./peak_o_mat/odr/__odrpack.c:70
70        Py_INCREF(odr_global.pyBeta);

Can someone advise me how to change the wrapper to be thread safe?

Christian



From raphael.langella at steria.com  Wed Mar 21 04:59:04 2007
From: raphael.langella at steria.com (raphael langella)
Date: Wed, 21 Mar 2007 09:59:04 +0100
Subject: [SciPy-user] building numpy/scipy on Solaris
Message-ID: <96b8d3f1.d3f196b8@steria.com>

---- Messages d?origine ----
De: "David M. Cooke" <cookedm at physics.mcmaster.ca>
Date: mardi, mars 20, 2007 6:56 pm
Objet: Re: [SciPy-user] building numpy/scipy on Solaris

> On Tue, Mar 20, 2007 at 12:08:14PM +0100, raphael langella wrote:
> > I'm trying to build numpy and scipy on Solaris 8.
> > The BLAS FAQ on netlib.org suggests using optimized BLAS librairies
> > provided by computer vendor, like the SUN Performance Library. This
> > library is supposed to provide enhanced and optimized version of 
> BLAS> and LAPACK. I happen to have Forte 7 installed, so I first 
> tried to
> > build against this library (libsunperf.a).
> > I tried several versions and different compilation options, but I 
> always> get undefined symbols. Is this supported? As anyone ever 
> succeeded in
> > using this library to compile scipy and numpy?
> 
> Haven't heard about anybody else trying it. What are the undefined
> symbols? One thing I can think of that I've run into before on a Alpha
> was that the Compaq Math Library only had LAPACK version 2, not 
> version3 (which had been already been out for several years when 
> our group got
> the Alpha...).

It's supposed to support LAPACK v3.0 and BLAS1, 2 & 3
(http://developers.sun.com/sunstudio/perflib_index.html).
It gives me this when I import numpy :
ImportError: ld.so.1: python: fatal: relocation error: file
/usr/lib/python2.3/site-packages/numpy/linalg/lapack_lite.so: symbol
__getenv_: referenced symbol not found

 
> > /home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-
> packages/scipy/linsolve/umfpack/tests/test_umfpack.py:17:
AttributeError: 'module' object has no attribute 'umfpack' (in ?)
> 
> UMFPACK not built?

right. I'm not sure if my users will need it, but I'll try building it
anyway.

> > Warning: FAILURE importing tests for <module 
> 'scipy.optimize.zeros' from '...kages/scipy/optimize/zeros.pyc'>
> > /home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-
> packages/scipy/optimize/lbfgsb.py:30: ImportError: ld.so.1: python: 
> fatal: relocation error: file 
>
/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/scipy/optimize/_lbfgsb.so:
symbol etime_: referenced symbol not found (in ?)
> 
> This one looks like the Sun Fortran compiler doesn't have etime. Not
> serious unless you use optimize.fmin_l_bfgs_b.

Well, let's say it's not critical for now...
> 
> > Warning: FAILURE importing tests for <module 'scipy.io.mio' from 
> '...site-packages/scipy/io/mio.pyc'>
> > /home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-
> packages/scipy/io/tests/test_mio.py:14: ImportError: cannot import 
> name loadmat (in ?)
> 
> Odd. loadmat comes from scipy.io.mio, and it's a pure python routine.
> 
> Did you remove the build/ directory before trying again? Distutils can
> get confused if it's interrupted (it doesn't do dependency tracking 
> thatwell).

Yes I did.
> > NameError: global name 'sorted' is not defined
> 
> Fixed in svn; a 2.4ism snuck in.

I'll try it.


> Huh, I'd say the Sun library isn't accurate enough. Sacrificing 
> accuracyfor speed, no doubt.

OK, but I didn't compile scipy and numpy against libsunperf, I used
netlib.org versions of blas and lapack. On the other hand, I did build
lapack against libsunperf (that's the default in make.inc.SUN4SOL2), so
I'm not sure which blas library is used in the end.

Thanks a lot for your support


From cookedm at physics.mcmaster.ca  Wed Mar 21 06:23:03 2007
From: cookedm at physics.mcmaster.ca (David M. Cooke)
Date: Wed, 21 Mar 2007 06:23:03 -0400
Subject: [SciPy-user] building numpy/scipy on Solaris
In-Reply-To: <96b8d3f1.d3f196b8@steria.com>
References: <96b8d3f1.d3f196b8@steria.com>
Message-ID: <20070321102303.GA32399@arbutus.physics.mcmaster.ca>

On Wed, Mar 21, 2007 at 09:59:04AM +0100, raphael langella wrote:
> > On Tue, Mar 20, 2007 at 12:08:14PM +0100, raphael langella wrote:
> > > I'm trying to build numpy and scipy on Solaris 8.
> > > The BLAS FAQ on netlib.org suggests using optimized BLAS librairies
> > > provided by computer vendor, like the SUN Performance Library. This
> > > library is supposed to provide enhanced and optimized version of 
> > BLAS> and LAPACK. I happen to have Forte 7 installed, so I first 
> > tried to
> > > build against this library (libsunperf.a).
> 
> It's supposed to support LAPACK v3.0 and BLAS1, 2 & 3
> (http://developers.sun.com/sunstudio/perflib_index.html).
> It gives me this when I import numpy :
> ImportError: ld.so.1: python: fatal: relocation error: file
> /usr/lib/python2.3/site-packages/numpy/linalg/lapack_lite.so: symbol
> __getenv_: referenced symbol not found

Odd. lapack_lite doesn't use getenv anywhere, so it must be the sunperf
library (and http://docs.sun.com/source/819-3692/plug_optimizing.html
shows that it does look at environment variables). I'm guessing there's
some link-time option that it needs.

> > > Warning: FAILURE importing tests for <module 'scipy.io.mio' from 
> > '...site-packages/scipy/io/mio.pyc'>
> > > /home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-
> > packages/scipy/io/tests/test_mio.py:14: ImportError: cannot import 
> > name loadmat (in ?)
> > 
> > Odd. loadmat comes from scipy.io.mio, and it's a pure python routine.
> > 
> > Did you remove the build/ directory before trying again? Distutils can
> > get confused if it's interrupted (it doesn't do dependency tracking 
> > thatwell).
> 
> Yes I did.
> 
> 
> > Huh, I'd say the Sun library isn't accurate enough. Sacrificing 
> > accuracyfor speed, no doubt.
> 
> OK, but I didn't compile scipy and numpy against libsunperf, I used
> netlib.org versions of blas and lapack. On the other hand, I did build
> lapack against libsunperf (that's the default in make.inc.SUN4SOL2), so
> I'm not sure which blas library is used in the end.

Hmm, then I'm going to blame the compiler :) Maybe it's doing some
optimization that it shouldn't? -fast, maybe?

-- 
|>|\/|<
/--------------------------------------------------------------------------\
|David M. Cooke                      http://arbutus.physics.mcmaster.ca/dmc/
|cookedm at physics.mcmaster.ca


From gruben at bigpond.net.au  Wed Mar 21 08:17:01 2007
From: gruben at bigpond.net.au (Gary Ruben)
Date: Wed, 21 Mar 2007 23:17:01 +1100
Subject: [SciPy-user] animate solitons
In-Reply-To: <45FFD427.3090900@aims.ac.za>
References: <45FFD427.3090900@aims.ac.za>
Message-ID: <4601223D.5080305@bigpond.net.au>

If it is in 1D, take a look at vpython. You can create a curve and as 
you modify it, vpython will take care of the animation for you.

Gary R.

Franck Kalala Mutombo wrote:
> hi,
> 
> I have solved the sine-cordon equation(which is a partial differential
> equation) and I found my solitons(the solution to that equation), it is
> easy to plot it. but want I need is to animate it in a certain range. do
> you know any python module which can allows me to animate my solitons?
> 
> ps: solitons a solution of a partial differential, it is a plane wave
> moving a certain speed, so I want to animate it as is moving.


From ryanlists at gmail.com  Wed Mar 21 08:51:58 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Wed, 21 Mar 2007 07:51:58 -0500
Subject: [SciPy-user] animate solitons
In-Reply-To: <4601223D.5080305@bigpond.net.au>
References: <45FFD427.3090900@aims.ac.za> <4601223D.5080305@bigpond.net.au>
Message-ID: <c5b438120703210551r5b3dd18dw1a79128b7baba938@mail.gmail.com>

I have some success using Mayavi to animate solutions of PDE's for the
mode shapes of a beam.

On 3/21/07, Gary Ruben <gruben at bigpond.net.au> wrote:
> If it is in 1D, take a look at vpython. You can create a curve and as
> you modify it, vpython will take care of the animation for you.
>
> Gary R.
>
> Franck Kalala Mutombo wrote:
> > hi,
> >
> > I have solved the sine-cordon equation(which is a partial differential
> > equation) and I found my solitons(the solution to that equation), it is
> > easy to plot it. but want I need is to animate it in a certain range. do
> > you know any python module which can allows me to animate my solitons?
> >
> > ps: solitons a solution of a partial differential, it is a plane wave
> > moving a certain speed, so I want to animate it as is moving.
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From raphael.langella at steria.com  Wed Mar 21 09:27:56 2007
From: raphael.langella at steria.com (raphael langella)
Date: Wed, 21 Mar 2007 14:27:56 +0100
Subject: [SciPy-user] building numpy/scipy on Solaris
Message-ID: <eec410113.10113eec4@steria.com>

---- Messages d?origine ----
De: "David M. Cooke" <cookedm at physics.mcmaster.ca>
Date: mercredi, mars 21, 2007 11:23 am
Objet: Re: [SciPy-user] building numpy/scipy on Solaris

> On Wed, Mar 21, 2007 at 09:59:04AM +0100, raphael langella wrote:
> > > On Tue, Mar 20, 2007 at 12:08:14PM +0100, raphael langella wrote:
> > > > I'm trying to build numpy and scipy on Solaris 8.
> > > > The BLAS FAQ on netlib.org suggests using optimized BLAS 
> librairies> > > provided by computer vendor, like the SUN 
> Performance Library. This
> > > > library is supposed to provide enhanced and optimized version 
> of 
> > > BLAS> and LAPACK. I happen to have Forte 7 installed, so I 
> first 
> > > tried to
> > > > build against this library (libsunperf.a).
> > 
> > It's supposed to support LAPACK v3.0 and BLAS1, 2 & 3
> > (http://developers.sun.com/sunstudio/perflib_index.html).
> > It gives me this when I import numpy :
> > ImportError: ld.so.1: python: fatal: relocation error: file
> > /usr/lib/python2.3/site-packages/numpy/linalg/lapack_lite.so: symbol
> > __getenv_: referenced symbol not found
> 
> Odd. lapack_lite doesn't use getenv anywhere, so it must be the 
> sunperflibrary (and http://docs.sun.com/source/819-
> 3692/plug_optimizing.htmlshows that it does look at environment 
> variables). I'm guessing there's
> some link-time option that it needs.

ok. I'll forget about the other errors I have and will focus on
libsunperf for now. It makes sense to me to use vendor provided
optimized BLAS/LAPACK libraries.

I compiled with these options :
python setup.py config_fc --f77flags='-xlic_lib=sunperf -xarch=v9b' install

and now, I've got this error when importing numpy :
ImportError: ld.so.1: python: fatal: relocation error: file
/usr/local/lib/python2.3/site-packages/numpy/linalg/lapack_lite.so:
symbol dgeev_: referenced symbol not found

Any idea?



From bryanv at enthought.com  Wed Mar 21 11:54:36 2007
From: bryanv at enthought.com (Bryan Van de Ven)
Date: Wed, 21 Mar 2007 10:54:36 -0500
Subject: [SciPy-user] animate solitons
In-Reply-To: <c5b438120703210551r5b3dd18dw1a79128b7baba938@mail.gmail.com>
References: <45FFD427.3090900@aims.ac.za> <4601223D.5080305@bigpond.net.au>
	<c5b438120703210551r5b3dd18dw1a79128b7baba938@mail.gmail.com>
Message-ID: <4601553C.7030809@enthought.com>

You can also do this sort of thing in Chaco if you only need 2D capabilities 
like line/scatter or colormap plots. Here is a simple example of a simulated 
updating data stream:

https://svn.enthought.com/enthought/browser/trunk/src/lib/enthought/chaco2/examples/advanced/data_stream.py

Ryan Krauss wrote:
> I have some success using Mayavi to animate solutions of PDE's for the
> mode shapes of a beam.
> 
> On 3/21/07, Gary Ruben <gruben at bigpond.net.au> wrote:
>> If it is in 1D, take a look at vpython. You can create a curve and as
>> you modify it, vpython will take care of the animation for you.
>>
>> Gary R.
>>
>> Franck Kalala Mutombo wrote:
>>> hi,
>>>
>>> I have solved the sine-cordon equation(which is a partial differential
>>> equation) and I found my solitons(the solution to that equation), it is
>>> easy to plot it. but want I need is to animate it in a certain range. do
>>> you know any python module which can allows me to animate my solitons?
>>>
>>> ps: solitons a solution of a partial differential, it is a plane wave
>>> moving a certain speed, so I want to animate it as is moving.
>> _______________________________________________
>> SciPy-user mailing list
>> SciPy-user at scipy.org
>> http://projects.scipy.org/mailman/listinfo/scipy-user
>>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
> 



From peridot.faceted at gmail.com  Wed Mar 21 12:40:00 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Wed, 21 Mar 2007 12:40:00 -0400
Subject: [SciPy-user] animate solitons
In-Reply-To: <4601553C.7030809@enthought.com>
References: <45FFD427.3090900@aims.ac.za> <4601223D.5080305@bigpond.net.au>
	<c5b438120703210551r5b3dd18dw1a79128b7baba938@mail.gmail.com>
	<4601553C.7030809@enthought.com>
Message-ID: <ce557a360703210940u75b288b3w2f782dd432d02346@mail.gmail.com>

Most of these solutions seem to be designed to allow animations to be
updated on the fly as new data comes in. Undoubtedly valuable, but
more sophisticated than one needs, often.  Is there a tidy package
that lets you take, say, a 2D array and treat one axis as x-values and
the other axis as time, and produce an animation? It would presumably
loop, and be user-controllable (pause, reverse, frame-forward,
frame-back, etc.)

Anne


From strawman at astraw.com  Wed Mar 21 12:58:41 2007
From: strawman at astraw.com (Andrew Straw)
Date: Wed, 21 Mar 2007 09:58:41 -0700
Subject: [SciPy-user] problem with the determinant of matrix
In-Reply-To: <1174417434.e88e400550ce2@mail.bg>
References: <1174417434.e88e400550ce2@mail.bg>
Message-ID: <46016441.5070208@astraw.com>

Could this be the issue that the SciPy build on the website was built 
with SSE2 support and you don't have SSE2? Do you have an older 
processor (other than Intel Pentium 3 or AMD something-or-other)? Not 
using Windows much, I don't keep track of the details, but there 
historically has been the occasional issue along these lines, and I'm 
not sure it's been addressed.

-Andrew

emilia12 at mail.bg wrote:
> Hi list,
> i have a problem with the determinant of matrix. Following
> the example ("http://www.scipy.org/SciPy_Tutorial"):
> #code
> from numpy import matrix
> from scipy.linalg import inv, det, eig
>
> A=matrix([[1,1,1],[4,4,3],[7,8,5]]) # 3 lines 3 rows
> b = matrix([1,2,1]).transpose()     # 3 lines 1 rows.
>
> print det(A)     # We can check, whether the matrix is
> regular
> #/code
>
> the python crashes with "Exit code: -1073741795"
>
> versions:
> Python 2.4.4 (#71, Oct 18 2006, 08:34:43) [MSC v.1310 32 bit
> (Intel)] on win32
> NumPy 1.0.1 for Python 2.4
> SciPy 0.5.2 for Python 2.4
>
> so, any ideas why this happens?
>
> e.
>
>
>
>
>
> -----------------------------
>
> SCENA - ???????????? ????????? ???????? ?? ??????? ??????????? ? ??????????.
> http://www.bgscena.com/
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   




From strawman at astraw.com  Wed Mar 21 13:01:46 2007
From: strawman at astraw.com (Andrew Straw)
Date: Wed, 21 Mar 2007 10:01:46 -0700
Subject: [SciPy-user] problem with the determinant of matrix
In-Reply-To: <46016441.5070208@astraw.com>
References: <1174417434.e88e400550ce2@mail.bg> <46016441.5070208@astraw.com>
Message-ID: <460164FA.8010101@astraw.com>

Andrew Straw wrote:
> Could this be the issue that the SciPy build on the website was built 
> with SSE2 support and you don't have SSE2? Do you have an older 
> processor (other than Intel Pentium 3 or AMD something-or-other)? Not 
>   
                          ^^^^^ older
> using Windows much, I don't keep track of the details, but there 
> historically has been the occasional issue along these lines, and I'm 
> not sure it's been addressed.
>
> -Andrew
>
> emilia12 at mail.bg wrote:
>   
>> Hi list,
>> i have a problem with the determinant of matrix. Following
>> the example ("http://www.scipy.org/SciPy_Tutorial"):
>> #code
>> from numpy import matrix
>> from scipy.linalg import inv, det, eig
>>
>> A=matrix([[1,1,1],[4,4,3],[7,8,5]]) # 3 lines 3 rows
>> b = matrix([1,2,1]).transpose()     # 3 lines 1 rows.
>>
>> print det(A)     # We can check, whether the matrix is
>> regular
>> #/code
>>
>> the python crashes with "Exit code: -1073741795"
>>
>> versions:
>> Python 2.4.4 (#71, Oct 18 2006, 08:34:43) [MSC v.1310 32 bit
>> (Intel)] on win32
>> NumPy 1.0.1 for Python 2.4
>> SciPy 0.5.2 for Python 2.4
>>
>> so, any ideas why this happens?
>>
>> e.
>>
>>
>>
>>
>>
>> -----------------------------
>>
>> SCENA - ???????????? ????????? ???????? ?? ??????? ??????????? ? ??????????.
>> http://www.bgscena.com/
>>
>> _______________________________________________
>> SciPy-user mailing list
>> SciPy-user at scipy.org
>> http://projects.scipy.org/mailman/listinfo/scipy-user
>>   
>>     
>
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   




From bryanv at enthought.com  Wed Mar 21 13:07:19 2007
From: bryanv at enthought.com (Bryan Van de Ven)
Date: Wed, 21 Mar 2007 12:07:19 -0500
Subject: [SciPy-user] animate solitons
In-Reply-To: <ce557a360703210940u75b288b3w2f782dd432d02346@mail.gmail.com>
References: <45FFD427.3090900@aims.ac.za>
	<4601223D.5080305@bigpond.net.au>	<c5b438120703210551r5b3dd18dw1a79128b7baba938@mail.gmail.com>	<4601553C.7030809@enthought.com>
	<ce557a360703210940u75b288b3w2f782dd432d02346@mail.gmail.com>
Message-ID: <46016647.5000401@enthought.com>

Hi Anne,

That chaco example I posted was made specifically in response to a question 
about realtime DAQ, so it does generate some random data "on the fly" to 
simulate a realtime data stream. However, it would be pretty trivial to have it, 
say, step through the rows of a preset 2d array each timer tick instead.

As for something straight out of the box, I'm not sure.. the only thing that 
comes to mind is VPython, possibly.


Anne Archibald wrote:
> Most of these solutions seem to be designed to allow animations to be
> updated on the fly as new data comes in. Undoubtedly valuable, but
> more sophisticated than one needs, often.  Is there a tidy package
> that lets you take, say, a 2D array and treat one axis as x-values and
> the other axis as time, and produce an animation? It would presumably
> loop, and be user-controllable (pause, reverse, frame-forward,
> frame-back, etc.)
> 
> Anne
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
> 



From pwang at enthought.com  Wed Mar 21 13:21:32 2007
From: pwang at enthought.com (Peter Wang)
Date: Wed, 21 Mar 2007 12:21:32 -0500
Subject: [SciPy-user] animate solitons
In-Reply-To: <46016647.5000401@enthought.com>
References: <45FFD427.3090900@aims.ac.za>
	<4601223D.5080305@bigpond.net.au>	<c5b438120703210551r5b3dd18dw1a79128b7baba938@mail.gmail.com>	<4601553C.7030809@enthought.com>
	<ce557a360703210940u75b288b3w2f782dd432d02346@mail.gmail.com>
	<46016647.5000401@enthought.com>
Message-ID: <9B72195F-2AC6-4F09-AB69-E80398163234@enthought.com>

On Mar 21, 2007, at 12:07 PM, Bryan Van de Ven wrote:

> That chaco example I posted was made specifically in response to a  
> question
> about realtime DAQ, so it does generate some random data "on the  
> fly" to
> simulate a realtime data stream. However, it would be pretty  
> trivial to have it,
> say, step through the rows of a preset 2d array each timer tick  
> instead.

Actually, the updating_plot demos in https://svn.enthought.com/ 
enthought/browser/trunk/src/lib/enthought/chaco2/examples/ 
updating_plot do precisely this - on each timer increment, they  
modify the plot's data to reflect a slightly different subsection of  
the total data.

> Is there a tidy package
> that lets you take, say, a 2D array and treat one axis as x-values and
> the other axis as time, and produce an animation?

You might also be interested in the scalar image function inspector:
https://svn.enthought.com/enthought/browser/trunk/src/lib/enthought/ 
chaco2/advanced/scalar_image_function_inspector

This interactively shows cross-cuts in X and Y through a 2D scalar  
data volume.  You can see a screenshot at:
http://code.enthought.com/chaco/gallery/ 
scalar_image_function_inspector.png

Much of the complexity in the code for this demo is in hooking up  
many of the interactive elements; if all that was desired was  
animation in a single dimension as the user moved a slider, then the  
code would be much simplified.


-peter


From robert.kern at gmail.com  Wed Mar 21 13:42:12 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 21 Mar 2007 12:42:12 -0500
Subject: [SciPy-user] odr thread safe?
In-Reply-To: <etq7vp$ldc$1@sea.gmane.org>
References: <etq7vp$ldc$1@sea.gmane.org>
Message-ID: <46016E74.9060309@gmail.com>

Christian wrote:
> Hi,
> I noticed that I cannot run odr more than once at a time using python threads.
> Is that true? Debugging yields:
> 
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread -1294730336 (LWP 9312)]
> 0xb6fe0934 in fcn_callback (n=0xb2d3e9ec, m=0xb2d3e9e8, np=0xb2d3e9e4,
> nq=0xb2d3e9e0,
>     ldn=0xb2d3e9ec, ldm=0xb2d3e9e8, ldnp=0xb2d3e9e4, beta=0x90e0350,
> xplusd=0xb19b4e58,
>     ifixb=0x90e6580, ifixx=0x90e65d8, ldfix=0xb2d3e9c4, ideval=0xb701a098,
> f=0xb19d76e8,
>     fjacb=0xb19d8ec0, fjacd=0xb19b2008, istop=0xb2d3e288) at
> ./peak_o_mat/odr/__odrpack.c:70
> 70        Py_INCREF(odr_global.pyBeta);
> 
> Can someone advise me how to change the wrapper to be thread safe?

Hmm. I'm not sure how this is happening. Calls to extension functions should be
holding the global interpreter lock unless if it is explicitly released, and I'm
pretty sure that I'm not. It's possible that the GIL gets released when the
Python callbacks execute; I'd have to check. In that case, we would need a lock
around the call to __odrpack.odr().

But yeah, there's a difficult problem with calling a FORTRAN subroutine multiple
concurrent times. We need to pass function pointer callbacks to the FORTRAN
subroutine. These need to be real C functions. They can't have state, so the
state needs to be global. C++ methods don't work. Consequently, the Python
functions that the C callbacks will actually call are stored in odr_global.

Looking at the Python C API docs, a possibility is to use
PyThreadState_GetDict() to get a thread-specific dictionary. If threads aren't
being used, then you only get NULL, though, so another mechanism will have to
exist for the unthreaded case. I think. I could be misinterpreting the
documentation.

  http://docs.python.org/api/threads.html

Let me know what you come up with!

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From emilia12 at mail.bg  Wed Mar 21 14:29:00 2007
From: emilia12 at mail.bg (emilia12 at mail.bg)
Date: Wed, 21 Mar 2007 20:29:00 +0200
Subject: [SciPy-user] problem with the determinant of matrix
In-Reply-To: <460164FA.8010101@astraw.com>
References: <1174417434.e88e400550ce2@mail.bg>
	<46016441.5070208@astraw.com> <460164FA.8010101@astraw.com>
Message-ID: <1174501740.6d53e1d4a465a@mail.bg>

Hi Andrew,

my computer is PIII on 1000MHz but i tried the same at my
work (PIV 2core ... etc) and it didn't crash but it
reported the following error:
RuntimeError: module compiled against version 1000002 of
C-API but this version of numpy is 1000009
RuntimeError: module compiled against version 1000002 of
C-API but this version of numpy is 1000009

and then I changed the 2dn line:
from scipy.linalg import inv, det, eig
with:
import numpy

and:
print numpy.linalg.det(A)
works OK

so probably there is a problem with numpy and scipi ?


regards
e.


????? ?? ????? ?? Andrew Straw <strawman at astraw.com>:

> Andrew Straw wrote:
> > Could this be the issue that the SciPy build on the
> website was built
> > with SSE2 support and you don't have SSE2? Do you have
> an older
> > processor (other than Intel Pentium 3 or AMD
> something-or-other)? Not
> >
>                           ^^^^^ older
> > using Windows much, I don't keep track of the details,
> but there
> > historically has been the occasional issue along these
> lines, and I'm
> > not sure it's been addressed.
> >
> > -Andrew
> >
> > emilia12 at mail.bg wrote:
> >
> >> Hi list,
> >> i have a problem with the determinant of matrix.
> Following
> >> the example ("http://www.scipy.org/SciPy_Tutorial"):
> >> #code
> >> from numpy import matrix
> >> from scipy.linalg import inv, det, eig
> >>
> >> A=matrix([[1,1,1],[4,4,3],[7,8,5]]) # 3 lines 3 rows
> >> b = matrix([1,2,1]).transpose()     # 3 lines 1 rows.
> >>
> >> print det(A)     # We can check, whether the matrix is
> >> regular
> >> #/code
> >>
> >> the python crashes with "Exit code: -1073741795"
> >>
> >> versions:
> >> Python 2.4.4 (#71, Oct 18 2006, 08:34:43) [MSC v.1310
> 32 bit
> >> (Intel)] on win32
> >> NumPy 1.0.1 for Python 2.4
> >> SciPy 0.5.2 for Python 2.4
> >>
> >> so, any ideas why this happens?
> >>
> >> e.
> >>
> >>
> >>
> >>
> >>
> >> -----------------------------
> >>
> >> SCENA - ???????????? ????????? ???????? ?? ???????
> ??????????? ? ??????????.
> >> http://www.bgscena.com/
> >>
> >> _______________________________________________
> >> SciPy-user mailing list
> >> SciPy-user at scipy.org
> >> http://projects.scipy.org/mailman/listinfo/scipy-user
> >>
> >>
> >
> >
> > _______________________________________________
> > SciPy-user mailing list
> > SciPy-user at scipy.org
> > http://projects.scipy.org/mailman/listinfo/scipy-user
> >
>
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>




-----------------------------

SCENA - ???????????? ????????? ???????? ?? ??????? ??????????? ? ??????????.
http://www.bgscena.com/



From oliphant at ee.byu.edu  Wed Mar 21 16:46:04 2007
From: oliphant at ee.byu.edu (Travis Oliphant)
Date: Wed, 21 Mar 2007 13:46:04 -0700
Subject: [SciPy-user] odr thread safe?
In-Reply-To: <46016E74.9060309@gmail.com>
References: <etq7vp$ldc$1@sea.gmane.org> <46016E74.9060309@gmail.com>
Message-ID: <4601998C.7010404@ee.byu.edu>

Robert Kern wrote:

>Hmm. I'm not sure how this is happening. Calls to extension functions should be
>holding the global interpreter lock unless if it is explicitly released, and I'm
>pretty sure that I'm not. It's possible that the GIL gets released when the
>Python callbacks execute; I'd have to check. In that case, we would need a lock
>around the call to __odrpack.odr().
>
>But yeah, there's a difficult problem with calling a FORTRAN subroutine multiple
>concurrent times. We need to pass function pointer callbacks to the FORTRAN
>subroutine. These need to be real C functions. They can't have state, so the
>state needs to be global. C++ methods don't work. Consequently, the Python
>functions that the C callbacks will actually call are stored in odr_global.
>
>Looking at the Python C API docs, a possibility is to use
>PyThreadState_GetDict() to get a thread-specific dictionary. If threads aren't
>being used, then you only get NULL, though, so another mechanism will have to
>exist for the unthreaded case. I think. I could be misinterpreting the
>documentation.
>  
>

In NumPy for the ufuncs, we use PyThreadState_GetDict and then if that 
is NULL use PyEval_GetBuiltins so that we always get a per-thread 
dictionary to store information in.

Something like that could be done.

-Travis





>  http://docs.python.org/api/threads.html
>
>Let me know what you come up with!
>
>  
>



From cookedm at physics.mcmaster.ca  Wed Mar 21 19:57:55 2007
From: cookedm at physics.mcmaster.ca (David M. Cooke)
Date: Wed, 21 Mar 2007 19:57:55 -0400
Subject: [SciPy-user] building numpy/scipy on Solaris
In-Reply-To: <eec410113.10113eec4@steria.com>
References: <eec410113.10113eec4@steria.com>
Message-ID: <20070321235755.GA4496@arbutus.physics.mcmaster.ca>

On Wed, Mar 21, 2007 at 02:27:56PM +0100, raphael langella wrote:
> ---- Messages d?origine ----
> De: "David M. Cooke" <cookedm at physics.mcmaster.ca>
> Date: mercredi, mars 21, 2007 11:23 am
> Objet: Re: [SciPy-user] building numpy/scipy on Solaris
> 
> > On Wed, Mar 21, 2007 at 09:59:04AM +0100, raphael langella wrote:
> > > > On Tue, Mar 20, 2007 at 12:08:14PM +0100, raphael langella wrote:
> > > > > I'm trying to build numpy and scipy on Solaris 8.
> > > > > The BLAS FAQ on netlib.org suggests using optimized BLAS 
> > librairies> > > provided by computer vendor, like the SUN 
> > Performance Library. This
> > > > > library is supposed to provide enhanced and optimized version 
> > of 
> > > > BLAS> and LAPACK. I happen to have Forte 7 installed, so I 
> > first 
> > > > tried to
> > > > > build against this library (libsunperf.a).
> > > 
> > > It's supposed to support LAPACK v3.0 and BLAS1, 2 & 3
> > > (http://developers.sun.com/sunstudio/perflib_index.html).
> > > It gives me this when I import numpy :
> > > ImportError: ld.so.1: python: fatal: relocation error: file
> > > /usr/lib/python2.3/site-packages/numpy/linalg/lapack_lite.so: symbol
> > > __getenv_: referenced symbol not found
> > 
> > Odd. lapack_lite doesn't use getenv anywhere, so it must be the 
> > sunperflibrary (and http://docs.sun.com/source/819-
> > 3692/plug_optimizing.htmlshows that it does look at environment 
> > variables). I'm guessing there's
> > some link-time option that it needs.
> 
> ok. I'll forget about the other errors I have and will focus on
> libsunperf for now. It makes sense to me to use vendor provided
> optimized BLAS/LAPACK libraries.
> 
> I compiled with these options :
> python setup.py config_fc --f77flags='-xlic_lib=sunperf -xarch=v9b' install
> 
> and now, I've got this error when importing numpy :
> ImportError: ld.so.1: python: fatal: relocation error: file
> /usr/local/lib/python2.3/site-packages/numpy/linalg/lapack_lite.so:
> symbol dgeev_: referenced symbol not found

It looks like the C versions of the Fortran routines in sunperf don't
have trailing underscores (they're not just the Fortran routine, they're
C wrappers). Try

CPPFLAGS='-DNO_APPEND_FORTRAN' python setup.py config_fc --f77flags='-xlic_lib=sunperf -xarch=v9b' install

so that lapack_litemodule.c uses dgeev instead of dgeev_.

-- 
|>|\/|<
/--------------------------------------------------------------------------\
|David M. Cooke                      http://arbutus.physics.mcmaster.ca/dmc/
|cookedm at physics.mcmaster.ca


From jim at well.com  Thu Mar 22 00:25:55 2007
From: jim at well.com (jim stockford)
Date: Wed, 21 Mar 2007 20:25:55 -0800
Subject: [SciPy-user] call for speakers (bayPIGgies)
Message-ID: <292db2ef401c498af6222c387abd0b2f@well.com>



The bayPIGgies group is (always) looking for speakers.
Our email discussion reflects a lot of interest in SciPy.
If any of you are interested in presenting SciPy issues,
you'll have an enthusiastic audience.

Let me know, I'll let you know what speaking entails (not
a lot, at minimum).

hopefully,
jim at well.com



From raphael.langella at steria.com  Thu Mar 22 04:31:04 2007
From: raphael.langella at steria.com (raphael langella)
Date: Thu, 22 Mar 2007 09:31:04 +0100
Subject: [SciPy-user] building numpy/scipy on Solaris
Message-ID: <a17bbc5e.bc5ea17b@steria.com>

---- Messages d?origine ----
De: "David M. Cooke" <cookedm at physics.mcmaster.ca>
Date: jeudi, mars 22, 2007 0:57 am
Objet: Re: [SciPy-user] building numpy/scipy on Solaris

> On Wed, Mar 21, 2007 at 02:27:56PM +0100, raphael langella wrote:
> > ---- Messages d?origine ----
> > De: "David M. Cooke" <cookedm at physics.mcmaster.ca>
> > Date: mercredi, mars 21, 2007 11:23 am
> > Objet: Re: [SciPy-user] building numpy/scipy on Solaris
> > 
> > > On Wed, Mar 21, 2007 at 09:59:04AM +0100, raphael langella wrote:
> > > > > On Tue, Mar 20, 2007 at 12:08:14PM +0100, raphael langella 
> wrote:> > > > > I'm trying to build numpy and scipy on Solaris 8.
> > > > > > The BLAS FAQ on netlib.org suggests using optimized BLAS 
> > > librairies> > > provided by computer vendor, like the SUN 
> > > Performance Library. This
> > > > > > library is supposed to provide enhanced and optimized 
> version 
> > > of 
> > > > > BLAS> and LAPACK. I happen to have Forte 7 installed, so I 
> > > first 
> > > > > tried to
> > > > > > build against this library (libsunperf.a).
> > > > 
> > > > It's supposed to support LAPACK v3.0 and BLAS1, 2 & 3
> > > > (http://developers.sun.com/sunstudio/perflib_index.html).
> > > > It gives me this when I import numpy :
> > > > ImportError: ld.so.1: python: fatal: relocation error: file
> > > > /usr/lib/python2.3/site-packages/numpy/linalg/lapack_lite.so: 
> symbol> > > __getenv_: referenced symbol not found
> > > 
> > > Odd. lapack_lite doesn't use getenv anywhere, so it must be the 
> > > sunperflibrary (and http://docs.sun.com/source/819-
> > > 3692/plug_optimizing.htmlshows that it does look at environment 
> > > variables). I'm guessing there's
> > > some link-time option that it needs.
> > 
> > ok. I'll forget about the other errors I have and will focus on
> > libsunperf for now. It makes sense to me to use vendor provided
> > optimized BLAS/LAPACK libraries.
> > 
> > I compiled with these options :
> > python setup.py config_fc --f77flags='-xlic_lib=sunperf -
> xarch=v9b' install
> > 
> > and now, I've got this error when importing numpy :
> > ImportError: ld.so.1: python: fatal: relocation error: file
> > /usr/local/lib/python2.3/site-packages/numpy/linalg/lapack_lite.so:
> > symbol dgeev_: referenced symbol not found
> 
> It looks like the C versions of the Fortran routines in sunperf don't
> have trailing underscores (they're not just the Fortran routine, 
> they'reC wrappers). Try
> 
> CPPFLAGS='-DNO_APPEND_FORTRAN' python setup.py config_fc --
> f77flags='-xlic_lib=sunperf -xarch=v9b' install
> 
> so that lapack_litemodule.c uses dgeev instead of dgeev_.

Well, now it does uses dgeev, but it's still not found :
ImportError: ld.so.1: python: fatal: relocation error: file
/home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/numpy/linalg/lapack_lite.so:
symbol dgeev: referenced symbol not found

So, the function is just absent from the library?
I set BLAS and LAPACK to libsunperf.a. Should I use a dynamic version
instead?


From cookedm at physics.mcmaster.ca  Thu Mar 22 04:41:01 2007
From: cookedm at physics.mcmaster.ca (David M. Cooke)
Date: Thu, 22 Mar 2007 04:41:01 -0400
Subject: [SciPy-user] building numpy/scipy on Solaris
In-Reply-To: <a17bbc5e.bc5ea17b@steria.com>
References: <a17bbc5e.bc5ea17b@steria.com>
Message-ID: <20070322084101.GA5508@arbutus.physics.mcmaster.ca>

On Thu, Mar 22, 2007 at 09:31:04AM +0100, raphael langella wrote:
> > > 
> > > I compiled with these options :
> > > python setup.py config_fc --f77flags='-xlic_lib=sunperf -
> > xarch=v9b' install
> > > 
> > > and now, I've got this error when importing numpy :
> > > ImportError: ld.so.1: python: fatal: relocation error: file
> > > /usr/local/lib/python2.3/site-packages/numpy/linalg/lapack_lite.so:
> > > symbol dgeev_: referenced symbol not found
> > 
> > It looks like the C versions of the Fortran routines in sunperf don't
> > have trailing underscores (they're not just the Fortran routine, 
> > they'reC wrappers). Try
> > 
> > CPPFLAGS='-DNO_APPEND_FORTRAN' python setup.py config_fc --
> > f77flags='-xlic_lib=sunperf -xarch=v9b' install
> > 
> > so that lapack_litemodule.c uses dgeev instead of dgeev_.
> 
> Well, now it does uses dgeev, but it's still not found :
> ImportError: ld.so.1: python: fatal: relocation error: file
> /home/user1/ctcils/poladmin/rla/root/lib/python2.3/site-packages/numpy/linalg/lapack_lite.so:
> symbol dgeev: referenced symbol not found
> 
> So, the function is just absent from the library?
> I set BLAS and LAPACK to libsunperf.a. Should I use a dynamic version
> instead?

Ahh, just realized. We can do this completely with the C compiler; the
-xlic_lib=sunperf -xarch=v9b is only being used for Fortran. Try

CFLAGS='-xlic_lib=sunperf -xarch=v9b' CPPFLAGS='-DNO_APPEND_FORTRAN' python setup.py install

Since sunperf has C bindings, we don't need the Fortran compiler at all
for Numpy (and I don't think it was being used in the first place).

-- 
|>|\/|<
/--------------------------------------------------------------------------\
|David M. Cooke                      http://arbutus.physics.mcmaster.ca/dmc/
|cookedm at physics.mcmaster.ca


From raphael.langella at steria.com  Thu Mar 22 05:39:42 2007
From: raphael.langella at steria.com (raphael langella)
Date: Thu, 22 Mar 2007 10:39:42 +0100
Subject: [SciPy-user] building numpy/scipy on Solaris
Message-ID: <d9bbd009.d009d9bb@steria.com>

---- Messages d?origine ----
De: "David M. Cooke" <cookedm at physics.mcmaster.ca>
Date: jeudi, mars 22, 2007 9:41 am
Objet: Re: [SciPy-user] building numpy/scipy on Solaris

> Ahh, just realized. We can do this completely with the C compiler; the
> -xlic_lib=sunperf -xarch=v9b is only being used for Fortran. Try
> 
> CFLAGS='-xlic_lib=sunperf -xarch=v9b' CPPFLAGS='-
> DNO_APPEND_FORTRAN' python setup.py install
> 
> Since sunperf has C bindings, we don't need the Fortran compiler at 
> allfor Numpy (and I don't think it was being used in the first place).

well, options for C don't have the same syntax as Fortran, so I used
'-mcpu=v9 -lsunperf' and got :

ImportError: ld.so.1: python: fatal: relocation error: file
/Produits/sun/forte/7/SUNWspro/lib/libsunperf.so.4: symbol __f95_sign:
referenced symbol not found

but then, I realized I was using the default libsunperf (v8). As I'm
building on sparc III, I want v9b optimizations, so I did :

LD_LIBRARY_PATH=/Produits/sun/forte/7/SUNWspro/lib/v9b:$LD_LIBRARY_PATH

and got :

ld: fatal: file /Produits/sun/forte/7/SUNWspro/lib/v9b/libsunperf.so:
wrong ELF class: ELFCLASS64

but I'm building 64 bits code, am I not? I added the -m64 option and got :

/Produits/publics/sparc.SunOS.5.8/python/2.3.4/include/python2.3/pyport.h:554:2:
#error "LONG_BIT definition appears wrong for platform (bad gcc/glibc
config?)."

Damn, it looks like my python was compiled in 32 bits. Well, if I have
to keep to 32 bits, so be it. Going back a few lines up, I have an
__f95_sign symbol not found (removing -mcpu=v9 doesn't change anything).


From bryanv at enthought.com  Thu Mar 22 10:00:57 2007
From: bryanv at enthought.com (Bryan Van de Ven)
Date: Thu, 22 Mar 2007 09:00:57 -0500
Subject: [SciPy-user] call for speakers (bayPIGgies)
In-Reply-To: <292db2ef401c498af6222c387abd0b2f@well.com>
References: <292db2ef401c498af6222c387abd0b2f@well.com>
Message-ID: <46028C19.3080106@enthought.com>

Hi Jim,

My name is Bryan Van de Ven and I work for Enthought.  I have actually been 
meaning to contact BayPiggies about the possibility of speaking since I will be 
coming out to the Bay Area later this Summer (July/August?).  I didn't imagine 
talking about SciPy directly, but about some combination of Traits (auto GUI 
creation, easy MVC framework), Chaco (2D interactive plotting), and maybe a 
little about mlab/mayavi (3D visualization). I'd certainly show some examples 
that used SciPy, though.  My schedule is flexible so if there is interest a 
short talk on these topics I can certainly make my trip coincide with one of 
your meetings. Please let me know!

Best Regards,

Bryan Van de Ven

jim stockford wrote:
> 
> The bayPIGgies group is (always) looking for speakers.
> Our email discussion reflects a lot of interest in SciPy.
> If any of you are interested in presenting SciPy issues,
> you'll have an enthusiastic audience.
> 
> Let me know, I'll let you know what speaking entails (not
> a lot, at minimum).
> 
> hopefully,
> jim at well.com
> 
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
> 



From nmarais at sun.ac.za  Thu Mar 22 10:34:12 2007
From: nmarais at sun.ac.za (Neilen Marais)
Date: Thu, 22 Mar 2007 16:34:12 +0200
Subject: [SciPy-user] linsolve.factorized was: Re: Using umfpack to
	calculate a incomplete LU factorisation (ILU)
References: <pan.2007.03.06.19.34.13.563384@sun.ac.za>
	<45EE899D.9040304@ntc.zcu.cz>
	<45F00D7A.4060603@iam.uni-stuttgart.de>
	<45F03884.3000101@ntc.zcu.cz> <45F04386.6010105@ntc.zcu.cz>
Message-ID: <pan.2007.03.22.14.34.11.503475@sun.ac.za>

Hi Robert!

On Thu, 08 Mar 2007 18:10:30 +0100, Robert Cimrman wrote:

> Robert Cimrman wrote:
> Well, I did it since I am going to need this, too :-)
> 
> In [3]:scipy.linsolve.factorized?
> ...
> Definition:     scipy.linsolve.factorized(A)
> Docstring:
>     Return a fuction for solving a linear system, with A pre-factorized.
> 
>     Example:
>       solve = factorized( A ) # Makes LU decomposition.
>       x1 = solve( rhs1 ) # Uses the LU factors.
>       x2 = solve( rhs2 ) # Uses again the LU factors.
> 
> This uses UMFPACK if available.

This is a useful improvement, thanks. But why not just extend
linsolve.splu to use umfpack so we can present a consistent interface? The
essential difference between factorized and splu is that you get to
explicity control the storage of the LU factorisation and get some
additional info (i.e. the number of nonzeros), whereas factorised only
gives you a solve function. The actual library used to do the sparse LU is
just an implementation detail that should abstracted wherever possible, no?

If nobody complains about the idea I'm willing to implement it.

Thanks
Neilen

> 
> cheers,
> r.

-- 
you know its kind of tragic 
we live in the new world
but we've lost the magic
-- Battery 9 (www.battery9.co.za)



From cimrman3 at ntc.zcu.cz  Thu Mar 22 10:46:07 2007
From: cimrman3 at ntc.zcu.cz (Robert Cimrman)
Date: Thu, 22 Mar 2007 15:46:07 +0100
Subject: [SciPy-user] linsolve.factorized was: Re: Using umfpack
 to	calculate a incomplete LU factorisation (ILU)
In-Reply-To: <pan.2007.03.22.14.34.11.503475@sun.ac.za>
References: <pan.2007.03.06.19.34.13.563384@sun.ac.za>	<45EE899D.9040304@ntc.zcu.cz>	<45F00D7A.4060603@iam.uni-stuttgart.de>	<45F03884.3000101@ntc.zcu.cz>
	<45F04386.6010105@ntc.zcu.cz>
	<pan.2007.03.22.14.34.11.503475@sun.ac.za>
Message-ID: <460296AF.4000405@ntc.zcu.cz>

Neilen Marais wrote:
>> Robert Cimrman wrote:
>> Well, I did it since I am going to need this, too :-)
>>
>> In [3]:scipy.linsolve.factorized?
>> ...
>> Definition:     scipy.linsolve.factorized(A)
>> Docstring:
>>     Return a fuction for solving a linear system, with A pre-factorized.
>>
>>     Example:
>>       solve = factorized( A ) # Makes LU decomposition.
>>       x1 = solve( rhs1 ) # Uses the LU factors.
>>       x2 = solve( rhs2 ) # Uses again the LU factors.
>>
>> This uses UMFPACK if available.
> 
> This is a useful improvement, thanks. But why not just extend
> linsolve.splu to use umfpack so we can present a consistent interface? The
> essential difference between factorized and splu is that you get to
> explicity control the storage of the LU factorisation and get some
> additional info (i.e. the number of nonzeros), whereas factorised only
> gives you a solve function. The actual library used to do the sparse LU is
> just an implementation detail that should abstracted wherever possible, no?
> 
> If nobody complains about the idea I'm willing to implement it.

Sure, splu is an exception, every effort making it consistent is
welcome. But note that umfpack always gives you complete LU factors,
there is no ILU (drop-off) support - how would you tackle this?

Maybe change its name to get_superlu_obj or something like that, use
use_solver( useUmfpack = False ) at its beginning, and restore the
use_solver setting at the end?

r.


From alexander.borghgraef.rma at gmail.com  Fri Mar 23 05:38:48 2007
From: alexander.borghgraef.rma at gmail.com (Alexander Borghgraef)
Date: Fri, 23 Mar 2007 10:38:48 +0100
Subject: [SciPy-user] Gaussian pyramid using ndimage
In-Reply-To: <BCD52B8D-1FE6-4616-85BF-4BE4103F02B9@stanford.edu>
References: <9e8c52a20703160942j3d1b4ca5ubbfab6dabe39f7c0@mail.gmail.com>
	<BCD52B8D-1FE6-4616-85BF-4BE4103F02B9@stanford.edu>
Message-ID: <9e8c52a20703230238m3465261ey84b3335fa4e6fe6f@mail.gmail.com>

On 3/16/07, Zachary Pincus <zpincus at stanford.edu> wrote:
>
> I *think* that all of the filters in ndimage proceed by pre-filtering
> and then doing whatever task is required, so even if there was a
> downsampling filter (maybe 'zoom' would work), it wouldn't be memory-
> efficient in the way you were hoping.
>
> Also, the spline filters used by ndimage for pre-filtering seem
> somewhat broken:
> http://projects.scipy.org/scipy/scipy/ticket/213


 Hmm. So ndimage isn't really ready for use yet? Pity, image processing is
my main use for scipy.
Documentation should be extended too, it's not always clear what exactly an
algorithm is doing, what
the valid parameters are and what they're doing. I also noticed
generic_filter requires the filter function
to work on an unraveled array, and has to return a scalar, so you can't
create tensor images with it
(motionfields, gradient fields, in my case the 3D normal field for an
elevation map...). Some missing
functionality there, and definitely in need of better documentation.


-- 
Alex Borghgraef
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070323/2fb244ac/attachment.html>

From acorriga at gmu.edu  Fri Mar 23 12:02:56 2007
From: acorriga at gmu.edu (Andrew Corrigan)
Date: Fri, 23 Mar 2007 16:02:56 +0000 (UTC)
Subject: [SciPy-user] spsolve not working
References: <45FD40BA.1020107@gmu.edu> <web-80528738@uni-stuttgart.de>
	<loom.20070318T202627-760@post.gmane.org>
Message-ID: <loom.20070323T170200-114@post.gmane.org>

Turns out I was just using spsolve incorrectly.  There was no problem with it.



From anand at soe.ucsc.edu  Fri Mar 23 12:23:34 2007
From: anand at soe.ucsc.edu (Anand Patil)
Date: Fri, 23 Mar 2007 09:23:34 -0700
Subject: [SciPy-user] Numpy distutils failing to report ndarrayobject.h
Message-ID: <4603FF06.2090206@cse.ucsc.edu>

Hi all,

The following in a Pyrex extension module:

cdef extern from "numpy/ndarrayobject.h":
    void* PyArray_DATA(object obj)

produces the following in C:

#include "numpy/ndarrayobject.h".

The numpy distutils apparently fail to compile the extension module 
under win2k, Python 2.4:

PyMC2/PyrexLazyFunction.c:15:33: numpy/ndarrayobject.h: No such file or directory
PyMC2/PyrexLazyFunction.c: In function
`__pyx_f_17PyrexLazyFunction_12LazyFunction_get_array_data':


but it compiles fine on my machine, running OSX 10.4 and Python 2.5. 
Unfortunately I can't experiment much because I don't have a Windows 
machine available. Can anyone help me out with this?

Thanks,
Anand


From gnurser at googlemail.com  Fri Mar 23 12:38:45 2007
From: gnurser at googlemail.com (George Nurser)
Date: Fri, 23 Mar 2007 16:38:45 +0000
Subject: [SciPy-user] sparse & odr failures on latest scipy svn
Message-ID: <1d1e6ea70703230938s7b774ea2p3df53123fdecccf1@mail.gmail.com>

I get one error in the sparse tests and two in the odr tests for scipy v 2869

python 2.3, numpy svn 3591, acml math, 64 bit Linux on opteron.
These errors were not there last time I updated (v 2324)

--George Nurser.


....

..............F..F..............
======================================================================
ERROR: check_normalize (scipy.sparse.tests.test_sparse.test_coo)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/noc/users/agn/ext/Linux/lib64/python2.3/site-packages/scipy/sparse/tests/test_sparse.py",
line 812, in check_normalize
    assert(zip(ncol,nrow,ndata) == sorted(zip(col,row,data))) #should
sort by cols, then rows
NameError: global name 'sorted' is not defined

======================================================================
FAIL: test_explicit (scipy.tests.test_odr.test_odr)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/noc/users/agn/ext/Linux/lib64/python2.3/site-packages/scipy/odr/tests/test_odr.py",
line 49, in test_explicit
    np.array([  1.2646548050648876e+03,  -5.4018409956678255e+01,
  File "/noc/users/agn/ext/Linux/lib64/python/numpy/testing/utils.py",
line 230, in assert_array_almost_equal
    header='Arrays are not almost equal')
  File "/noc/users/agn/ext/Linux/lib64/python/numpy/testing/utils.py",
line 215, in assert_array_compare
    assert cond, msg
AssertionError:
Arrays are not almost equal

(mismatch 100.0%)
 x: array([  1.26462971e+03,  -5.42545890e+01,  -8.64250389e-02])
 y: array([  1.26465481e+03,  -5.40184100e+01,  -8.78497122e-02])

======================================================================
FAIL: test_multi (scipy.tests.test_odr.test_odr)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/noc/users/agn/ext/Linux/lib64/python2.3/site-packages/scipy/odr/tests/test_odr.py",
line 190, in test_multi
    np.array([ 4.3799880305938963,  2.4333057577497703,  8.0028845899503978,
  File "/noc/users/agn/ext/Linux/lib64/python/numpy/testing/utils.py",
line 230, in assert_array_almost_equal
    header='Arrays are not almost equal')
  File "/noc/users/agn/ext/Linux/lib64/python/numpy/testing/utils.py",
line 215, in assert_array_compare
    assert cond, msg
AssertionError:
Arrays are not almost equal

(mismatch 100.0%)
 x: array([ 4.31272063,  2.44289312,  7.76215871,  0.55995622,  0.46423343])
 y: array([ 4.37998803,  2.43330576,  8.00288459,  0.51011472,  0.51739023])

----------------------------------------------------------------------
Ran 1620 tests in 4.210s

FAILED (failures=2, errors=1)
<unittest.TextTestRunner object at 0x2a9bdee1d0>


From koara at atlas.cz  Fri Mar 23 13:17:33 2007
From: koara at atlas.cz (koara at atlas.cz)
Date: Fri, 23 Mar 2007 18:17:33 +0100
Subject: [SciPy-user] sparse mmio bug?
Message-ID: <b3f08489d19d4f94b311a895771a8f33@atlas.cz>

Hello all,

When writing a sparse matrix to disk with io.mmwrite(sparse_matrix), the line with 

typecode = a.gettypecode() 

fails. Changing this to 

typecode = a.dtype.char

the same way it is used couple of line above for the dense matrix case solves the problem.


I use python2.5 with scipy 0.5.2 from Feisty repositories.
------------------------------------------

www.icqsms.cz/



From oliphant at ee.byu.edu  Fri Mar 23 13:25:45 2007
From: oliphant at ee.byu.edu (Travis Oliphant)
Date: Fri, 23 Mar 2007 10:25:45 -0700
Subject: [SciPy-user] Numpy distutils failing to report ndarrayobject.h
In-Reply-To: <4603FF06.2090206@cse.ucsc.edu>
References: <4603FF06.2090206@cse.ucsc.edu>
Message-ID: <46040D99.6020509@ee.byu.edu>

Anand Patil wrote:

>Hi all,
>
>The following in a Pyrex extension module:
>
>cdef extern from "numpy/ndarrayobject.h":
>    void* PyArray_DATA(object obj)
>
>produces the following in C:
>
>#include "numpy/ndarrayobject.h".
>
>The numpy distutils apparently fail to compile the extension module 
>under win2k, Python 2.4:
>
>PyMC2/PyrexLazyFunction.c:15:33: numpy/ndarrayobject.h: No such file or directory
>PyMC2/PyrexLazyFunction.c: In function
>`__pyx_f_17PyrexLazyFunction_12LazyFunction_get_array_data':
>
>
>but it compiles fine on my machine, running OSX 10.4 and Python 2.5. 
>Unfortunately I can't experiment much because I don't have a Windows 
>machine available. Can anyone help me out with this?
>  
>

The include directory is not in the list of directories to search for 
the header files and so it is not being found.  I'm not sure what would 
cause this.  It's possible that the windows system doesn't have NumPy 
installed (or else the headers are not installed).  Or, it's possible 
that numpy.distutils is not adding the NumPy directory to the list of 
directories to search for header files.

I'm not sure why it's not adding the location of the NumPy headers to 
the compile line.

-Travis



>Thanks,
>Anand
>_______________________________________________
>SciPy-user mailing list
>SciPy-user at scipy.org
>http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>  
>



From openopt at ukr.net  Fri Mar 23 14:39:16 2007
From: openopt at ukr.net (dmitrey)
Date: Fri, 23 Mar 2007 20:39:16 +0200
Subject: [SciPy-user] can basearray using somehow be tried already?
In-Reply-To: <20070319165534.GG31245@encolpuis>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>	<45FE6806.4040003@ias.u-psud.fr>
	<20070319104238.GK8808@aims.ac.za>	<45FE8EBE.1040002@enthought.com>
	<20070319165534.GG31245@encolpuis>
Message-ID: <46041ED4.9080307@ukr.net>

Hi all,
is currently any way to try using basearray?
first of all I'm very interested in operators (matmult, dotmult etc) - 
will they have MATLAB/Octave/omatrix/etc -like style
dot is present => dotwise operation
dot is absent => matrix operation
or other?

And what will creation way look like? I mean something like
x = [1 2 3; 4 5 6]

I spent some hours using web search and scipy.org but found nothing.
WBR, D.



From robert.kern at gmail.com  Fri Mar 23 15:08:18 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 23 Mar 2007 14:08:18 -0500
Subject: [SciPy-user] can basearray using somehow be tried already?
In-Reply-To: <46041ED4.9080307@ukr.net>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>	<45FE6806.4040003@ias.u-psud.fr>	<20070319104238.GK8808@aims.ac.za>	<45FE8EBE.1040002@enthought.com>	<20070319165534.GG31245@encolpuis>
	<46041ED4.9080307@ukr.net>
Message-ID: <460425A2.8040703@gmail.com>

dmitrey wrote:
> Hi all,
> is currently any way to try using basearray?

It does not exist. It was a proposal to add a minimal array object like that
from numpy to the Python core. That proposal is dead now. If you want an array
package for Python that is currently available, you want numpy.

  http://numpy.scipy.org

> first of all I'm very interested in operators (matmult, dotmult etc) - 
> will they have MATLAB/Octave/omatrix/etc -like style
> dot is present => dotwise operation
> dot is absent => matrix operation
> or other?

No, we do not have the ability to change Python's syntax to add operators. For
the array object, operators all act element-wise. numpy also provides a matrix
object based on the array object which implements the relevant operators as
matrix operations (e.g. * is a matrix multiplication rather than element-wise
multiplication).

> And what will creation way look like? I mean something like
> x = [1 2 3; 4 5 6]

x = numpy.array([[1, 2, 3], [4, 5, 6]])

Again, we don't have the ability to change Python's syntax to support other such
syntaxes.

If you want some documentation in terms that are familiar to Matlab and Octave
users, you should read this page:

  http://www.scipy.org/NumPy_for_Matlab_Users

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From perry at stsci.edu  Fri Mar 23 15:33:56 2007
From: perry at stsci.edu (Perry Greenfield)
Date: Fri, 23 Mar 2007 15:33:56 -0400
Subject: [SciPy-user] can basearray using somehow be tried already?
In-Reply-To: <460425A2.8040703@gmail.com>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>	<45FE6806.4040003@ias.u-psud.fr>	<20070319104238.GK8808@aims.ac.za>	<45FE8EBE.1040002@enthought.com>	<20070319165534.GG31245@encolpuis>
	<46041ED4.9080307@ukr.net> <460425A2.8040703@gmail.com>
Message-ID: <66EE74B6-330B-4603-970A-104824585FE1@stsci.edu>


On Mar 23, 2007, at 3:08 PM, Robert Kern wrote:

> dmitrey wrote:
>
>> first of all I'm very interested in operators (matmult, dotmult  
>> etc) -
>> will they have MATLAB/Octave/omatrix/etc -like style
>> dot is present => dotwise operation
>> dot is absent => matrix operation
>> or other?
>
> No, we do not have the ability to change Python's syntax to add  
> operators. For
> the array object, operators all act element-wise. numpy also  
> provides a matrix
> object based on the array object which implements the relevant  
> operators as
> matrix operations (e.g. * is a matrix multiplication rather than  
> element-wise
> multiplication).

While strictly true, there was a cool hack a while back that had  
nearly the same effect. I forget if it was looked into and found  
wanting for this purpose. Perhaps someone remembers if there was a  
reason this couldn't be used (or wouldn't be considered symbolic  
enough). For example, one could define an effective operator |mul|"   
and use it like:

mat1 |mul| mat2

Two drawbacks were that it prevents the use of the text of the  
operator (e.g., mul in this case) for other purposes, for example for  
a variable name, and there is no way to change the precedence of the  
bracketing operator ('|' in this case). Another is that I guess you  
would have to explicitly import the operator names into your  
namespace (instead of having to refer to |numpy.mat|)

http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/384122

Perry



From openopt at ukr.net  Fri Mar 23 15:21:29 2007
From: openopt at ukr.net (dmitrey)
Date: Fri, 23 Mar 2007 21:21:29 +0200
Subject: [SciPy-user] can basearray using somehow be tried already?
In-Reply-To: <460425A2.8040703@gmail.com>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>	<45FE6806.4040003@ias.u-psud.fr>	<20070319104238.GK8808@aims.ac.za>	<45FE8EBE.1040002@enthought.com>	<20070319165534.GG31245@encolpuis>	<46041ED4.9080307@ukr.net>
	<460425A2.8040703@gmail.com>
Message-ID: <460428B9.10309@ukr.net>

Thank you  Robert, I had read its before. As I understood from
some links like
http://wiki.python.org/moin/CodingProjectIdeas/PythonCore ,
basearray was intended to be added in Python 2.6.
So, you know for sure that it will not happen?
And what about other versions - 2.7 etc?
WBR, D.

Robert Kern wrote:
> dmitrey wrote:
>   
>> Hi all,
>> is currently any way to try using basearray?
>>     
>
> It does not exist. It was a proposal to add a minimal array object like that
> from numpy to the Python core. That proposal is dead now. If you want an array
> package for Python that is currently available, you want numpy.
>
>   http://numpy.scipy.org
>
>   
>> first of all I'm very interested in operators (matmult, dotmult etc) - 
>> will they have MATLAB/Octave/omatrix/etc -like style
>> dot is present => dotwise operation
>> dot is absent => matrix operation
>> or other?
>>     
>
> No, we do not have the ability to change Python's syntax to add operators. For
> the array object, operators all act element-wise. numpy also provides a matrix
> object based on the array object which implements the relevant operators as
> matrix operations (e.g. * is a matrix multiplication rather than element-wise
> multiplication).
>
>   
>> And what will creation way look like? I mean something like
>> x = [1 2 3; 4 5 6]
>>     
>
> x = numpy.array([[1, 2, 3], [4, 5, 6]])
>
> Again, we don't have the ability to change Python's syntax to support other such
> syntaxes.
>
> If you want some documentation in terms that are familiar to Matlab and Octave
> users, you should read this page:
>
>   http://www.scipy.org/NumPy_for_Matlab_Users
>
>   



From robert.kern at gmail.com  Fri Mar 23 15:47:21 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 23 Mar 2007 14:47:21 -0500
Subject: [SciPy-user] can basearray using somehow be tried already?
In-Reply-To: <460428B9.10309@ukr.net>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>	<45FE6806.4040003@ias.u-psud.fr>	<20070319104238.GK8808@aims.ac.za>	<45FE8EBE.1040002@enthought.com>	<20070319165534.GG31245@encolpuis>	<46041ED4.9080307@ukr.net>	<460425A2.8040703@gmail.com>
	<460428B9.10309@ukr.net>
Message-ID: <46042EC9.7040405@gmail.com>

dmitrey wrote:
> Thank you  Robert, I had read its before. As I understood from
> some links like
> http://wiki.python.org/moin/CodingProjectIdeas/PythonCore ,
> basearray was intended to be added in Python 2.6.

Ah. That project idea seems to have been kept over from last year where this was
actually a project that someone attempted. You linked to the student's
application previously. See my response to that post for why it's not quite the
project you think it is, why I think it is dead, and what is replacing it.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From robert.kern at gmail.com  Fri Mar 23 15:48:23 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 23 Mar 2007 14:48:23 -0500
Subject: [SciPy-user] can basearray using somehow be tried already?
In-Reply-To: <66EE74B6-330B-4603-970A-104824585FE1@stsci.edu>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>	<45FE6806.4040003@ias.u-psud.fr>	<20070319104238.GK8808@aims.ac.za>	<45FE8EBE.1040002@enthought.com>	<20070319165534.GG31245@encolpuis>	<46041ED4.9080307@ukr.net>
	<460425A2.8040703@gmail.com>
	<66EE74B6-330B-4603-970A-104824585FE1@stsci.edu>
Message-ID: <46042F07.5020701@gmail.com>

Perry Greenfield wrote:
> On Mar 23, 2007, at 3:08 PM, Robert Kern wrote:
> 
>> dmitrey wrote:
>>
>>> first of all I'm very interested in operators (matmult, dotmult  
>>> etc) -
>>> will they have MATLAB/Octave/omatrix/etc -like style
>>> dot is present => dotwise operation
>>> dot is absent => matrix operation
>>> or other?
>> No, we do not have the ability to change Python's syntax to add  
>> operators. For
>> the array object, operators all act element-wise. numpy also  
>> provides a matrix
>> object based on the array object which implements the relevant  
>> operators as
>> matrix operations (e.g. * is a matrix multiplication rather than  
>> element-wise
>> multiplication).
> 
> While strictly true, there was a cool hack a while back that had  
> nearly the same effect. I forget if it was looked into and found  
> wanting for this purpose. Perhaps someone remembers if there was a  
> reason this couldn't be used (or wouldn't be considered symbolic  
> enough).

What do you mean by "used"? There's no reason an individual couldn't use it, no.
It's entirely decoupled from anything else; i.e. no one else has to modify
anything in order to support it. Personally, I think the hack is pretty cool,
but its verbosity and precedence problems prevent me from actually using it.
Unfortunately, I think it simply doesn't solve the problem and creates more
magic in the process. I wouldn't want to see such pseudo-operators added to
numpy, for example.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From perry at stsci.edu  Fri Mar 23 15:54:36 2007
From: perry at stsci.edu (Perry Greenfield)
Date: Fri, 23 Mar 2007 15:54:36 -0400
Subject: [SciPy-user] can basearray using somehow be tried already?
In-Reply-To: <46042F07.5020701@gmail.com>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>	<45FE6806.4040003@ias.u-psud.fr>	<20070319104238.GK8808@aims.ac.za>	<45FE8EBE.1040002@enthought.com>	<20070319165534.GG31245@encolpuis>	<46041ED4.9080307@ukr.net>
	<460425A2.8040703@gmail.com>
	<66EE74B6-330B-4603-970A-104824585FE1@stsci.edu>
	<46042F07.5020701@gmail.com>
Message-ID: <7E1D131F-353B-4F16-9DEE-F194192E9F35@stsci.edu>


On Mar 23, 2007, at 3:48 PM, Robert Kern wrote:

> Perry Greenfield wrote:
>>
>> While strictly true, there was a cool hack a while back that had
>> nearly the same effect. I forget if it was looked into and found
>> wanting for this purpose. Perhaps someone remembers if there was a
>> reason this couldn't be used (or wouldn't be considered symbolic
>> enough).
>
> What do you mean by "used"? There's no reason an individual  
> couldn't use it, no.
> It's entirely decoupled from anything else; i.e. no one else has to  
> modify
> anything in order to support it. Personally, I think the hack is  
> pretty cool,
> but its verbosity and precedence problems prevent me from actually  
> using it.
> Unfortunately, I think it simply doesn't solve the problem and  
> creates more
> magic in the process. I wouldn't want to see such pseudo-operators  
> added to
> numpy, for example.

Yes, in the sense of being a standard part of numpy (or matrix add- 
ons, etc.). Since we aren't heavy matrix users (yet anyway) it  
doesn't matter that much to us. This solution is not as good as being  
able to define new operators in Python itself. On the other hand,  
perhaps it is a better solution than having '*' have different  
meanings for different flavors of arrays. That's magic of a different  
sort, and just as prone to causing problems. So I think it is  
arguable which is worse. I agree it brings it's own set of problems.

Perry



From openopt at ukr.net  Fri Mar 23 16:05:57 2007
From: openopt at ukr.net (dmitrey)
Date: Fri, 23 Mar 2007 22:05:57 +0200
Subject: [SciPy-user] can basearray using somehow be tried already?
In-Reply-To: <46042F07.5020701@gmail.com>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>	<45FE6806.4040003@ias.u-psud.fr>	<20070319104238.GK8808@aims.ac.za>	<45FE8EBE.1040002@enthought.com>	<20070319165534.GG31245@encolpuis>	<46041ED4.9080307@ukr.net>	<460425A2.8040703@gmail.com>	<66EE74B6-330B-4603-970A-104824585FE1@stsci.edu>
	<46042F07.5020701@gmail.com>
Message-ID: <46043325.3010000@ukr.net>

Ok, but isn't it possible to add derive class to numpy, based on 
ndarray, with MATLAB-like dotwise & matrix operations with any of 
ndarray & matrix instances? I.e. x*y (x^y, x/y etc) is matrix & x .* y 
(x.^y, x./y etc) is dotwise, if any of x & y are something like mat_array?
Currently I use it in my code things like
    def _matmult(self, x, y):
        return asarray(x) ** asarray(y)
 
    def _dotmult(self, x, y):
        return asarray(x) * asarray(y)

>See my response to that post for why it's not quite the
project you think it is, why I think it is dead, and what is replacing it.

Can't you give exact URL, please? I'm not familiar enough with mailing 
lists interface & browsing so I can't find it by myself.
WBR, D.


Robert Kern wrote:
> Perry Greenfield wrote:
>   
>> On Mar 23, 2007, at 3:08 PM, Robert Kern wrote:
>>
>>     
>>> dmitrey wrote:
>>>
>>>       
>>>> first of all I'm very interested in operators (matmult, dotmult  
>>>> etc) -
>>>> will they have MATLAB/Octave/omatrix/etc -like style
>>>> dot is present => dotwise operation
>>>> dot is absent => matrix operation
>>>> or other?
>>>>         
>>> No, we do not have the ability to change Python's syntax to add  
>>> operators. For
>>> the array object, operators all act element-wise. numpy also  
>>> provides a matrix
>>> object based on the array object which implements the relevant  
>>> operators as
>>> matrix operations (e.g. * is a matrix multiplication rather than  
>>> element-wise
>>> multiplication).
>>>       
>> While strictly true, there was a cool hack a while back that had  
>> nearly the same effect. I forget if it was looked into and found  
>> wanting for this purpose. Perhaps someone remembers if there was a  
>> reason this couldn't be used (or wouldn't be considered symbolic  
>> enough).
>>     
>
> What do you mean by "used"? There's no reason an individual couldn't use it, no.
> It's entirely decoupled from anything else; i.e. no one else has to modify
> anything in order to support it. Personally, I think the hack is pretty cool,
> but its verbosity and precedence problems prevent me from actually using it.
> Unfortunately, I think it simply doesn't solve the problem and creates more
> magic in the process. I wouldn't want to see such pseudo-operators added to
> numpy, for example.
>
>   



From robert.kern at gmail.com  Fri Mar 23 16:20:16 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 23 Mar 2007 15:20:16 -0500
Subject: [SciPy-user] can basearray using somehow be tried already?
In-Reply-To: <46043325.3010000@ukr.net>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>	<45FE6806.4040003@ias.u-psud.fr>	<20070319104238.GK8808@aims.ac.za>	<45FE8EBE.1040002@enthought.com>	<20070319165534.GG31245@encolpuis>	<46041ED4.9080307@ukr.net>	<460425A2.8040703@gmail.com>	<66EE74B6-330B-4603-970A-104824585FE1@stsci.edu>	<46042F07.5020701@gmail.com>
	<46043325.3010000@ukr.net>
Message-ID: <46043680.8070805@gmail.com>

dmitrey wrote:
> Ok, but isn't it possible to add derive class to numpy, based on 
> ndarray, with MATLAB-like dotwise & matrix operations with any of 
> ndarray & matrix instances? I.e. x*y (x^y, x/y etc) is matrix & x .* y 
> (x.^y, x./y etc) is dotwise, if any of x & y are something like mat_array?

You can't add new operators. You can only redefine the behavior of the existing
operators, and that's exactly what numpy.matrix does.

> Currently I use it in my code things like
>     def _matmult(self, x, y):
>         return asarray(x) ** asarray(y)
>  
>     def _dotmult(self, x, y):
>         return asarray(x) * asarray(y)
> 
>> See my response to that post for why it's not quite the
> project you think it is, why I think it is dead, and what is replacing it.
> 
> Can't you give exact URL, please? I'm not familiar enough with mailing 
> lists interface & browsing so I can't find it by myself.

http://projects.scipy.org/pipermail/numpy-discussion/2007-March/026618.html

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From lorenzo.isella at gmail.com  Fri Mar 23 21:46:57 2007
From: lorenzo.isella at gmail.com (Lorenzo Isella)
Date: Sat, 24 Mar 2007 02:46:57 +0100
Subject: [SciPy-user] Post-Processing 3D Array
Message-ID: <46048311.30308@gmail.com>

Dear All,
I am currently running a code to solve Navier-Stokes equations in a 
cylinder. The equations are solved in cylindrical coordinates 
(theta,r,z). The output is fundamentally made up of several 3D arrays 
containing e.g. the axial component of the velocity in space v_z[i,j,k] 
and so on.
I am using Python to post-process the data, but there are a few problems 
I do not know how to tackle.
(1) It would be very nice to be able to cut several cross-sections of 
the tube along which I'd like to plot v_z. I can easily obtain the 2D 
array v_z_cross[i,j], function of theta and r only, but then I am not 
sure about how to plot it in 2D. Also, the geometry of the resulting 
plot has to look like a circle and absolutely not a square.
(2)Would it also be possible to make a full 3D plot of v_z[i,j,k]? As in 
(1), the geometry of the resulting plot has to look like a cylinder and 
not a cube or a square duct.

Kind Regards

Lorenzo


From mwojc at p.lodz.pl  Sat Mar 24 08:00:42 2007
From: mwojc at p.lodz.pl (Marek Wojciechowski)
Date: Sat, 24 Mar 2007 13:00:42 +0100
Subject: [SciPy-user] ffnet 0.6 released
In-Reply-To: <mailman.6838.1174680925.29466.scipy-user@scipy.org>
References: <mailman.6838.1174680925.29466.scipy-user@scipy.org>
Message-ID: <200703241300.42900.mwojc@p.lodz.pl>

Hallo scipy users!
I'd like to announce new release of ffnet (feed-forward neural network for 
python). Project depends heavily on scipy/numpy, it's written partially in 
fortran and wrapped with f2py. Below i'm pasting original announce message 
posted to comp.lang.python.announce:

ffnet version 0.6 is now released. Source packages, 
Gentoo ebuilds and Windows binaries are now available 
for download at:

? ? http://ffnet.sourceforge.net

The last public release was 0.5.

If you are unfamiliar with this package, see the end of
this message for a description.

NEW FEATURES
? ? - trained network can be now exported to fortran source
? ? ? code and compiled
? ? - added new architecture generator (imlgraph)
? ? - added rprop training algorithm
? ? - added draft network plotting facility (based on networkx 
? ? ? and matplotlib)

CHANGES & BUG FIXES
? ? - fixed bug preventing ffnet form working with networkx-0.33
? ? - training data can be now numpy arrays
? ? - ffnet became a package now, but API should be compatibile 
? ? ? with previous version

DOCUMENTATION
? ? - docstrings of all objects have been improved
? ? - docs (automatically generated with epydoc) are avilable
? ? ? online and have been included to source distribution


What is ffnet?
--------------
ffnet is a fast and easy-to-use feed-forward neural
network training solution for python.

Unique features
---------------
1. Any network connectivity without cycles is allowed.
2. Training can be performed with use of several optimization 
? ?schemes including: standard backpropagation with momentum, rprop,
? ?conjugate gradient, bfgs, tnc, genetic alorithm based optimization.
3. There is access to exact partial derivatives of network outputs 
? ?vs. its inputs.
4. Automatic normalization of data.

Basic assumptions and limitations:
----------------------------------
1. Network has feed-forward architecture.
2. Input units have identity activation function, 
? ?all other units have sigmoid activation function.
3. Provided data are automatically normalized, both input and output, 
? ?with a linear mapping to the range (0.15, 0.85).
? ?Each input and output is treated separately (i.e. linear map is 
? ?unique for each input and output).
4. Function minimized during training is a sum of squared errors 
? ?of each output for each training pattern.
? ?
Performance
-----------
Excellent computational performance is achieved implementing core
functions in fortran 77 and wrapping them with f2py. ffnet outstands
in performance pure python training packages and is competitive to
'compiled language' software. Moreover, a trained network can be 
exported to fortran sources, compiled and called from many
programming languages.

Usage
-----
Basic usage of the package is outlined below:

? ? from ffnet import ffnet, mlgraph, savenet, loadnet, exportnet
? ? conec = mlgraph( (2,2,1) )
? ? net = ffnet(conec)
? ? input = [ [0.,0.], [0.,1.], [1.,0.], [1.,1.] ]
? ? target ?= [ [1.], [0.], [0.], [1.] ]
? ? net.train_tnc(input, target, maxfun = 1000)
? ? net.test(input, target, iprint = 2)
? ? savenet(net, "xor.net")
? ? exportnet(net, "xor.f")
? ? net = loadnet("xor.net")
? ? answer = net( [ 0., 0. ] )
? ? partial_derivatives = net.derivative( [ 0., 0. ] )

Usage examples with full description can be found in 
examples directory of the source distribution or browsed
at http://ffnet.sourceforge.net.


From karol.langner at kn.pl  Sat Mar 24 08:39:40 2007
From: karol.langner at kn.pl (Karol Langner)
Date: Sat, 24 Mar 2007 13:39:40 +0100
Subject: [SciPy-user] can basearray using somehow be tried already?
In-Reply-To: <46042EC9.7040405@gmail.com>
References: <5f56302b0703190328s33362261n37871db9cdeeb2a7@mail.gmail.com>
	<460428B9.10309@ukr.net> <46042EC9.7040405@gmail.com>
Message-ID: <200703241339.40222.karol.langner@kn.pl>

On Friday 23 of March 2007 20:47, Robert Kern wrote:
> dmitrey wrote:
> > Thank you  Robert, I had read its before. As I understood from
> > some links like
> > http://wiki.python.org/moin/CodingProjectIdeas/PythonCore ,
> > basearray was intended to be added in Python 2.6.
>
> Ah. That project idea seems to have been kept over from last year where
> this was actually a project that someone attempted. You linked to the
> student's application previously. See my response to that post for why it's
> not quite the project you think it is, why I think it is dead, and what is
> replacing it.

I am the one that attempted that project, without success. I added a short 
annotation to the SciPy wiki project page from last year.

Karol

-- 
written by Karol Langner
Sat Mar 24 13:38:08 CET 2007


From nwagner at iam.uni-stuttgart.de  Sat Mar 24 15:13:20 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Sat, 24 Mar 2007 20:13:20 +0100
Subject: [SciPy-user] sparse & odr failures on latest scipy svn
In-Reply-To: <1d1e6ea70703230938s7b774ea2p3df53123fdecccf1@mail.gmail.com>
References: <1d1e6ea70703230938s7b774ea2p3df53123fdecccf1@mail.gmail.com>
Message-ID: <web-80801996@uni-stuttgart.de>

On Fri, 23 Mar 2007 16:38:45 +0000
  "George Nurser" <gnurser at googlemail.com> wrote:
> I get one error in the sparse tests and two in the odr 
>tests for scipy v 2869
> 
> python 2.3, numpy svn 3591, acml math, 64 bit Linux on 
>opteron.
> These errors were not there last time I updated (v 2324)
> 
> --George Nurser.
> 
> 
> ....
> 
> ..............F..F..............
> ======================================================================
> ERROR: check_normalize 
>(scipy.sparse.tests.test_sparse.test_coo)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>  File 
>"/noc/users/agn/ext/Linux/lib64/python2.3/site-packages/scipy/sparse/tests/test_sparse.py",
> line 812, in check_normalize
>    assert(zip(ncol,nrow,ndata) == 
>sorted(zip(col,row,data))) #should
> sort by cols, then rows
> NameError: global name 'sorted' is not defined
> 
> ======================================================================
>FAIL: test_explicit (scipy.tests.test_odr.test_odr)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>  File 
>"/noc/users/agn/ext/Linux/lib64/python2.3/site-packages/scipy/odr/tests/test_odr.py",
> line 49, in test_explicit
>    np.array([  1.2646548050648876e+03, 
> -5.4018409956678255e+01,
>  File 
>"/noc/users/agn/ext/Linux/lib64/python/numpy/testing/utils.py",
> line 230, in assert_array_almost_equal
>    header='Arrays are not almost equal')
>  File 
>"/noc/users/agn/ext/Linux/lib64/python/numpy/testing/utils.py",
> line 215, in assert_array_compare
>    assert cond, msg
> AssertionError:
> Arrays are not almost equal
> 
> (mismatch 100.0%)
> x: array([  1.26462971e+03,  -5.42545890e+01, 
> -8.64250389e-02])
> y: array([  1.26465481e+03,  -5.40184100e+01, 
> -8.78497122e-02])
> 
> ======================================================================
>FAIL: test_multi (scipy.tests.test_odr.test_odr)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>  File 
>"/noc/users/agn/ext/Linux/lib64/python2.3/site-packages/scipy/odr/tests/test_odr.py",
> line 190, in test_multi
>    np.array([ 4.3799880305938963,  2.4333057577497703, 
> 8.0028845899503978,
>  File 
>"/noc/users/agn/ext/Linux/lib64/python/numpy/testing/utils.py",
> line 230, in assert_array_almost_equal
>    header='Arrays are not almost equal')
>  File 
>"/noc/users/agn/ext/Linux/lib64/python/numpy/testing/utils.py",
> line 215, in assert_array_compare
>    assert cond, msg
> AssertionError:
> Arrays are not almost equal
> 
> (mismatch 100.0%)
> x: array([ 4.31272063,  2.44289312,  7.76215871, 
> 0.55995622,  0.46423343])
> y: array([ 4.37998803,  2.43330576,  8.00288459, 
> 0.51011472,  0.51739023])
> 
> ----------------------------------------------------------------------
> Ran 1620 tests in 4.210s
> 
>FAILED (failures=2, errors=1)
> <unittest.TextTestRunner object at 0x2a9bdee1d0>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user

See

http://projects.scipy.org/scipy/scipy/ticket/357

Nils
  


From bryan at cole.uklinux.net  Sat Mar 24 16:49:04 2007
From: bryan at cole.uklinux.net (Bryan Cole)
Date: Sat, 24 Mar 2007 20:49:04 +0000
Subject: [SciPy-user] Post-Processing 3D Array
In-Reply-To: <46048311.30308@gmail.com>
References: <46048311.30308@gmail.com>
Message-ID: <1174769344.6343.31.camel@pc1.cole.uklinux.net>


> I am currently running a code to solve Navier-Stokes equations in a 
> cylinder. The equations are solved in cylindrical coordinates 
> (theta,r,z). The output is fundamentally made up of several 3D arrays 
> containing e.g. the axial component of the velocity in space v_z[i,j,k] 
> and so on.
> I am using Python to post-process the data, but there are a few problems 
> I do not know how to tackle.
> (1) It would be very nice to be able to cut several cross-sections of 
> the tube along which I'd like to plot v_z. I can easily obtain the 2D 
> array v_z_cross[i,j], function of theta and r only, but then I am not 
> sure about how to plot it in 2D. Also, the geometry of the resulting 
> plot has to look like a circle and absolutely not a square.
> (2)Would it also be possible to make a full 3D plot of v_z[i,j,k]? As in 
> (1), the geometry of the resulting plot has to look like a cylinder and 
> not a cube or a square duct.

Sounds like a job for VTK (www.vtk.org). In VTK-speak, your dataset is
best represented as a "structured grid". This means your data points can
be indexed as i,j,k (i.e. a grid), but it's not a Cartesian grid, so the
3D location of each data point must be specified directly. At each data
point, you can have an arbitrary number of scalar, vector or tensor
attributes.

I would say the easiest way to start is to write out your data as a VTK
format file (this can be ASCII or binary) with format description given
in http://www.vtk.org/pdf/file-formats.pdf , then load this up in either
MayaVi (http://mayavi.sourceforge.net/) or Paraview
(http://www.paraview.org/HTML/Index.html). 

There is a python module for writing VTK files called pyvtk
(http://cens.ioc.ee/projects/pyvtk/ ) which may simplify the task
further.

HTH
BC

> 
> Kind Regards
> 
> Lorenzo




From gael.varoquaux at normalesup.org  Sun Mar 25 08:33:31 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Sun, 25 Mar 2007 14:33:31 +0200
Subject: [SciPy-user] Post-Processing 3D Array
In-Reply-To: <1174769344.6343.31.camel@pc1.cole.uklinux.net>
References: <46048311.30308@gmail.com>
	<1174769344.6343.31.camel@pc1.cole.uklinux.net>
Message-ID: <20070325123331.GA12920@clipper.ens.fr>

On Sat, Mar 24, 2007 at 08:49:04PM +0000, Bryan Cole wrote:
> > (1) It would be very nice to be able to cut several cross-sections of 
> > the tube along which I'd like to plot v_z. I can easily obtain the 2D 
> > array v_z_cross[i,j], function of theta and r only, but then I am not 
> > sure about how to plot it in 2D. Also, the geometry of the resulting 
> > plot has to look like a circle and absolutely not a square.
> > (2)Would it also be possible to make a full 3D plot of v_z[i,j,k]? As in 
> > (1), the geometry of the resulting plot has to look like a cylinder and 
> > not a cube or a square duct.

> Sounds like a job for VTK (www.vtk.org). In VTK-speak, your dataset is
> best represented as a "structured grid". This means your data points can
> be indexed as i,j,k (i.e. a grid), but it's not a Cartesian grid, so the
> 3D location of each data point must be specified directly. At each data
> point, you can have an arbitrary number of scalar, vector or tensor
> attributes.

> I would say the easiest way to start is to write out your data as a VTK
> format file (this can be ASCII or binary) with format description given
> in http://www.vtk.org/pdf/file-formats.pdf , then load this up in either
> MayaVi (http://mayavi.sourceforge.net/) or Paraview
> (http://www.paraview.org/HTML/Index.html). 

> There is a python module for writing VTK files called pyvtk
> (http://cens.ioc.ee/projects/pyvtk/ ) which may simplify the task
> further.

Indeed I would say that vtk is very suited for this task. Tvtk and
mayavi2 are python wrapper for vtk that are very pythonic and nice for
interactiv work. I am currently working on a module to drive mayavi2 from
"ipython -wthread" for scripting and interactive work. It still needs
some work before I can even submit it for inclusion to mayavi2 but I am
sending it along just in case in can be of some use. It can be of much
help as it deals with the building of the vtk data objects from numpy
arrays for most cases.

Tvtk and mayavi2 are part of the enthought tools suite, if you want to
use the attached module you will have to use the svn:
https://svn.enthought.com/enthought/wiki/GrabbingAndBuilding

For more info on mayavi2 and tvtk see the scipy wiki at
http://www.scipy.org (I cannot provide a direct link to the proper page,
as the wiki is down right now).

HTH

Ga?l



From fperez.net at gmail.com  Sun Mar 25 13:53:08 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Sun, 25 Mar 2007 11:53:08 -0600
Subject: [SciPy-user] Post-Processing 3D Array
In-Reply-To: <20070325123331.GA12920@clipper.ens.fr>
References: <46048311.30308@gmail.com>
	<1174769344.6343.31.camel@pc1.cole.uklinux.net>
	<20070325123331.GA12920@clipper.ens.fr>
Message-ID: <db6b5ecc0703251053v6ca89d87r4ecc2604dec5fc81@mail.gmail.com>

On 3/25/07, Gael Varoquaux <gael.varoquaux at normalesup.org> wrote:

[...]
> "ipython -wthread" for scripting and interactive work. It still needs
> some work before I can even submit it for inclusion to mayavi2 but I am
> sending it along just in case in can be of some use.

Did you forget to click 'attach' ?

Cheers,

f

ps - I'm interested :)


From zunzun at zunzun.com  Sun Mar 25 14:33:00 2007
From: zunzun at zunzun.com (zunzun at zunzun.com)
Date: Sun, 25 Mar 2007 14:33:00 -0400
Subject: [SciPy-user] www.scipy.org appears to be down
Message-ID: <20070325183259.GA1028@zunzun.com>

Quick note: www.scipy.org appears to be down.

     James



From gael.varoquaux at normalesup.org  Sun Mar 25 18:23:59 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Mon, 26 Mar 2007 00:23:59 +0200
Subject: [SciPy-user] Post-Processing 3D Array
In-Reply-To: <db6b5ecc0703251053v6ca89d87r4ecc2604dec5fc81@mail.gmail.com>
References: <46048311.30308@gmail.com>
	<1174769344.6343.31.camel@pc1.cole.uklinux.net>
	<20070325123331.GA12920@clipper.ens.fr>
	<db6b5ecc0703251053v6ca89d87r4ecc2604dec5fc81@mail.gmail.com>
Message-ID: <20070325222356.GA153@clipper.ens.fr>

On Sun, Mar 25, 2007 at 11:53:08AM -0600, Fernando Perez wrote:
> On 3/25/07, Gael Varoquaux <gael.varoquaux at normalesup.org> wrote:

> [...]
> > "ipython -wthread" for scripting and interactive work. It still needs
> > some work before I can even submit it for inclusion to mayavi2 but I am
> > sending it along just in case in can be of some use.

> Did you forget to click 'attach' ?

Why do I always do this !! The good thing about this is that I spent
three hours in a train hacking on this module this afternoon, so you get
more features and less code !

> ps - I'm interested :)

I am interested in comments, eventhought this is really work in progress.

Ga?l
-------------- next part --------------
"""A simple wrapper around tvtk.tools.mlab suitable for MayaVi2!  This
is meant to be used from the embedded Python interpreter in MayaVi2 or
from IPython with the "-wthread" switch.

There are several test functions at the end of this file that are
illustrative to look at.

"""

# Author: Prabhu Ramachandran <prabhu_r at users.sf.net>
# Copyright (c) 2007, Enthought, Inc.
# License: BSD Style.

#TODO:  * Add optionnal scalars to plot3d
#       * Make streamline display colors by default
#       * Ask Prabhu what the difference is between surf and mesh.
#         Propose mesh fall removal.
#       * Find out if tri_mesh is still needed. If not kill it.
#       * Propose surf_regular_c for removal

# Standard library imports.
import scipy

# Enthought library imports.
from enthought.envisage import get_application
from enthought.tvtk.api import tvtk
from enthought.tvtk.tools import mlab
from enthought.traits.api import HasTraits, Instance
from enthought.traits.ui.api import View, Item, Group

# MayaVi related imports.
from enthought.mayavi.services import IMAYAVI
from enthought.mayavi.sources.vtk_data_source import VTKDataSource
from enthought.mayavi.filters.filter_base import FilterBase
from enthought.mayavi.modules.surface import Surface
from enthought.mayavi.modules.vectors import Vectors
from enthought.mayavi.modules.iso_surface import IsoSurface
from enthought.mayavi.modules.streamline import Streamline
from enthought.mayavi.modules.glyph import Glyph
from enthought.mayavi.app import Mayavi
from enthought.mayavi.core.source import Source
from enthought.mayavi.core.module import Module
from enthought.mayavi.core.module_manager import ModuleManager
from enthought.mayavi.sources.array_source import ArraySource


######################################################################
# Application and mayavi instances.
application = get_application()
mayavi = None
if application is not None:
    mayavi = application.get_service(IMAYAVI)


######################################################################
# `ImageActor` class

# This should be added as a new MayaVi module.  It is here for testing
# and further improvements.
class ImageActor(Module):

    # An image actor.
    actor = Instance(tvtk.ImageActor, allow_none=False)

    view = View(Group(Item(name='actor', style='custom',
                           resizable=True),
                      show_labels=False),
                width=500,
                height=500,
                resizable=True)

    def setup_pipeline(self):
        self.actor = tvtk.ImageActor()
        
    def update_pipeline(self):
        """Override this method so that it *updates* the tvtk pipeline
        when data upstream is known to have changed.
        """
        mm = self.module_manager
        if mm is None:
            return
        src = mm.source
        self.actor.input = src.outputs[0]
        self.pipeline_changed = True

    def update_data(self):
        """Override this method so that it flushes the vtk pipeline if
        that is necessary.
        """
        # Just set data_changed, the component should do the rest.
        self.data_changed = True

    def _actor_changed(self, old, new):
        if old is not None:
            self.actors.remove(old)
        self.actors.append(new)

######################################################################
# Utility functions.
def _make_glyph_data(points, vectors=None, scalars=None):
    """Makes the data for glyphs using mlab.
    """
    g = mlab.Glyphs(points, vectors, scalars)
    return g.poly_data

def _make_default_figure():
    """Checks to see if a valid mayavi instance is running.  If not
    creates a new one.
    """
    global mayavi
    if mayavi is None or application.stopped is not None:
        mayavi = figure()
    return mayavi

def _add_data(tvtk_data, name=''):
    """Add a TVTK data object `tvtk_data` to the mayavi pipleine.
    Give the object a name of `name`.
    """
    if isinstance(tvtk_data, tvtk.Object):
        d = VTKDataSource()
        d.data = tvtk_data
    elif isinstance(tvtk_data, Source):
        d = tvtk_data
    else:
        raise TypeError, \
              "first argument should be either a TVTK object"\
              " or a mayavi source"

    if len(name) > 0:
        d.name = name
    mayavi = _make_default_figure()
    mayavi.add_source(d)
    return d

def _traverse(node):
    """Traverse a tree accessing the nodes children attribute.
    """
    try:
        for leaf in node.children:
            for leaflet in _traverse(leaf):
                yield leaflet
    except AttributeError:
        pass
    yield node

def _find_data(object):
    """Goes up the vtk pipeline to find the data sources of a given
    object.
    """
    if isinstance(object, ModuleManager):
        inputs = [object.source]
    elif hasattr(object, 'module_manager'):
        inputs = [object.module_manager.source]
    elif hasattr(object, 'data') or isinstance(object, ArraySource):
        inputs = [object]
    else:
        raise TypeError, 'Cannot find data source for given object'
    data_sources = []
    try:
        while True:
            input = inputs.pop()
            if hasattr(input, 'inputs'):
                inputs += input.inputs
            elif hasattr(input, 'image_data'):
                data_sources.append(input.image_data)
            else:
                data_sources.append(input.data)
    except IndexError:
        pass
    return data_sources

def _has_scalar_data(object):
    """Tests if an object has scalar data.
    """
    data_sources = _find_data(object)
    for source in data_sources:
        if source.point_data.scalars is not None:
            return True
        elif source.cell_data.scalars is not None:
            return True
    return False

def _has_vector_data(object):
    """Tests if an object has vector data.
    """
    data_sources = _find_data(object)
    for source in data_sources:
        if source.point_data.vectors is not None:
            return True
        elif source.cell_data.vectors is not None:
            return True
    return False

def _has_tensor_data(object):
    """Tests if an object has tensor data.
    """
    data_sources = _find_data(object)
    for source in data_sources:
        if source.point_data.tensors is not None:
            return True
        elif source.cell_data.tensors is not None:
            return True
    return False

def _find_module_manager(object=None, data_type=None):
    """If an object is specified, returns its module_manager, elsewhere finds
    the first module_manager in the scene.
    """
    if object is None:
        for object in _traverse(gcf()):
            if isinstance(object, ModuleManager):
                if ((data_type == 'scalar' and not _has_scalar_data(object))    
                  or (data_type == 'vector' and not _has_vector_data(object))
                  or (data_type == 'tensor' and not _has_tensor_data(object))):
                    continue
                return object
        else:
            print("No object in the scene has a color map")
    else:
        if hasattr(object, 'module_manager'):
            if ((data_type == 'scalar' and _has_scalar_data(object))
               or (data_type == 'vector' and _has_vector_data(object))
               or (data_type == 'tensor' and _has_tensor_data(object))
                or data_type is None):
                return object.module_manager
            else:
                print("This object has no %s data" % data_type)
        else:
            print("This object has no color map")
    return None

def _orient_colorbar(colorbar, orientation):
    """Orients the given colorbar (make it horizontal or vertical).
    """
    if orientation == "vertical":
        colorbar.orientation = "vertical"
        colorbar.width = 0.1
        colorbar.height = 0.8
        colorbar.position = (0.01, 0.15)
    elif orientation == "horizontal":
        colorbar.orientation = "horizontal"
        colorbar.width = 0.8
        colorbar.height = 0.17
        colorbar.position = (0.1, 0.01)
    else:
        print "Unknown orientation"
    gcf().render()

def _typical_distance(data_obj):
    """ Returns a typical distance in a cloud of points.
        This is done by taking the size of the bounding box, and dividing it
        by the cubic root of the number of points.
    """
    x_min, x_max, y_min, y_max, z_min, z_max = data_obj.bounds
    distance = scipy.sqrt(((x_max-x_min)**2 + (y_max-y_min)**2 +
                           (z_max-z_min)**2)/(4*
                           data_obj.number_of_points**(0.33)))
    if distance == 0:
        return 1
    else:
        return 0.4*distance

######################################################################
# Data creation

def scalarscatter(*args, **kwargs):
    """
    Creates scattered scalar data.

    Function signatures
    -------------------

        vectorscatter(s, ...)
        vectorscatter(x, y, z, s, ...)
        vectorscatter(x, y, z, f, ...)

    If only 1 array s is passed the x, y and z arrays are assumed to be
    made from the indices of vectors.

    If 4 positional arguments are passed the last one must be an array s, or
    acallable, f, that returns an array.

    Arguments
    ---------
 
        x -- x coordinates of the points of the mesh (optional).

        y -- y coordinates of the points of the mesh (optional).

        z -- z coordinates of the points of the mesh (optional).
        
        s -- scalar value

        f -- callable that is used to build the scalar data (only if 4 
             positional arguments are passed).

    Keyword arguments
    -----------------

        name -- The name of the vtk object created. Default: 'Scattered scalars'

        extent --  [xmin, xmax, ymin, ymax, zmin, zmax] 
                   Default is the x, y, z arrays extents.

    """
    if len(args)==1:
        s = args[0]
        x, y, z = scipy.indices(s.shape)
    elif len(args)==4:
        x, y, z, s = args
        if callable(s):
            s = f(x, y, z)
    else:
        raise ValueError, "wrong number of arguments"
    
    assert ( x.shape == y.shape and
             y.shape == z.shape and
             s.shape == z.shape ), "argument shape are not equal"

    if 'extent' in kwargs:
        xmin, xmax, ymin, ymax, zmin, zmax = kwargs.pop('extent')
        x = xmin + x*(xmax - xmin)/float(x.max() - x.min()) -x.min()
        y = ymin + y*(ymax - ymin)/float(y.max() - y.min()) -y.min()
        z = zmin + z*(zmax - zmin)/float(z.max() - z.min()) -z.min()

    points = scipy.c_[x.ravel(), y.ravel(), z.ravel()]
    scalars = s.ravel()
    name = kwargs.pop('name', 'Scattered scalars')

    data = _make_glyph_data(points, None, scalars)
    data_obj = _add_data(data, name)
    return data_obj

def vectorscatter(*args, **kwargs):
    """
    Creates scattered vector data.

    Function signatures
    -------------------

        vectorscatter(u, v, w, ...)
        vectorscatter(x, y, z, u, v, w, ...)
        vectorscatter(x, y, z, f, ...)

    If only 3 arrays u, v, w are passed the x, y and z arrays are assumed to be
    made from the indices of vectors.

    If 4 positional arguments are passed the last one must be a callable, f, 
    that returns vectors.

    Arguments
    ---------
 
        x -- x coordinates of the points of the mesh (optional).

        y -- y coordinates of the points of the mesh (optional).

        z -- z coordinates of the points of the mesh (optional).
        
        u -- x coordinnate of the vector field

        v -- y coordinnate of the vector field

        w -- z coordinnate of the vector field

        f -- callable that is used to build the vector field (only if 4 
             positional arguments are passed).

    Keyword arguments
    -----------------

        name -- The name of the vtk object created. Default: 'Scattered vector'

        extent --  [xmin, xmax, ymin, ymax, zmin, zmax] 
                   Default is the x, y, z arrays extents.

        scalars -- The scalars associated to the vectors. Defaults to none. 

    """
    if len(args)==3:
        u, v, w = args
        x, y, z = scipy.indices(u.shape)
    elif len(args)==6:
        x, y, z, u, v, w = args
    elif len(args)==4:
        x, y, z, f = args
        assert callable(f), "when used with 4 arguments last argument must be callable"
        u, v, w = f(x, y, z)
    else:
        raise ValueError, "wrong number of arguments"
    
    assert ( x.shape == y.shape and
             y.shape == z.shape and
             z.shape == u.shape and
             u.shape == v.shape and
             v.shape == w.shape ), "argument shape are not equal"

    if 'extent' in kwargs:
        xmin, xmax, ymin, ymax, zmin, zmax = kwargs.pop('extent')
        x = xmin + x*(xmax - xmin)/float(x.max() - x.min()) -x.min()
        y = ymin + y*(ymax - ymin)/float(y.max() - y.min()) -y.min()
        z = zmin + z*(zmax - zmin)/float(z.max() - z.min()) -z.min()

    points = scipy.c_[x.ravel(), y.ravel(), z.ravel()]
    vectors = scipy.c_[u.ravel(), v.ravel(), w.ravel()]
    if 'scalars' in kwargs:
        scalars = kwargs['scalars'].ravel()
    else:
        scalars = None
    name = kwargs.pop('name', 'Scattered vectors')

    data = _make_glyph_data(points, vectors, scalars)
    data_obj = _add_data(data, name)
    return data_obj

def scalarfield(*args, **kwargs):
    """
    Creates a scalar field data.

    Function signatures
    -------------------

        scalarfield(s, ...)
        scalarfield(x, y, z, s, ...)
        scalarfield(x, y, z, f, ...)

    If only 1 array s is passed the x, y and z arrays are assumed to
    be made from the indices of the s array.

    If the x, y and z arrays are passed they are supposed to have been
    generated by  `numpy.mgrid`.  The function builds a scalar field assuming 
    the points are regularily spaced.

    Arguments
    ---------
 
        x -- x coordinates of the points of the mesh (optional).

        y -- y coordinates of the points of the mesh (optional).

        z -- z coordinates of the points of the mesh (optional).
        
        s -- scalar values.

        f -- callable that is used to build the scalar field (only if 4 
             positional arguments are passed).

    Keyword arguments
    -----------------

        name -- The name of the vtk object created. Default: 'Scalar field'

        extent --  [xmin, xmax, ymin, ymax, zmin, zmax] 
                   Default is the x, y, z arrays extents.
    """

    # Get the keyword args.
    name = kwargs.get('name', 'Scalar field')

    if len(args)==1:
        s = args[0]
        x, y, z = scipy.indices(s.shape)
    elif len(args)==4:
        x, y, z, s = args
        if callable(s):
            s = f(x, y, z)
    else:
        raise ValueError, "wrong number of arguments"
    
    assert ( x.shape == y.shape and
             y.shape == z.shape and
             s.shape == z.shape ), "argument shape are not equal"

    if 'extent' in kwargs:
        xmin, xmax, ymin, ymax, zmin, zmax = kwargs.pop('extent')
        x = xmin + x*(xmax - xmin)/float(x.max() - x.min()) -x.min()
        y = ymin + y*(ymax - ymin)/float(y.max() - y.min()) -y.min()
        z = zmin + z*(zmax - zmin)/float(z.max() - z.min()) -z.min()

    points = scipy.c_[x.ravel(), y.ravel(), z.ravel()]
    dx = x[1, 0, 0] - x[0, 0, 0]
    dy = y[0, 1, 0] - y[0, 0, 0]
    dz = z[0, 0, 1] - z[0, 0, 0]

    data = ArraySource(scalar_data=s,
                      origin=[points[0, 0], points[0, 1], points[0, 2]],
                      spacing=[dx, dy, dz])
    data_obj = _add_data(data, name)
    return data_obj

def vectorfield(*args, **kwargs):
    """
    Creates a vector field data.

    Function signatures
    -------------------

        vectorfield(u, v, w, ...)
        vectorfield(x, y, z, u, v, w, ...)
        vectorfield(x, y, z, f, ...)

    If only 3 arrays u, v, w are passed the x, y and z arrays are assumed to
    be made from the indices of the u, v, w arrays.
        
    If the x, y and z arrays are passed they are supposed to have been
    generated by  `numpy.mgrid`.  The function builds a vector field assuming 
    the points are regularily spaced.

    Arguments
    ---------
 
        x -- x coordinates of the points of the mesh (optional).

        y -- y coordinates of the points of the mesh (optional).

        z -- z coordinates of the points of the mesh (optional).
        
        u -- x coordinnate of the vector field

        v -- y coordinnate of the vector field

        w -- z coordinnate of the vector field

        f -- callable that is used to build the vector field (only if 4 
             positional arguments are passed).

    Keyword arguments
    -----------------

        name -- The name of the vtk object created. Default: 'Vector field'

        extent --  [xmin, xmax, ymin, ymax, zmin, zmax] 
                   Default is the x, y, z arrays extents.

        scalars -- The scalars associated to the vectors. Defaults to none. 

        transpose_vectors -- If the additional argument
                             transpose_vectors is passed, then the
                             input vectors array is suitably
                             transposed.  By default transpose_vectors
                             is True so that the array is in the
                             correct format that VTK expects.
                             However, a transposed array is not
                             contiguous and thus a copy is made, this
                             also means that any changes to the users
                             input array will will not be reflected in
                             the renderered object (provided you know
                             how to do this).  Thus, sometimes users
                             might want to provide already transposed
                             data suitably formatted.  In these cases
                             one should set transpose_vectors to
                             False.
                             Default value: True
    """
    # Get the keyword args.
    transpose_vectors = kwargs.get('transpose_vectors', True)

    if len(args)==3:
        u, v, w = args
        x, y, z = scipy.indices(u.shape)
    elif len(args)==6:
        x, y, z, u, v, w = args
    elif len(args)==4:
        x, y, z, f = args
        assert callable(f), "when used with 4 arguments last argument must be callable"
        u, v, w = f(x, y, z)
    else:
        raise ValueError, "wrong number of arguments"
    
    assert ( x.shape == y.shape and
             y.shape == z.shape and
             z.shape == u.shape and
             u.shape == v.shape and
             v.shape == w.shape ), "argument shape are not equal"

    if 'extent' in kwargs:
        xmin, xmax, ymin, ymax, zmin, zmax = kwargs.pop('extent')
        x = xmin + x*(xmax - xmin)/float(x.max() - x.min()) -x.min()
        y = ymin + y*(ymax - ymin)/float(y.max() - y.min()) -y.min()
        z = zmin + z*(zmax - zmin)/float(z.max() - z.min()) -z.min()

    points = scipy.c_[x.ravel(), y.ravel(), z.ravel()]
    vectors = scipy.concatenate([u[..., scipy.newaxis], 
                                 v[..., scipy.newaxis],
                                 w[..., scipy.newaxis] ],
                                 axis=3)
    if 'scalars' in kwargs:
        scalars = kwargs['scalars']
    else:
        scalars = None
    name = kwargs.pop('name', 'Vector field')
    dx = x[1, 0, 0] - x[0, 0, 0]
    dy = y[0, 1, 0] - y[0, 0, 0]
    dz = z[0, 0, 1] - z[0, 0, 0]

    if not transpose_vectors:
        vectors.shape = vectors.shape[::-1]
    data = ArraySource(transpose_input_array=transpose_vectors,
                      vector_data=vectors,
                      scalar_data=scalars,
                      origin=[points[0, 0], points[0, 1], points[0, 2]],
                      spacing=[dx, dy, dz])
    data_obj = _add_data(data, name)
    return data_obj

######################################################################
# Module creation

def isosurface(data_obj, name='IsoSurface', transparent=True,
                    contours=5):
    """ Applies the IsoSsurface mayavi module to the given VTK data object.
    """
    mayavi.engine.current_object = data_obj

    iso = IsoSurface()

    # Check what type the 'contours' are and do whatever is needed.
    contour_list = True
    try:
        len(contours)
    except TypeError:
        contour_list = False

    if contour_list:
        iso.contour.contours = contours
    else: 
        assert type(contours) == int, "The contours argument must be an integer"
        assert contours > 1, "The contours argument must be positiv"
        iso.contour.set(auto_contours=True,
                            number_of_contours=contours)

    mayavi.add_module(iso)

    if transparent:
        data_range = iso.module_manager.scalar_lut_manager.data_range
        iso.module_manager.scalar_lut_manager.lut.alpha_range = \
                (0.2, 0.8)
        data_range = ( scipy.mean(data_range)
                        + 0.4 * ( data_range.max() - data_range.min()) 
                               * scipy.array([-1, 1])) 
        iso.scene.render()

    return iso

def vectors(data_obj, color=None,  name='Vectors', mode='2d',
                scale_factor=1.):
    """ Applies the Vectors mayavi module to the given VTK data object.
    """
    v = Vectors(name=name)
    mayavi.engine.current_object = data_obj
    mayavi.add_module(v)
    mode_map = {'2d': 0, 'arrow': 1, 'cone': 2, 'cylinder': 3,
                'sphere': 4, 'cube': 5, 'point': 6}
    if mode == 'point':
        v.glyph.glyph_source = tvtk.PointSource(radius=0,
                                                number_of_points=1)
    else:
         v.glyph.glyph_source = v.glyph.glyph_list[mode_map[mode]]
    if color:
        v.glyph.color_mode = 'no_coloring'
        v.actor.property.color = color
    elif _has_scalar_data(data_obj) :
        v.glyph.color_mode = 'color_by_scalar'
    else:
        v.glyph.color_mode = 'color_by_vector'
    v.glyph.glyph.scale_factor = scale_factor
    return v

def glyph(data_obj, color=None, name='Glyph', mode='2d'):
    """ Applies the Glyph mayavi module to the given VTK data object.
    """
    g = Glyph()
    mayavi.engine.current_object = data_obj
    mayavi.add_module(g)
    mode_map = {'2d': 0, 'arrow': 1, 'cone': 2, 'cylinder': 3,
                'sphere': 4, 'cube': 5, 'point': 6}
    if mode == 'point':
        g.glyph.glyph_source = tvtk.PointSource(radius=0,
                                                number_of_points=1)
    else:
         g.glyph.glyph_source = g.glyph.glyph_list[mode_map[mode]]
    if color:
        g.actor.property.color = color
    if _has_scalar_data(data_obj) :
        g.glyph.color_mode = 'color_by_scalar'
    return g

#FIXME : just started this procedure !! Need to modify the color so that if
# none it warps a scalar. Need to add a kwarg for the source.
def streamline(data_obj, color=None,  name='Streamline', ):
    """ Applies the Streamline mayavi module to the given VTK data object.
    """
    st = Streamline()
    mayavi.engine.current_object = data_obj
    mayavi.add_module(st)
    if color:
        st.actor.property.color = color
    elif _has_scalar_data(data_obj) :
        st.actor.mapper.scalar_visibility = True
    else:
        st.actor.mapper.interpolate_scalars_before_mapping = True
        st.actor.mapper.scalar_visibility = True
    return st

######################################################################
# Helper functions

def quiver3d(*args, **kwargs):
    """
    Plots glyphs (like arrows) indicating the direction of the vectors
    for a 3D volume of data supplied as arguments.

    Function signatures
    -------------------

        quiver3d(vectordata, ...)
        quiver3d(u, v, w, ...)
        quiver3d(x, y, z, u, v, w, ...)
        quiver3d(x, y, z, f, ...)

    If only one positional argument is passed, it should be VTK data
    object with vector data.

    If only 3 arrays u, v, w are passed the x, y and z arrays are assumed to be
    made from the indices of vectors.

    If 4 positional arguments are passed the last one must be a callable, f, 
    that returns vectors.

    Arguments
    ---------
        
        vectordata -- VTK data object with vector data, such as created
                      by vectorscatter of vectorfield.

        x -- x coordinates of the points of the mesh (optional).

        y -- y coordinates of the points of the mesh (optional).

        z -- z coordinates of the points of the mesh (optional).
        
        u -- x coordinnate of the vector field

        v -- y coordinnate of the vector field

        w -- z coordinnate of the vector field

        f -- callable that is used to build the vector field (only if 4 
             positional arguments are passed).

    Keyword arguments
    -----------------

        name -- The name of the vtk object created. Default: 'Quiver3D'

        mode -- This should be one of ['2d' (2d arrows),
                'arrow', 'cone', 'cylinder', 'sphere', 'cube',
                'point'] and depending on what is passed shows an
                appropriate glyph.  It defaults to a 3d arrow.

        extent --  [xmin, xmax, ymin, ymax, zmin, zmax] 
                   Default is the x, y, z arrays extents.

        scalars -- The scalars used to display the color of the glyphs. 
                  Defaults to the norm of the vectors.

        color -- The color of the glyphs in the absence of scalars. 
                 Default: (1., 0., 0.)

        autoscale -- Autoscale the size of the glyph.
                     Default: True

        scale_factor -- Default 1
    """
    if len(args)==1:
        data_obj = args[0]
    else:
        data_kwargs = kwargs.copy()
        data_kwargs.pop('name','')
        if len(args)==6:
            data_obj = vectorscatter(*args, **data_kwargs)
        else:
            data_obj = vectorfield(*args, **data_kwargs)
    
    if not 'name' in kwargs:
            kwargs['name'] = 'Quiver3D'

    if not 'mode' in kwargs:
            kwargs['mode'] = 'arrow'

    if not 'autoscale' in kwargs or kwargs['autoscale']:
        scale_factor = kwargs.get('scale_facotr', 1.)
        kwargs['scale_factor'] = (scale_factor * 
                        _typical_distance(_find_data(data_obj)[0]) )

    return vectors(data_obj, **kwargs)

def contour3d(*args, **kwargs):
    """
    Plots iso-surfaces for a 3D volume of data suplied as arguments.

    Function signatures
    -------------------

        contour3d(scalars, ...)
        contour3d(scalarfield, ...)

    Arguments
    ---------
 
        scalars -- A 3D array giving the value of the scalar on a grid.
        
        scalarfield -- VTK data object with scalar field data, such as
                       created by scalarfield.


    Keyword arguments
    -----------------

        name -- The name of the vtk object created. Default: 'Contour3D'

        transpose_scalars -- If the additional argument
                             transpose_scalars is passed, then the
                             input scalar array is suitably
                             transposed.  By default transpose_scalars
                             is True so that the array is in the
                             correct format that VTK expects.
                             However, a transposed array is not
                             contiguous and thus a copy is made, this
                             also means that any changes to the users
                             input array will will not be reflected in
                             the renderered object (provided you know
                             how to do this).  Thus, sometimes users
                             might want to provide already transposed
                             data suitably formatted.  In these cases
                             one should set transpose_scalars to
                             False.
                             Default value: True

        contours -- Integer/list specifying number/list of
                    iso-surfaces. Specifying 0 shows no contours.
                    Specifying a list of values will only give the
                    requested contours asked for.  Default: 3

        extent --  [xmin, xmax, ymin, ymax, zmin, zmax] 
                   Default is the shape of the scalars

        transparent -- Boolean to make the opacity of the isosurfaces depend
                       on the scalar.  Default: True
    """
    if len(args)==1:
        if hasattr(args[0], 'shape'):
            scalars = args[0]
            assert len(scalars.shape) == 3, 'Only 3D arrays are supported.'
            data_kwargs = kwargs.copy()
            data_kwargs.pop('contours', '')
            data_kwargs.pop('transparent', '')
            if not 'name' in kwargs:
                data_kwargs['name'] = 'Contour3D'
            data_obj = scalarfield(scalars, **data_kwargs)
        else:
            data_obj = args[0]
    else:
        raise TypeError, "contour3d takes only one argument"

    # Remove extra kwargs that are not needed for the isosurface.
    kwargs.pop('extent', '')
    kwargs.pop('name', '')

    return isosurface(data_obj, **kwargs)

######################################################################
# The Mlab functionality.

def plot3d(x, y, z, radius=0.01, use_tubes=True, color=(1., 0., 0.) ,
          name='Plot3D'):
    """Draws lines between points

    Arguments
    ---------

    x -- x coordinates of the points of the line

    y -- y coordinates of the points of the line

    z -- z coordinates of the points of the line

    Keyword arguments
    -----------------

    color -- color of the line. Default: (1., 0., 0.)

    use_tubes -- Enables the drawing of the lines as tubes. Default: True

    radius -- radius of the tubes created. Default: 0.01

    name -- The name of the vtk object created. Default: 'Line3D'

    """
    assert ( x.shape == y.shape and
             y.shape == z.shape and
             s.shape == z.shape ), "argument shape are not equal"

    points = c_[x, y, z]
    np = len(points) - 1
    lines = scipy.zeros((np, 2), 'l')
    lines[:,0] = scipy.arange(0, np-0.5, 1, 'l')
    lines[:,1] = scipy.arange(1, np+0.5, 1, 'l')
    pd = tvtk.PolyData(points=points, lines=lines)
    _add_data(pd, name)
    if use_tubes:
        filter = tvtk.TubeFilter(number_of_sides=6)
        filter.radius = radius
        f = FilterBase(filter=filter, name='TubeFilter')
        mayavi.add_filter(f)
    s = Surface()
    s.actor.mapper.scalar_visibility = False
    mayavi.add_module(s)
    s.actor.property.color = color
    return s

def surf_regular(x, y, z, warp=True, scale=[1.0, 1.0, 1.0],
                 name='SurfRegular', f_args=(), **f_kwargs):
    """Creates a surface given regularly spaced values of x, y and the
    corresponding z as arrays.  Also works if z is a function.
    Currently works only for regular data - can be enhanced later. The x
    and y arrays give the grid line positions along x and y.

    Arguments
    ---------
    
        x -- 1D array of x points (regularly spaced)

        y -- 1D array of y points (regularly spaced)

        z -- A 2D array for the x and y points with x varying fastest
        and y next.  Also will work if z is a callable which supports
        x and y arrays as the arguments.

    Keyword arguments
    -----------------

        warp -- If true, warp the data to show a 3D surface
        (default = 1).        

        scale -- Scale the x, y and z axis as per passed values.
        Defaults to [1.0, 1.0, 1.0].

        f_args -- additional positional arguments for func()
        (default is empty)

        f_kwargs -- a dict of additional keyword arguments for func()
        (default is empty)
    """
    data = mlab.make_surf_actor(x, y, z, warp, scale,
                                make_actor=False, *f_args, **f_kwargs)
    _add_data(data, name)
    if warp:
        from enthought.mayavi.filters.warp_scalar import WarpScalar
        from enthought.mayavi.filters.poly_data_normals import PolyDataNormals
        w = WarpScalar()
        w.filter.scale_factor=scale[2]
        mayavi.add_filter(w)
        n = PolyDataNormals()
        n.filter.feature_angle = 45
        mayavi.add_filter(n)
        
    s = Surface()
    mayavi.add_module(s)
    return s

def surf_regular_c(x, y, z, warp=True, scale=[1.0, 1.0, 1.0],
                   number_of_contours=10, name='SurfRegularC',
                   f_args=(), **f_kwargs):
    """ MayaVi1's `imv.surf` like functionality that plots surfaces
    given x (1D), y(1D) and z (or a callable) arrays.  Also plots
    contour lines.
    """
    s1 = surf_regular(x, y, z, warp, scale, name, *f_args, **f_kwargs)
    
    data_src = s1.module_manager.source.inputs[0].inputs[0]
    s2 = Surface(name='Contours')
    data_src.add_child(s2)
    s2.enable_contours = True
    s2.contour.number_of_contours = number_of_contours
    return s1, s2

def tri_mesh(triangles, points, scalars=None, scalar_visibility=False,
             surface=False, color=(0.5, 1.0, 0.5), name='TriMesh'):
    """Given triangle connectivity and points, plots a mesh of them.
    """
    data = mlab.make_triangle_polydata(triangles, points, scalars)
    _add_data(data, name)
    s = Surface()
    representation = 'w'
    if surface:
        representation = 's'
    s.actor.mapper.scalar_visibility = scalar_visibility
    mayavi.add_module(s)
    if surface:
        s.actor.property.set(diffuse=0.0, ambient=1.0, color=color,
                             representation=representation)
    else:
        s.actor.property.set(diffuse=1.0, ambient=0.0, color=color,
                             representation=representation)
    return s

def mesh(x, y, z, scalars=None, scalar_visibility=False,
         surface=False, color=(0.5, 1.0, 0.5), name='Mesh'):
    """Given x, y generated from scipy.mgrid, and a z to go with it.
    Along with optional scalars.  This class builds the triangle
    connectivity (assuming that x, y are from scipy.mgrid) and builds
    a mesh and shows it.

    Arguments 
    ---------

        x -- An array of x coordinate values formed using scipy.mgrid.

        y -- An array of y coordinate values formed using scipy.mgrid.
        
        z -- An array of z coordinate values formed using scipy.mgrid.

    Keyword arguments
    -----------------

        scalars -- An optional array of scalars to associate with the points.

        scalar_visibility -- A boolean switching the visibility of the scalars.

        color -- The color of the mesh in the absence of scalars.

        name -- The name of the vtk object created. Default is "Mesh"
    """
    triangles, points = mlab.make_triangles_points(x, y, z, scalars)        
    return tri_mesh(triangles, points, scalars,
                    scalar_visibility=scalar_visibility,
                    surface=surface, color=color, name=name)

def surf(*args, **kwargs):
    """
    Plots a surface using grid spaced data supplied as 2D arrays.

    Function signatures
    -------------------

        surf(z, scalars=z, ...)
        surf(x, y, z, scalars=z, ...)
        
    If only one array z is passed the x and y arrays are assumed to be made 
    of the indices of z.
    z is the elevation matrix. If no `scalars` argument is passed the color 
    of the surface also represents the z matrix. Setting the `scalars` argument
    to None prevents this.

    Arguments
    ---------
 
        x -- x coordinates of the points of the mesh (optional).

        y -- y coordinates of the points of the mesh (optional).

        z -- A 2D array giving the elevation of the mesh.
             Also will work if z is a callable which supports x and y arrays 
             as the arguments, but x and y must then be supplied.

    Keyword arguments
    -----------------

        extent --  [xmin, xmax, ymin, ymax, zmin, zmax] 
                   Default is the x, y, z arrays extents.

        scalars -- An array of the same shape as z that gives the color of the
                   surface. This can also be a function that takes x and
                   y as arguments.

        represention -- can be 'surface', 'wireframe' or 'points'.
                        Default is 'surface'

        color -- The color of the mesh in the absence of scalars.

        name -- The name of the vtk object created. Default is "Surf"
    """
    if len(args)==1:
        z = args[0]
        x, y = scipy.indices(z.shape)
    else:
        x, y, z = args
    if callable(z):
        z = z(x, y)
    if not 'scalars' in kwargs:
        kwargs['scalars'] = z
    if callable(kwargs['scalars']):
        kwargs['scalars'] = kwargs['scalars'](x, y)
    if 'color' in kwargs and kwargs['color']:
        kwargs['scalar_visibility'] = False
    if 'extent' in kwargs:
        xmin, xmax, ymin, ymax, zmin, zmax = kwargs.pop('extent')
        x = xmin + x*(xmax - xmin)/float(x.max() - x.min()) -x.min()
        y = ymin + y*(ymax - ymin)/float(y.max() - y.min()) -y.min()
        z = zmin + z*(zmax - zmin)/float(z.max() - z.min()) -z.min()
    return _surf(x, y, z, **kwargs)

def _surf(x, y, z, scalars=None, scalar_visibility=True,
          color=(0.5, 1.0, 0.5), name='Surf', representation='surface'):
    """ Functions that does the work for "surf". It is called with the right 
        number of arguments after the "surf" fonction does the magic to 
        translate the user-supplied arguments into something this function
        understands. """
    triangles, points = mlab.make_triangles_points(x, y, z, scalars)
    data = mlab.make_triangle_polydata(triangles, points, scalars)
    _add_data(data, name)
    
    s = Surface()
    s.actor.mapper.scalar_visibility = scalar_visibility
    mayavi.add_module(s)
    s.actor.property.color = color
    s.actor.property.represention = representation
    return s

def contour_surf(*args, **kwargs):
    """ Plots the contours of a surface using grid spaced data supplied as 
    2D arrays.

    Function signatures
    -------------------

        contour_surf(z, scalars=z, ...)
        contour_surf(x, y, z, scalars=z, ...)
        
    If only one array z is passed the x and y arrays are assumed to be made 
    of the indices of z.
    z is the elevation matrix. If no `scalars` argument is passed the
    contours are contour lines of the elevation, elsewhere they are
    contour lines of the scalar array.

    Arguments
    ---------
 
        x -- x coordinates of the points of the mesh (optional).

        y -- y coordinates of the points of the mesh (optional).

        z -- A 2D array giving the elevation of the mesh.
             Also will work if z is a callable which supports x and y arrays 
             as the arguments, but x and y must then be supplied.

    Keyword arguments
    -----------------

        extent --  [xmin, xmax, ymin, ymax, zmin, zmax] 
                   Default is the x, y, z arrays extents.

        contours -- Integer/list specifying number/list of
                    iso-surfaces. Specifying 0 shows no contours.
                    Specifying a list of values will only give the
                    requested contours asked for.  Default: 10

        scalars -- An array of the same shape as z that gives the scalar
                   data to plot the contours of. This can also be a function 
                   that takes x and y as arguments.

        color -- The color of the contour lines. If None, this is given by
                 the scalars.

        name -- The name of the vtk object created. Default is "Contour Surf"
    """
    
    contours = kwargs.pop('contours', 10)
    if not 'name' in kwargs:
        kwargs['name'] = "Contour Surf"
    s = surf(*args, **kwargs)
    s.enable_contours = True
    
    # Check what type the 'contours' are and do whatever is needed.
    contour_list = True
    try:
        len(contours)
    except TypeError:
        contour_list = False

    if contour_list:
        s.contour.contours = contours
        s.contour.set(auto_contours=False)
    else: 
        assert type(contours) == int, "The contours argument must be an integer"
        assert contours > 1, "The contours argument must be positiv"
        s.contour.set(auto_contours=True,
                            number_of_contours=contours)

    return s

def fancy_tri_mesh(triangles, points, scalars=None, scalar_visibility=False,
                   tube_radius=0.05, sphere_radius=0.05,
                   color=(0.5, 1.0, 0.5), name='FancyTriMesh'):
    """Plots the mesh using tubes and spheres so its fancier.
    """
    data = mlab.make_triangle_polydata(triangles, points, scalars)
    _add_data(data, name)

    # Extract the edges.
    ef = tvtk.ExtractEdges()
    extract_edges = FilterBase(filter=ef, name='ExtractEdges')
    mayavi.add_filter(extract_edges)

    # Now show the lines with tubes.
    tf = tvtk.TubeFilter(radius=tube_radius)
    tube = FilterBase(filter=tf, name='TubeFilter')
    mayavi.add_filter(tube)
    s = Surface(name='Tubes')
    s.actor.mapper.scalar_visibility = scalar_visibility
    mayavi.add_module(s)
    s.actor.property.color = color

    # Show the points with glyphs.
    g = Glyph(name='Spheres')
    g.glyph.glyph_source = g.glyph.glyph_list[4]
    g.glyph.glyph_source.radius = sphere_radius
    extract_edges.add_child(g)
    g.glyph.scale_mode = 'data_scaling_off'
    g.actor.mapper.scalar_visibility=scalar_visibility
    g.actor.property.color = color

def fancy_mesh(x, y, z, scalars=None, scalar_visibility=False,
               tube_radius=0.05, sphere_radius=0.05,
               color=(0.5, 1.0, 0.5), name='FancyMesh'):
    """Like mesh but shows the mesh using tubes and spheres.
    
    Arguments 
    ---------

        x -- An array of x coordinate values formed using scipy.mgrid.

        y -- An array of y coordinate values formed using scipy.mgrid.
        
        z -- An array of z coordinate values formed using scipy.mgrid.

    Keyword arguments
    -----------------

        scalars -- An optional array of scalars to associate with the
                   points.
    """
    triangles, points = mlab.make_triangles_points(x, y, z, scalars)
    return fancy_tri_mesh(triangles, points, scalars, scalar_visibility,
                          tube_radius, sphere_radius,
                          color, name=name)

def imshow(arr, scale=[1.0, 1.0, 1.0], interpolate=False,
           lut_mode='blue-red', file_name='',
           name='ImShow'):
    """Allows one to view a 2D Numeric array as an image.  This works
    best for very large arrays (like 1024x1024 arrays).

    Arguments
    ---------
    
    arr -- Array to be viewed.

    Keyword arguments
    -----------------

    scale -- Scale the x, y and z axis as per passed values.
             Defaults to [1.0, 1.0, 1.0].

    interpolate -- Boolean to interpolate the data in the image.
    """
    # FIXME
    assert len(arr.shape) == 2, "Only 2D arrays can be viewed!"

    ny, nx = arr.shape
    dx, dy, junk = scipy.array(scale)*1.0
    xa = scipy.arange(0, nx*scale[0] - 0.1*dx, dx, 'f')
    ya = scipy.arange(0, ny*scale[1] - 0.1*dy, dy, 'f')

    arr_flat = scipy.ravel(arr)
    min_val = min(arr_flat)
    max_val = max(arr_flat)

    sp = mlab._create_structured_points_direct(xa, ya)

    from enthought.mayavi.core.lut_manager import LUTManager
    lut = LUTManager(lut_mode=lut_mode, file_name=file_name)    
    lut.data_range = min_val, max_val
    a = lut.lut.map_scalars(arr_flat, 0, 0)
    a.name = 'scalars'
    sp.point_data.scalars = a
    sp.scalar_type = 'unsigned_char'
    sp.number_of_scalar_components = 4
    d = _add_data(sp, name)

    ia = ImageActor()
    ia.actor.interpolate = interpolate
    mayavi.add_module(ia)

######################################################################
# Non data-related drawing elements
def outline(object=None, color=None, name='Outline'):
    """Creates an outline for the current data.

    Keyword arguments
    -----------------
    
        object -- the object for which we create the outline
                       (default None).

        color -- The color triplet, eg: ( 1., 0., 0.)
    """
    from enthought.mayavi.modules.outline import Outline
    mayavi = _make_default_figure()
    scene = gcf()
    for obj in _traverse(scene):
        if isinstance(obj, Outline) and obj.name == name:
            o = obj
            break
    else:
        o = Outline(name=name)
        if object is not None:
            object.add_child(o)
        else:
            mayavi.add_module(o)
        if color is None:
            color = scene.scene.foreground
    if not color is None:
        o.actor.property.color = color
    return o

def axes(color=None, xlabel=None, ylabel=None, zlabel=None,
         object=None, name='Axes'):
    """Creates an axes for the current data.

    Keyword arguments
    -----------------
    
        color -- The color triplet, eg: (1., 0., 0.)

        xlabel -- the label of the x axis, default: ''

        ylabel -- the label of the y axis, default: ''

        zlabel -- the label of the z axis, default: ''

        object -- the object for which we create the axes.
    """
    from enthought.mayavi.modules.axes import Axes
    mayavi = _make_default_figure()
    scene = gcf()
    for obj in _traverse(scene):
        if isinstance(obj, Axes) and obj.name == name:
            a = obj
            break
    else:
        a = Axes(name=name)
        if object is not None:
            object.add_child(a)
        else:
            mayavi.add_module(a)
        if color is None:
            color = scene.scene.foreground
        if xlabel is None:
            xlabel = ''
        if ylabel is None:
            ylabel = ''
        if zlabel is None:
            zlabel = ''
    if color is not None:
        a.property.color = color
    if xlabel is not None:
        a.axes.x_label = xlabel
    if ylabel is not None:
        a.axes.y_label = ylabel
    if zlabel is not None:
        a.axes.z_label = zlabel
    return a

def figure():
    """If you are running from IPython this will start up mayavi for
    you!  This returns the current running MayaVi script instance.
    """
    global mayavi, application
    if mayavi is not None and application.stopped is None:
        mayavi.new_scene()
        return mayavi
    m = Mayavi()
    m.main()
    m.script.new_scene()
    engine = m.script.engine
    mayavi = m.script
    application = m.application
    return mayavi

def gcf():
    """Return a handle to the current figure.
    """
    return mayavi.engine.current_scene

def clf():
    """Clear the current figure.
    """
    try:
        scene = gcf()
        scene.children[:] = []
    except AttributeError:
        pass

def xlabel(text):
    """Creates a set of axes if there isn't already one, and sets the x label
    """
    return axes(xlabel=text)

def ylabel(text):
    """Creates a set of axes if there isn't already one, and sets the y label
    """
    return axes(ylabel=text)

def zlabel(text):
    """ Creates a set of axes if there isn't already one, and sets the z label
    """
    return axes(zlabel=text)

def title(text=None, color=None, size=None, name='Title'):
    """Creates a title for the figure.

    Keyword arguments
    -----------------

        text -- The text of the title, default: ''
    
        color -- The color triplet, eg: ( 1., 0., 0.)

        size -- The size, default: 1
    """
    from enthought.mayavi.modules.text import Text
    scene = gcf()
    for object in _traverse(scene):
        if isinstance(object, Text) and object.name==name:
            t = object
            break
    else:
        t = Text(name=name)
        mayavi.add_module(t)
        if color is None:
            color = scene.scene.foreground
        if text is None:
            text = 'title'
    if color is not None:
        t.property.color = color
    if text is not None:
        t.text = text
    if text is not None or size is not None:
        t.width = min(0.05*size*len(t.text), 1)
        t.x_position = 0.5*(1 - t.width)
        t.y_position = 0.8
    return t

def scalarbar(object=None, title=None, orientation=None):
    """Adds a colorbar for the scalar color mapping of the given object.

    If no object is specified, the first object with scalar data in the scene 
    is used.

    Keyword arguments
    -----------------
    
        title -- The title string 

        orientation -- Can be 'horizontal' or 'vertical'
    """
    module_manager = _find_module_manager(object=object, data_type="scalar")
    if module_manager is None:
        return
    if not module_manager.scalar_lut_manager.show_scalar_bar:
        if title is None:
            title = ''
        if orientation is None:
            orientation = 'horizontal'
    colorbar = module_manager.scalar_lut_manager.scalar_bar
    if title is not None:
        colorbar.title = title
    if orientation is not None:
        _orient_colorbar(colorbar, orientation)
    module_manager.scalar_lut_manager.show_scalar_bar = True
    return colorbar

def vectorbar(object=None, title=None, orientation=None):
    """Adds a colorbar for the vector color mapping of the given object.

    If no object is specified, the first object with vector data in the scene 
    is used.

    Keyword arguments
    -----------------

        object -- Optional object to get the vector lut from
    
        title -- The title string 

        orientation -- Can be 'horizontal' or 'vertical'
    """
    module_manager = _find_module_manager(object=object, data_type="vector")
    if module_manager is None:
        return
    if not module_manager.vector_lut_manager.show_scalar_bar:
        title = ''
        orientation = 'horizontal'
    colorbar = module_manager.vector_lut_manager.scalar_bar
    if title is not None:
        colorbar.title = title
    if orientation is not None:
        _orient_colorbar(colorbar, orientation)
    module_manager.vector_lut_manager.show_scalar_bar = True
    return colorbar

def colorbar(object=None, title=None, orientation=None):
    """Adds a colorbar for the color mapping of the given object. 
    
    If the object has scalar data, the scalar color mapping is 
    represented. Elsewhere the vector color mapping is represented, if 
    available.
    If no object is specified, the first object with a color map in the scene 
    is used.

    Keyword arguments
    -----------------

        object -- Optional object to get the vector lut from
    
        title       -- The title string 

        orientation -- Can be 'horizontal' or 'vertical'
    """
    colorbar = scalarbar(object=object, title=title, orientation=orientation)
    if colorbar is None:
        colorbar = vectorbar(object=object, title=title, orientation=orientation)
    return colorbar


######################################################################
# Test functions.
######################################################################
def test_arrows():
    a = arrows([[-1,-1,-1],[1,0,0]], [[1,1,1],[0,1,0]],
               color=(1,0,0))
    return a
    
def test_plot3d():
    """Generates a pretty set of lines."""
    n_mer, n_long = 6, 11
    pi = scipy.pi
    dphi = pi/1000.0
    phi = scipy.arange(0.0, 2*pi + 0.5*dphi, dphi, 'd')
    mu = phi*n_mer
    x = scipy.cos(mu)*(1+scipy.cos(n_long*mu/n_mer)*0.5)
    y = scipy.sin(mu)*(1+scipy.cos(n_long*mu/n_mer)*0.5)
    z = scipy.sin(n_long*mu/n_mer)*0.5

    l = plot3d(x, y, z, radius=0.05, color=(0.0, 0.0, 0.8))
    return l

def test_molecule():
    """Generates and shows a Caffeine molecule."""
    o = [[30, 62, 19],[8, 21, 10]]
    n = [[31, 21, 11], [18, 42, 14], [55, 46, 17], [56, 25, 13]]
    c = [[5, 49, 15], [30, 50, 16], [42, 42, 15], [43, 29, 13], [18, 28, 12],
         [32, 6, 8], [63, 36, 15], [59, 60, 20]]
    h = [[23, 5, 7], [32, 0, 16], [37, 5, 0], [73, 36, 16], [69, 60, 20],
         [54, 62, 28], [57, 66, 12], [6, 59, 16], [1, 44, 22], [0, 49, 6]]

    oxygen = spheres(o, radius=8, color=(1,0,0), name='Oxygen')
    nitrogen = spheres(n, radius=10, color=(0,0,1), name='Nitrogen')
    carbon = spheres(c, radius=10, color=(0,1,0), name='Carbon')
    hydrogen = spheres(h, radius=5, color=(1,1,1), name='Hydrogen')

def test_trimesh():
    """Test for simple triangle mesh."""
    pts = scipy.array([[0.0,0,0], [1.0,0.0,0.0], [1,1,0]], 'd')
    triangles = [[0, 1, 2]]
    t1 = tri_mesh(triangles, pts)
    pts1 = pts.copy()
    pts1[:,2] = 1.0
    t2 = fancy_tri_mesh(triangles, pts1)

def test_surf_regular(contour=1):
    """Test Surf on regularly spaced co-ordinates like MayaVi."""
    def f(x, y):
        sin, cos = scipy.sin, scipy.cos
        return sin(x+y) + sin(2*x - y) + cos(3*x+4*y)
        #return scipy.sin(x*y)/(x*y)

    x = scipy.arange(-7., 7.05, 0.1)
    y = scipy.arange(-5., 5.05, 0.05)
    if contour:
        s = surf_regular_c(x, y, f)
    else:
        s = surf_regular(x, y, f)
    return s

def test_simple_surf():
    """Test Surf with a simple collection of points."""
    x, y = scipy.mgrid[0:3:1,0:3:1]
    z = x
    s = surf(x, y, z, scipy.asarray(z, 'd'))

def test_surf():
    """A very pretty picture of spherical harmonics translated from
    the octaviz example."""
    pi = scipy.pi
    cos = scipy.cos
    sin = scipy.sin
    dphi, dtheta = pi/250.0, pi/250.0
    [phi,theta] = scipy.mgrid[0:pi+dphi*1.5:dphi,0:2*pi+dtheta*1.5:dtheta] 
    m0 = 4; m1 = 3; m2 = 2; m3 = 3; m4 = 6; m5 = 2; m6 = 6; m7 = 4;
    r = sin(m0*phi)**m1 + cos(m2*phi)**m3 + sin(m4*theta)**m5 + cos(m6*theta)**m7
    x = r*sin(phi)*cos(theta)
    y = r*cos(phi)
    z = r*sin(phi)*sin(theta);

    s = surf(x, y, z)

def test_mesh_sphere():
    """Create a simple sphere and test the mesh."""
    pi = scipy.pi
    cos = scipy.cos
    sin = scipy.sin    
    du, dv = pi/20.0, pi/20.0
    phi, theta = scipy.mgrid[0.01:pi+du*1.5:du, 0:2*pi+dv*1.5:dv]
    r = 1.0
    x = r*sin(phi)*cos(theta)
    y = r*sin(phi)*sin(theta)
    z = r*cos(phi)
    s = fancy_mesh(x, y, z, z, scalar_visibility=True,
                   tube_radius=0.01, sphere_radius=0.025)

def test_mesh():
    """Create a fancy looking mesh (example taken from octaviz)."""
    pi = scipy.pi
    cos = scipy.cos
    sin = scipy.sin
    du, dv = pi/20.0, pi/20.0
    u, v = scipy.mgrid[0.01:pi+du*1.5:du, 0:2*pi+dv*1.5:dv]
    x = (1- cos(u))*cos(u+2*pi/3) * cos(v + 2*pi/3.0)*0.5
    y = (1- cos(u))*cos(u+2*pi/3) * cos(v - 2*pi/3.0)*0.5
    z = cos(u-2*pi/3.)

    m = fancy_mesh(x, y, z, z, scalar_visibility=True,
                   tube_radius=0.0075, sphere_radius=0.02)

def test_imshow():
    """Show a large random array."""
    import enthought.util.randomx as RandomArray
    z_large = RandomArray.random((1024, 512))
    i = imshow(z_large)

def test_contour3d():
    dims = [64, 64, 64]
    xmin, xmax, ymin, ymax, zmin, zmax = [-5,5,-5,5,-5,5]
    x, y, z = scipy.ogrid[xmin:xmax:dims[0]*1j,
                          ymin:ymax:dims[1]*1j,
                          zmin:zmax:dims[2]*1j]
    x = x.astype('f')
    y = y.astype('f')
    z = z.astype('f')

    sin = scipy.sin
    scalars = x*x*0.5 + y*y + z*z*2.0

    contour3d(scalars, contours=4, show_slices=True)

    # Show an outline and zoom appropriately.
    outline()
    mayavi.engine.current_scene.scene.isometric_view()
    
def test_quiver3d():
    dims = [16, 16, 16]
    xmin, xmax, ymin, ymax, zmin, zmax = [-5,5,-5,5,-5,5]
    x, y, z = scipy.mgrid[xmin:xmax:dims[0]*1j,
                          ymin:ymax:dims[1]*1j,
                          zmin:zmax:dims[2]*1j]
    x = x.astype('f')
    y = y.astype('f')
    z = z.astype('f')

    sin = scipy.sin
    cos = scipy.cos
    u = cos(x)
    v = sin(y)
    w = sin(x*z)
    
    # All these work!
    #quiver3d(u, v, w)
    quiver3d(x, y, z, u, v, w)

    # Show an outline and zoom appropriately.
    outline()
    mayavi.engine.current_scene.scene.isometric_view()

def test_quiver3d_2d_data():
    dims = [32, 32]
    xmin, xmax, ymin, ymax = [-5,5,-5,5]
    x, y = scipy.mgrid[xmin:xmax:dims[0]*1j,
                       ymin:ymax:dims[1]*1j]
    x = x.astype('f')
    y = y.astype('f')

    sin = scipy.sin
    cos = scipy.cos
    u = cos(x)
    v = sin(y)
    w = scipy.zeros_like(x)
    
    quiver3d(x, y, w, u, v, w)

    # Show an outline and zoom appropriately.
    outline()
    mayavi.engine.current_scene.scene.isometric_view()

From jstrunk at enthought.com  Sun Mar 25 20:52:49 2007
From: jstrunk at enthought.com (Jeff Strunk)
Date: Sun, 25 Mar 2007 19:52:49 -0500
Subject: [SciPy-user] www.scipy.org appears to be down
In-Reply-To: <20070325183259.GA1028@zunzun.com>
References: <20070325183259.GA1028@zunzun.com>
Message-ID: <200703251952.49248.jstrunk@enthought.com>

Thank you. I restarted it at 5:30pm central.

-Jeff

On Sunday 25 March 2007 1:33 pm, zunzun at zunzun.com wrote:
> Quick note: www.scipy.org appears to be down.
>
>      James
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user


From ckkart at hoc.net  Sun Mar 25 21:57:43 2007
From: ckkart at hoc.net (Christian)
Date: Mon, 26 Mar 2007 10:57:43 +0900
Subject: [SciPy-user] odr thread safe?
In-Reply-To: <4601998C.7010404@ee.byu.edu>
References: <etq7vp$ldc$1@sea.gmane.org> <46016E74.9060309@gmail.com>
	<4601998C.7010404@ee.byu.edu>
Message-ID: <eu79a3$sdp$1@sea.gmane.org>

Travis Oliphant wrote:
> 
> In NumPy for the ufuncs, we use PyThreadState_GetDict and then if that 
> is NULL use PyEval_GetBuiltins so that we always get a per-thread 
> dictionary to store information in.

So I should store the members of the odr_global structure in the thread or
builtin dict and read it from there when needed? I am not really experienced
with the Python C API but I always looked for an argument to start digging into
it. Currently however my tests segfault and I cannot compile the extension with
 debugging symbols. Isn't build --debug supposed to do that?

Thanks for your help, Christian



From emilia12 at mail.bg  Mon Mar 26 06:29:35 2007
From: emilia12 at mail.bg (emilia12 at mail.bg)
Date: Mon, 26 Mar 2007 13:29:35 +0300
Subject: [SciPy-user] examples in the scipy Example List
Message-ID: <1174904975.ae8068fc3cd66@mail.bg>

hi,

is it possible to post only working examples in the "scipy
Example List" (http://www.scipy.org/scipy_Example_List)

for eg in "convolve()", variable "h1" is missing !

e.



-----------------------------

SCENA - ???????????? ????????? ???????? ?? ??????? ??????????? ? ??????????.
http://www.bgscena.com/



From raphael.langella at steria.com  Mon Mar 26 09:27:33 2007
From: raphael.langella at steria.com (raphael langella)
Date: Mon, 26 Mar 2007 15:27:33 +0200
Subject: [SciPy-user] building numpy/scipy on Solaris
Message-ID: <10737139e1.139e110737@steria.com>

---- Messages d?origine ----
De: raphael langella <raphael.langella at steria.com>
Date: jeudi, mars 22, 2007 10:39 am
Objet: Re: [SciPy-user] building numpy/scipy on Solaris

> ---- Messages d?origine ----
> De: "David M. Cooke" <cookedm at physics.mcmaster.ca>
> Date: jeudi, mars 22, 2007 9:41 am
> Objet: Re: [SciPy-user] building numpy/scipy on Solaris
> 
> > Ahh, just realized. We can do this completely with the C 
> compiler; the
> > -xlic_lib=sunperf -xarch=v9b is only being used for Fortran. Try
> > 
> > CFLAGS='-xlic_lib=sunperf -xarch=v9b' CPPFLAGS='-
> > DNO_APPEND_FORTRAN' python setup.py install
> > 
> > Since sunperf has C bindings, we don't need the Fortran compiler 
> at 
> > allfor Numpy (and I don't think it was being used in the first 
> place).

> well, options for C don't have the same syntax as Fortran, so I used
> '-mcpu=v9 -lsunperf' and got :
> 
> ImportError: ld.so.1: python: fatal: relocation error: file
> /Produits/sun/forte/7/SUNWspro/lib/libsunperf.so.4: symbol __f95_sign:
> referenced symbol not found

Well, any idea about this error? It would be neat to be able to build
numpy with libsunperf.


From cookedm at physics.mcmaster.ca  Mon Mar 26 09:33:57 2007
From: cookedm at physics.mcmaster.ca (David M. Cooke)
Date: Mon, 26 Mar 2007 09:33:57 -0400
Subject: [SciPy-user] building numpy/scipy on Solaris
In-Reply-To: <10737139e1.139e110737@steria.com>
References: <10737139e1.139e110737@steria.com>
Message-ID: <A7EF01B7-AA1E-4562-8D9E-4C6F25698202@physics.mcmaster.ca>

On Mar 26, 2007, at 09:27 , raphael langella wrote:
> ---- Messages d?origine ----
> De: raphael langella <raphael.langella at steria.com>
> Date: jeudi, mars 22, 2007 10:39 am
> Objet: Re: [SciPy-user] building numpy/scipy on Solaris
>
>> ---- Messages d?origine ----
>> De: "David M. Cooke" <cookedm at physics.mcmaster.ca>
>> Date: jeudi, mars 22, 2007 9:41 am
>> Objet: Re: [SciPy-user] building numpy/scipy on Solaris
>>
>>> Ahh, just realized. We can do this completely with the C
>> compiler; the
>>> -xlic_lib=sunperf -xarch=v9b is only being used for Fortran. Try
>>>
>>> CFLAGS='-xlic_lib=sunperf -xarch=v9b' CPPFLAGS='-
>>> DNO_APPEND_FORTRAN' python setup.py install
>>>
>>> Since sunperf has C bindings, we don't need the Fortran compiler
>> at
>>> allfor Numpy (and I don't think it was being used in the first
>> place).
>
>> well, options for C don't have the same syntax as Fortran, so I used
>> '-mcpu=v9 -lsunperf' and got :
>>
>> ImportError: ld.so.1: python: fatal: relocation error: file
>> /Produits/sun/forte/7/SUNWspro/lib/libsunperf.so.4: symbol  
>> __f95_sign:
>> referenced symbol not found
>
> Well, any idea about this error? It would be neat to be able to build
> numpy with libsunperf.

Looks like you may have to link with the Fortran compiler, or link  
with the appropriate Fortran support libraries. You could try setting  
the LD environment variable to your Fortran compiler, and building  
again.

-- 
|>|\/|<
/------------------------------------------------------------------\
|David M. Cooke              http://arbutus.physics.mcmaster.ca/dmc/
|cookedm at physics.mcmaster.ca

-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 186 bytes
Desc: This is a digitally signed message part
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070326/6c901d03/attachment.sig>

From nowar00 at hotmail.com  Mon Mar 26 09:59:49 2007
From: nowar00 at hotmail.com (javi markez bigara)
Date: 26 Mar 2007 06:59:49 -0700
Subject: [SciPy-user] problems with scipy.gplt.pyPlot
Message-ID: <BAY130-F15B70E49A8D9258D0000BCAA6F0@phx.gbl>

Date: Mon, 26 Mar 2007 13:59:49 +0000
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1; format=flowed

hi, i am new in this type of forums, i am looking for the solution to a 
problem coming from my low knoledge of python. well, so i would like to plot 
a list of coordenates using scipy this is wat i write:

draw=scipy.gplt.pyPlot.Plot()
draw.plot(newlist1)


newlist is  a list of longitude latitude

like  [-4.6925871,45.36987745] each line
i have like 140000 lines like this one. so whenever i run the program
i get a plot but the x axe is the line number
althought i get in the y axe two lines one with the longitude ane other one 
with the latitude  .
i would like to get one point 4 each line
can anybody help me please.......

thank in advance

_________________________________________________________________
Grandes ?xitos, superh?roes, imitaciones, cine y TV... 
http://es.msn.kiwee.com/ Lo mejor para tu m?vil.



From wachao at cse.ohio-state.edu  Mon Mar 26 10:04:38 2007
From: wachao at cse.ohio-state.edu (chao wang)
Date: Mon, 26 Mar 2007 10:04:38 -0400 (EDT)
Subject: [SciPy-user] Two questions about maxentropy module
Message-ID: <Pine.GSO.4.40.0703261003470.26445-100000@kappa.cse.ohio-state.edu>

Good day!

It's nice to see the scipy.maxentropy module. I have two questions about
this module.

Q1) The current model fitting algorithms include CG, BFGS, LBFGSB, Powell,
and Nelder-Mead. How about L-BFGS (unbounded optimization)? Is it
implemented or is there any specific concern of not using it here? Also,
it seems that LBFGSB does not work ...

Q2) bergerexamplesimulated.py example shows how to use bigmodel. However,
it requires the event space to be explicitly enumerated (the enumerated
event space is used when defining a uniform instrumental distribution
for sampling), which is intractable for large event space. E.g., if I
wanna model a maxent distribution over 100 binary variables. What change
should I make to the example code?

Any advice appreciated.

Thanks!



From nowar00 at hotmail.com  Mon Mar 26 11:09:43 2007
From: nowar00 at hotmail.com (javi markez bigara)
Date: Mon, 26 Mar 2007 15:09:43 +0000
Subject: [SciPy-user] FW:  problems with scipy.gplt.pyPlot
Message-ID: <BAY130-F233C3CBECC2A6A751FF63CAA6F0@phx.gbl>




>From: "javi markez bigara" <nowar00 at hotmail.com>
>Reply-To: SciPy Users List <scipy-user at scipy.org>
>To: SciPy-user at scipy.org
>Subject: [SciPy-user] problems with scipy.gplt.pyPlot
>Date: 26 Mar 2007 06:59:49 -0700
>
>Date: Mon, 26 Mar 2007 13:59:49 +0000
>Mime-Version: 1.0
>Content-Type: text/plain; charset=iso-8859-1; format=flowed
>
>hi, i am new in this type of forums, i am looking for the solution to a
>problem coming from my low knoledge of python. well, so i would like to 
>plot
>a list of coordenates using scipy this is wat i write:
>
>draw=scipy.gplt.pyPlot.Plot()
>draw.plot(newlist1)
>
>
>newlist is  a list of longitude latitude
>
>like  [-4.6925871,45.36987745] each line
>i have like 140000 lines like this one. so whenever i run the program
>i get a plot but the x axe is the line number
>althought i get in the y axe two lines one with the longitude ane other one
>with the latitude  .
>i would like to get one point 4 each line
>can anybody help me please.......
>
>thank in advance
>
>_________________________________________________________________
>Grandes ?xitos, superh?roes, imitaciones, cine y TV...
>http://es.msn.kiwee.com/ Lo mejor para tu m?vil.
>

_________________________________________________________________
Descarga gratis la Barra de Herramientas de MSN 
http://www.msn.es/usuario/busqueda/barra?XAPID=2031&DI=1055&SU=http%3A//www.hotmail.com&HL=LINKTAG1OPENINGTEXT_MSNBH
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: text3.txt
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070326/40c11cb1/attachment.txt>

From schaouette at free.fr  Mon Mar 26 11:35:28 2007
From: schaouette at free.fr (Gilles =?utf-8?q?Gr=C3=A9goire?=)
Date: Mon, 26 Mar 2007 17:35:28 +0200
Subject: [SciPy-user] examples in the scipy Example List
In-Reply-To: <1174904975.ae8068fc3cd66@mail.bg>
References: <1174904975.ae8068fc3cd66@mail.bg>
Message-ID: <200703261735.29178.schaouette@free.fr>

Le Lundi 26 Mars 2007 12:29, emilia12 at mail.bg a ?crit?:
> hi,
Hello!
>
> is it possible to post only working examples in the "scipy
> Example List" (http://www.scipy.org/scipy_Example_List)
I changed the examples so that they should all work using copy/paste now.
--Gilles

>
> for eg in "convolve()", variable "h1" is missing !
>
> e.
>
>
>
> -----------------------------
>
> SCENA - ???????????? ????????? ???????? ?? ??????? ??????????? ?
> ??????????. http://www.bgscena.com/
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user

From scipy-user at monte-stello.com  Mon Mar 26 12:22:38 2007
From: scipy-user at monte-stello.com (Antoine Sirinelli)
Date: Mon, 26 Mar 2007 18:22:38 +0200
Subject: [SciPy-user] [Cookbook] FiltFilt does not work
Message-ID: <20070326162238.GA10284@monte-stello.com>

Hi,

If you try to run the FiltFilt script presented in the Cookbook
(http://scipy.org/Cookbook/FiltFilt) you will see that the result is not
the one presented. I have the problem with these versions:
Python: 2.4.4
numpy: 1.0.1
scipy: 0.5.1

Do you experiment the same problem ?

Antoine


From Axel.Kowald at rub.de  Mon Mar 26 12:36:51 2007
From: Axel.Kowald at rub.de (Axel Kowald)
Date: Mon, 26 Mar 2007 18:36:51 +0200
Subject: [SciPy-user] Scipy and Multiple Regression ??
Message-ID: <4607F6A3.5030105@rub.de>

Hi everybody,

I'm new to scipy and wondered it I can also do multiple regression with it ?
I only found routines for normal linear regression.

Thanx,

 Axel


From robert.kern at gmail.com  Mon Mar 26 12:58:32 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Mon, 26 Mar 2007 11:58:32 -0500
Subject: [SciPy-user] Scipy and Multiple Regression ??
In-Reply-To: <4607F6A3.5030105@rub.de>
References: <4607F6A3.5030105@rub.de>
Message-ID: <4607FBB8.9060809@gmail.com>

Axel Kowald wrote:
> Hi everybody,
> 
> I'm new to scipy and wondered it I can also do multiple regression with it ?

What exactly do you mean by "multiple regression"? If you mean linear
least-squares with models of the form

  y = a0 + a1*x + a2*x**2 + ...

or

  y = a0 + a1*x1 + a2*x2 + ...

then scipy.linalg.lstsq() should work for you.

> I only found routines for normal linear regression.

There's scipy.optimize.leastsq() for nonlinear least-squares regression. There's
scipy.odr for nonlinear ordinary least-squares, orthogonal distance regression,
and implicit regression (models of the form f(x)=0), all with as many input
dimensions as you like.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From steveire at gmail.com  Mon Mar 26 12:59:16 2007
From: steveire at gmail.com (Stephen Kelly)
Date: Mon, 26 Mar 2007 17:59:16 +0100
Subject: [SciPy-user] Arrayfns in numpy?
In-Reply-To: <18fbbe5a0703140833o4cc196a4pa653439f60a6bd4e@mail.gmail.com>
References: <18fbbe5a0703051224i5eba702ey8b9ee3b51da8b528@mail.gmail.com>
	<20070305205644.GA18701@virginia.edu>
	<18fbbe5a0703060254v4dd06c17td5d266a1a772ae1e@mail.gmail.com>
	<45ED948A.2040708@ee.byu.edu>
	<18fbbe5a0703060825nd8b9783ib2e9eea18da41645@mail.gmail.com>
	<45ED9A2F.3050803@ee.byu.edu>
	<e76aa17f0703070617j6aeeccf7g19f81896eba2e9d9@mail.gmail.com>
	<45EED6DF.9020502@american.edu> <45EEF5A5.6080202@ee.byu.edu>
	<18fbbe5a0703140833o4cc196a4pa653439f60a6bd4e@mail.gmail.com>
Message-ID: <18fbbe5a0703260959q60c34428r9dc5b62e49a499ae@mail.gmail.com>

Here is the function I'm using after the advice of Alok Singhal above.
Just thought I'd post it in case someone else is looking for similar.

def interp(y_values, x_values, new_x_values):
  """ Perform interpolation similar to deprecated arrayfns.interp in Numeric.

  interp(y, x, z) = y(z) interpolated by treating y(x) as piecewise fcn.
  """
  x = array(x_values, dtype='float')
  y = array(y_values, dtype='float')

  xx = array(new_x_values, dtype='float')
  # High indices
  hi = searchsorted(x, xx)

  # Low indices
  lo = hi - 1

  slopes = (y[hi] - y[lo])/(x[hi] - x[lo])

  # Interpolated data
  yy = y[lo] + slopes*(xx - x[lo])
  return yy

It gives almost identical results for my data, except for points out
of range for which it creates zeros unlike arrayfns.

Thanks again for the help.

Stephen.


From jonathan.taylor at stanford.edu  Mon Mar 26 13:01:21 2007
From: jonathan.taylor at stanford.edu (Jonathan Taylor)
Date: Mon, 26 Mar 2007 13:01:21 -0400
Subject: [SciPy-user] Scipy and Multiple Regression ??
In-Reply-To: <4607F6A3.5030105@rub.de>
References: <4607F6A3.5030105@rub.de>
Message-ID: <4607FC61.6030900@stanford.edu>


There's a package in the sandbox called models that allows you to do
this -- hopefully it will move out of the sandbox soon.

There is also code for:

- generalized linear models
- robust regression (M-estimators)
- Cox proportional hazard
- mixed models
- smoothing splines / generalized additive models

The last two items need work.
Axel Kowald wrote:
> Hi everybody,
>
> I'm new to scipy and wondered it I can also do multiple regression with it ?
> I only found routines for normal linear regression.
>
> Thanx,
>
>  Axel
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   


From lorenzo.isella at gmail.com  Mon Mar 26 13:07:04 2007
From: lorenzo.isella at gmail.com (Lorenzo Isella)
Date: Mon, 26 Mar 2007 19:07:04 +0200
Subject: [SciPy-user] Automatic Indentation
Message-ID: <a2b3004b0703261007v48743615w1c6cfbdd34695f3e@mail.gmail.com>

Dear All,
Again a newbie question: I am postprocessing some data files with Python.
Let us say they are named {filea001...filea100},
{fileb001...fileb100}, {filec001...filec100} and so on.
I need to perform different operations on the three different data
sets, and I wrote a code for that.
The problem is that in the future the three data sets will not be
present simultaneously and I thought about simply defining a parameter
which assumes values {1,2,3} depending on which datasets are present
and using some if-conditions to tell the code which data it should
read and process, without changing dramatically the code structure.
Now, the problem is that I cannot simply encapsulate large parts of
the code into an if condition without also changing the indentation
and this is getting tedious and error prone.
I am sure I cannot be the first Python user to have come across this,
so I would like to know how more experienced users deal with this.
Kind Regards

Lorenzo Isella


From zunzun at zunzun.com  Mon Mar 26 13:37:21 2007
From: zunzun at zunzun.com (zunzun at zunzun.com)
Date: Mon, 26 Mar 2007 13:37:21 -0400
Subject: [SciPy-user] Scipy and Multiple Regression ??
In-Reply-To: <4607FC61.6030900@stanford.edu>
References: <4607F6A3.5030105@rub.de> <4607FC61.6030900@stanford.edu>
Message-ID: <20070326173720.GA28763@zunzun.com>

That will be very, very useful - thanks to all
who are working on it.  I'd sure like to add the
M-estimators to my site.

I'm working on adding covariance matrix estimators
to Robert Kern's existing odr fortran wrappers if
you need them.  This will allow for good standard
fit statistics for nonlinear models when combined
with http://www.scipy.org/Cookbook/OLS

     James Phillips
     http://zunzun.com


On Mon, Mar 26, 2007 at 01:01:21PM -0400, Jonathan Taylor wrote:
> 
> There is also code for:
> 
> - generalized linear models
> - robust regression (M-estimators)
> - Cox proportional hazard


From zunzun at zunzun.com  Mon Mar 26 13:44:41 2007
From: zunzun at zunzun.com (zunzun at zunzun.com)
Date: Mon, 26 Mar 2007 13:44:41 -0400
Subject: [SciPy-user] Automatic Indentation
In-Reply-To: <a2b3004b0703261007v48743615w1c6cfbdd34695f3e@mail.gmail.com>
References: <a2b3004b0703261007v48743615w1c6cfbdd34695f3e@mail.gmail.com>
Message-ID: <20070326174441.GC28763@zunzun.com>

I wound up wrapping my data in an abstracted Data Object class
and changing my code to use that.  Kind of like the Data class
in the ODR code, sorta.  Doing so greatly simplified my
programming as the data is all encapsulated in the object,
and it allows me to change the data handling separately from
the other code.

     James Phillips
     http://zunzun.com


On Mon, Mar 26, 2007 at 07:07:04PM +0200, Lorenzo Isella wrote:
> I need to perform different operations on the three different data
> sets, and I wrote a code for that.
> The problem is that in the future the three data sets will not be
> present simultaneously...


From emilia12 at mail.bg  Mon Mar 26 14:12:35 2007
From: emilia12 at mail.bg (emilia12 at mail.bg)
Date: Mon, 26 Mar 2007 21:12:35 +0300
Subject: [SciPy-user] [Cookbook] FiltFilt does not work
Message-ID: <1174932755.baa896d35359e@mail.bg>



Hi Antoine,

i am with
Python 2.4.4 (#71, Oct 18 2006, 08:34:43) [MSC v.1310 32 bit
(Intel)] on win32
numpy: 1.0.1
scipy: 0.5.1

and my results are very close to curves on the picture from
the example
the difference is only in the noise (the signal have a small
random part, which is unique for each calculation)

e.

-----------------------------

SCENA - ???????????? ????????? ???????? ?? ??????? ??????????? ? ??????????.
http://www.bgscena.com/



From joseph.a.crider at boeing.com  Mon Mar 26 14:24:21 2007
From: joseph.a.crider at boeing.com (Crider, Joseph A)
Date: Mon, 26 Mar 2007 13:24:21 -0500
Subject: [SciPy-user] Automatic Indentation
In-Reply-To: <a2b3004b0703261007v48743615w1c6cfbdd34695f3e@mail.gmail.com>
References: <a2b3004b0703261007v48743615w1c6cfbdd34695f3e@mail.gmail.com>
Message-ID: <A995E69F377A7D42B7E6CF23D3A300B102829775@XCH-SE-1V1.se.nos.boeing.com>

I frequently use jEdit (www.jedit.org) for editing code, and it has a
command on the Edit menu to move indentation of a selected block either
left or right, which has been very useful for me in similar situations.
I've encountered similar features on some other editors, but I can't
recall which ones right off.


J. Allen Crider

-----Original Message-----
From: Lorenzo Isella [mailto:lorenzo.isella at gmail.com] 
Sent: Monday, March 26, 2007 12:07 PM
To: scipy-user at scipy.org
Subject: [SciPy-user] Automatic Indentation

Dear All,
Again a newbie question: I am postprocessing some data files with
Python.
Let us say they are named {filea001...filea100}, {fileb001...fileb100},
{filec001...filec100} and so on.
I need to perform different operations on the three different data sets,
and I wrote a code for that.
The problem is that in the future the three data sets will not be
present simultaneously and I thought about simply defining a parameter
which assumes values {1,2,3} depending on which datasets are present and
using some if-conditions to tell the code which data it should read and
process, without changing dramatically the code structure.
Now, the problem is that I cannot simply encapsulate large parts of the
code into an if condition without also changing the indentation and this
is getting tedious and error prone.
I am sure I cannot be the first Python user to have come across this, so
I would like to know how more experienced users deal with this.
Kind Regards

Lorenzo Isella
_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user


From elcorto at gmx.net  Mon Mar 26 15:41:12 2007
From: elcorto at gmx.net (Steve Schmerler)
Date: Mon, 26 Mar 2007 21:41:12 +0200
Subject: [SciPy-user] Automatic Indentation
In-Reply-To: <A995E69F377A7D42B7E6CF23D3A300B102829775@XCH-SE-1V1.se.nos.boeing.com>
References: <a2b3004b0703261007v48743615w1c6cfbdd34695f3e@mail.gmail.com>
	<A995E69F377A7D42B7E6CF23D3A300B102829775@XCH-SE-1V1.se.nos.boeing.com>
Message-ID: <460821D8.40309@gmx.net>

Crider, Joseph A wrote:
> I frequently use jEdit (www.jedit.org) for editing code, and it has a
> command on the Edit menu to move indentation of a selected block either
> left or right, which has been very useful for me in similar situations.
> I've encountered similar features on some other editors, but I can't
> recall which ones right off.

Almost every text editor offers some functionallity of this kind. I
favour vim and ther it is CTRL-v for "visual block mode", then mark your
block and > for moving the block 'shiftwidth' characters to the right,
which should be equal to your tab-setting (set shiftwidth=4, set
tabstop=4 in my case).

Anyway if you have to indent very large code chunks, you're probably
writing  "spaghetti code" and should consider packing parts of it in
functions and classes to keep the clarity :)

-- 
cheers,
steve

I love deadlines. I like the whooshing sound they make as they fly by.
-- Douglas Adams



From scipy-user at monte-stello.com  Mon Mar 26 16:48:55 2007
From: scipy-user at monte-stello.com (Antoine Sirinelli)
Date: Mon, 26 Mar 2007 21:48:55 +0100
Subject: [SciPy-user] [Cookbook] FiltFilt does not work
In-Reply-To: <1174932755.baa896d35359e@mail.bg>
References: <1174932755.baa896d35359e@mail.bg>
Message-ID: <20070326204851.GB15905@localhost.localdomain>

On Mon, Mar 26, 2007 at 09:12:35PM +0300, emilia12 at mail.bg wrote:
> and my results are very close to curves on the picture from
> the example
> the difference is only in the noise (the signal have a small
> random part, which is unique for each calculation)

I agree that the random part have to be different but as you can see on
the figure attached, the last points show a sligtly big difference. The
filtered signal goes away from the noisy signal.

Antoine

-------------- next part --------------
A non-text attachment was scrubbed...
Name: filtfilt.png
Type: image/png
Size: 32066 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070326/b8dd682f/attachment.png>

From dbecker at alum.dartmouth.org  Mon Mar 26 21:13:35 2007
From: dbecker at alum.dartmouth.org (Dan Becker)
Date: Tue, 27 Mar 2007 01:13:35 +0000 (UTC)
Subject: [SciPy-user] Finding lots of roots (probably with brentq)
Message-ID: <loom.20070327T025954-315@post.gmane.org>

Hi,

I have a single equation whose roots I would like to find with many
sets of arguments.  I can do this by iterating over the sets of arguments, but
that doesn't feel very elegant.  I'd rather send the arrays or a matrix of
arguments to brentq once.  To be concrete, consider this code:

---
import time
from scipy.optimize import brentq
from numpy import ones

def tominimize(x,y):
    return x**2-y

t2=time.time()
for thisArg in xrange(1000):
    c=brentq(tominimize,0,100,thisArg)
print(time.time()-t2)


I wonder if this might be faster (and prettier) if I could call brentq once with
a whole set of arguments as in

brentq(tominimize,0,100,[thisArg for thisArg in xrange(1000))

I haven't figured out how to do this without getting type errors.

Thanks for any help!
Dan



From Axel.Kowald at rub.de  Tue Mar 27 03:32:56 2007
From: Axel.Kowald at rub.de (Axel Kowald)
Date: Tue, 27 Mar 2007 09:32:56 +0200
Subject: [SciPy-user] Scipy and Multiple Regression ??
In-Reply-To: <mailman.9.1174928407.11350.scipy-user@scipy.org>
References: <mailman.9.1174928407.11350.scipy-user@scipy.org>
Message-ID: <4608C8A8.5060703@rub.de>

Hi,

thanks for the quick reply. Yes, I want to do linear least-squares and 
it seems scipy.linalg.lstsq() is what I'm looking for.

Many thanks,

       axel

> What exactly do you mean by "multiple regression"? If you mean linear
> least-squares with models of the form
>
>   y = a0 + a1*x + a2*x**2 + ...
>
> or
>
>   y = a0 + a1*x1 + a2*x2 + ...
>
> then scipy.linalg.lstsq() should work for you.


From nowar00 at hotmail.com  Tue Mar 27 08:22:15 2007
From: nowar00 at hotmail.com (javi markez bigara)
Date: Tue, 27 Mar 2007 12:22:15 +0000
Subject: [SciPy-user] plotting with scipy
Message-ID: <BAY130-F75A9DF69A34FCE0D02C0DAA6E0@phx.gbl>

hi,

i am looking for the easy way to plot some coordinates with 
scipy.gplt.pyPlot this is how i try it but is not working as i would like 
to....
the new list is a list of pair coordinates something like this:

[-6.654644,65.55555]
[-6.645654,65.98798]
[-6.987813,65.28951]
[-6.645654,65.65987]
[-6.634895,65.36985]

>>>import scipy.plt
>>>from scipy import gplt
>>>draw=scipy.gplt.pyPlot.Plot()
>>>draw.plot(newlist)
>>>draw.title('little map')
>>>draw.xtitle('longitude')
>>>draw.ytitle('latitude')

the thing is that i get in the x axes the number of list and then two strait 
lines


65_________________________________
50
35
20
5
-10-------------------------------------------------
0      1          2         3         4       5     ............

this is wat it plots more or less
can anybody help me.
thanks in advance
Nowar00

_________________________________________________________________
Descarga gratis la Barra de Herramientas de MSN 
http://www.msn.es/usuario/busqueda/barra?XAPID=2031&DI=1055&SU=http%3A//www.hotmail.com&HL=LINKTAG1OPENINGTEXT_MSNBH



From ryanlists at gmail.com  Tue Mar 27 08:30:48 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Tue, 27 Mar 2007 07:30:48 -0500
Subject: [SciPy-user] plotting with scipy
In-Reply-To: <BAY130-F75A9DF69A34FCE0D02C0DAA6E0@phx.gbl>
References: <BAY130-F75A9DF69A34FCE0D02C0DAA6E0@phx.gbl>
Message-ID: <c5b438120703270530p505df831jd8817cce5f69042b@mail.gmail.com>

I think almost all users of scipy use matplotlib instead of scipy's
plotting capabilities.  It is really a great package especially if you
only need 2D plotting.  Check out matplotlib.sourceforge.net.

Ryan

On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> hi,
>
> i am looking for the easy way to plot some coordinates with
> scipy.gplt.pyPlot this is how i try it but is not working as i would like
> to....
> the new list is a list of pair coordinates something like this:
>
> [-6.654644,65.55555]
> [-6.645654,65.98798]
> [-6.987813,65.28951]
> [-6.645654,65.65987]
> [-6.634895,65.36985]
>
> >>>import scipy.plt
> >>>from scipy import gplt
> >>>draw=scipy.gplt.pyPlot.Plot()
> >>>draw.plot(newlist)
> >>>draw.title('little map')
> >>>draw.xtitle('longitude')
> >>>draw.ytitle('latitude')
>
> the thing is that i get in the x axes the number of list and then two strait
> lines
>
>
> 65_________________________________
> 50
> 35
> 20
> 5
> -10-------------------------------------------------
> 0      1          2         3         4       5     ............
>
> this is wat it plots more or less
> can anybody help me.
> thanks in advance
> Nowar00
>
> _________________________________________________________________
> Descarga gratis la Barra de Herramientas de MSN
> http://www.msn.es/usuario/busqueda/barra?XAPID=2031&DI=1055&SU=http%3A//www.hotmail.com&HL=LINKTAG1OPENINGTEXT_MSNBH
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From nowar00 at hotmail.com  Tue Mar 27 08:51:11 2007
From: nowar00 at hotmail.com (javi markez bigara)
Date: Tue, 27 Mar 2007 12:51:11 +0000
Subject: [SciPy-user] plotting with scipy
Message-ID: <BAY130-F1A1F9A0058CCF2B0AA1DBAA6E0@phx.gbl>




>From: "javi markez bigara" <nowar00 at hotmail.com>
>Reply-To: SciPy Users List <scipy-user at scipy.org>
>To: SciPy-user at scipy.org
>Subject: [SciPy-user] plotting with scipy
>Date: Tue, 27 Mar 2007 12:22:15 +0000
>
>hi,
>
>i am looking for the easy way to plot some coordinates with
>scipy.gplt.pyPlot this is how i try it but is not working as i would like
>to....
>the new list is a list of pair coordinates something like this:
>
>[-6.654644,65.55555]
>[-6.645654,65.98798]
>[-6.987813,65.28951]
>[-6.645654,65.65987]
>[-6.634895,65.36985]
>
> >>>import scipy.plt
> >>>from scipy import gplt
> >>>draw=scipy.gplt.pyPlot.Plot()
> >>>draw.plot(newlist)
> >>>draw.title('little map')
> >>>draw.xtitle('longitude')
> >>>draw.ytitle('latitude')
>
>the thing is that i get in the x axes the number of list and then two 
>strait
>lines
>
>
>65_________________________________
>50
>35
>20
>5
>-10-------------------------------------------------
>0      1          2         3         4       5     ............
>(i would like to get a point for each line.......)any sugestions
>this is wat it plots more or less
>can anybody help me.
>thanks in advance
>Nowar00
>
>_________________________________________________________________
>Descarga gratis la Barra de Herramientas de MSN
>http://www.msn.es/usuario/busqueda/barra?XAPID=2031&DI=1055&SU=http%3A//www.hotmail.com&HL=LINKTAG1OPENINGTEXT_MSNBH
>
>_______________________________________________
>SciPy-user mailing list
>SciPy-user at scipy.org
>http://projects.scipy.org/mailman/listinfo/scipy-user

_________________________________________________________________
Dale rienda suelta a tu tiempo libre. Mil ideas para exprimir tu ocio con 
MSN Entretenimiento. http://entretenimiento.msn.es/



From raphael.langella at steria.com  Tue Mar 27 09:21:54 2007
From: raphael.langella at steria.com (raphael langella)
Date: Tue, 27 Mar 2007 15:21:54 +0200
Subject: [SciPy-user] building numpy/scipy on Solaris
Message-ID: <d8d6e83b.e83bd8d6@steria.com>

OK, I tried so many different ways of compiling numpy with libsunperf, I
think I'm heading nowhere.
Let's try with Atlas. It's much more documented and tested.
I get this when compiling numpy :

gcc -shared
build/temp.solaris-2.8-sun4u-2.5/numpy/core/blasdot/_dotblas.o
-L/tmp/ATLAS/lib/SunOS_SunUSIII_8 -llapack -lf77blas -lcblas -latlas -o
build/lib.solaris-2.8-sun4u-2.5/numpy/core/_dotblas.so
Text relocation remains                         referenced
    against symbol                  offset      in file
ATL_zupKBmm32_8_1_b0                0x514      
/tmp/ATLAS/lib/SunOS_SunUSIII_8/libatlas.a(ATL_zupKBmm_b0.o)
...
(hundreds of similar lines)
...
ATL_zaxpby_aX_bXi0                  0x1a8      
/tmp/ATLAS/lib/SunOS_SunUSIII_8/libatlas.a(ATL_zaxpby.o)
ld: fatal: relocations remain against allocatable but non-writable sections
collect2: ld returned 1 exit status
error: Command "gcc -shared
build/temp.solaris-2.8-sun4u-2.5/numpy/core/blasdot/_dotblas.o
-L/tmp/ATLAS/lib/SunOS_SunUSIII_8 -llapack -lf77blas -lcblas -latlas -o
build/lib.solaris-2.8-sun4u-2.5/numpy/core/_dotblas.so" failed with exit
status 1

by the way, my site.cfg looks like this :
[DEFAULT]
library_dirs = /tmp/ATLAS/lib/SunOS_SunUSIII_8
include_dirs = /tmp/ATLAS/include/SunOS_SunUSIII_8

[atlas]
atlas_libs = lapack, f77blas, cblas, atlas

I also tried to add search_static_first = 0
but it didn't change anything. So, I converted all the atlas library to
dynamic versions and tried again :

building 'numpy.core._dotblas' extension
compiling C sources
C compiler: gcc -fno-strict-aliasing -DNDEBUG -g -O3 -Wall
-Wstrict-prototypes -fPIC

creating build/temp.solaris-2.8-sun4u-2.5/numpy/core/blasdot
compile options: '-DNO_ATLAS_INFO=2 -Inumpy/core/blasdot
-Inumpy/core/include -Ibuild/src.solaris-2.8-sun4u-2.5/numpy/core
-Inumpy/core/src -Inumpy/core/include
-I/Produits/publics/sparc.SunOS.5.8/python/2.5.0/include/python2.5 -c'
gcc: numpy/core/blasdot/_dotblas.c
gcc -shared
build/temp.solaris-2.8-sun4u-2.5/numpy/core/blasdot/_dotblas.o
-L/tmp/ATLAS/lib/SunOS_SunUSIII_8 -llapack -lf77blas -lcblas -latlas -o
build/lib.solaris-2.8-sun4u-2.5/numpy/core/_dotblas.so
ld: fatal: symbol `lsame_' is multiply-defined:
        (file /tmp/ATLAS/lib/SunOS_SunUSIII_8/liblapack.so type=FUNC;
file /tmp/ATLAS/lib/SunOS_SunUSIII_8/libf77blas.so type=FUNC);
ld: fatal: symbol `xerbla_' is multiply-defined:
        (file /tmp/ATLAS/lib/SunOS_SunUSIII_8/liblapack.so type=FUNC;
file /tmp/ATLAS/lib/SunOS_SunUSIII_8/libf77blas.so type=FUNC);
ld: fatal: File processing errors. No output written to
build/lib.solaris-2.8-sun4u-2.5/numpy/core/_dotblas.so
collect2: ld returned 1 exit status
ld: fatal: symbol `lsame_' is multiply-defined:
        (file /tmp/ATLAS/lib/SunOS_SunUSIII_8/liblapack.so type=FUNC;
file /tmp/ATLAS/lib/SunOS_SunUSIII_8/libf77blas.so type=FUNC);
ld: fatal: symbol `xerbla_' is multiply-defined:
        (file /tmp/ATLAS/lib/SunOS_SunUSIII_8/liblapack.so type=FUNC;
file /tmp/ATLAS/lib/SunOS_SunUSIII_8/libf77blas.so type=FUNC);
ld: fatal: File processing errors. No output written to
build/lib.solaris-2.8-sun4u-2.5/numpy/core/_dotblas.so
collect2: ld returned 1 exit status
error: Command "gcc -shared
build/temp.solaris-2.8-sun4u-2.5/numpy/core/blasdot/_dotblas.o
-L/tmp/ATLAS/lib/SunOS_SunUSIII_8 -llapack -lf77blas -lcblas -latlas -o
build/lib.solaris-2.8-sun4u-2.5/numpy/core/_dotblas.so" failed with exit
status 1

I tried with both GNU and Sun ld with same result.
Any idea ?
-------------- next part --------------
_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user

From nowar00 at hotmail.com  Tue Mar 27 09:22:27 2007
From: nowar00 at hotmail.com (javi markez bigara)
Date: Tue, 27 Mar 2007 13:22:27 +0000
Subject: [SciPy-user] plotting with scipy
In-Reply-To: <c5b438120703270530p505df831jd8817cce5f69042b@mail.gmail.com>
Message-ID: <BAY130-F40B834B3CF13EE0908473AAA6E0@phx.gbl>


thanks a lot for ur reply i will download it but the thig is that i dont 
know were to put it them to import it from python.....
thanks a lot Ryan
nowar

>From: "Ryan Krauss" <ryanlists at gmail.com>
>Reply-To: SciPy Users List <scipy-user at scipy.org>
>To: "SciPy Users List" <scipy-user at scipy.org>
>Subject: Re: [SciPy-user] plotting with scipy
>Date: Tue, 27 Mar 2007 07:30:48 -0500
>
>I think almost all users of scipy use matplotlib instead of scipy's
>plotting capabilities.  It is really a great package especially if you
>only need 2D plotting.  Check out matplotlib.sourceforge.net.
>
>Ryan
>
>On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > hi,
> >
> > i am looking for the easy way to plot some coordinates with
> > scipy.gplt.pyPlot this is how i try it but is not working as i would 
>like
> > to....
> > the new list is a list of pair coordinates something like this:
> >
> > [-6.654644,65.55555]
> > [-6.645654,65.98798]
> > [-6.987813,65.28951]
> > [-6.645654,65.65987]
> > [-6.634895,65.36985]
> >
> > >>>import scipy.plt
> > >>>from scipy import gplt
> > >>>draw=scipy.gplt.pyPlot.Plot()
> > >>>draw.plot(newlist)
> > >>>draw.title('little map')
> > >>>draw.xtitle('longitude')
> > >>>draw.ytitle('latitude')
> >
> > the thing is that i get in the x axes the number of list and then two 
>strait
> > lines
> >
> >
> > 65_________________________________
> > 50
> > 35
> > 20
> > 5
> > -10-------------------------------------------------
> > 0      1          2         3         4       5     ............
> >
> > this is wat it plots more or less
> > can anybody help me.
> > thanks in advance
> > Nowar00
> >
> > _________________________________________________________________
> > Descarga gratis la Barra de Herramientas de MSN
> > 
>http://www.msn.es/usuario/busqueda/barra?XAPID=2031&DI=1055&SU=http%3A//www.hotmail.com&HL=LINKTAG1OPENINGTEXT_MSNBH
> >
> > _______________________________________________
> > SciPy-user mailing list
> > SciPy-user at scipy.org
> > http://projects.scipy.org/mailman/listinfo/scipy-user
> >
>_______________________________________________
>SciPy-user mailing list
>SciPy-user at scipy.org
>http://projects.scipy.org/mailman/listinfo/scipy-user

_________________________________________________________________
Acepta el reto MSN Premium: Correos m?s divertidos con fotos y textos 
incre?bles en MSN Premium. Desc?rgalo y pru?balo 2 meses gratis. 
http://join.msn.com?XAPID=1697&DI=1055&HL=Footer_mailsenviados_correosmasdivertidos



From ryanlists at gmail.com  Tue Mar 27 09:37:52 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Tue, 27 Mar 2007 08:37:52 -0500
Subject: [SciPy-user] plotting with scipy
In-Reply-To: <BAY130-F40B834B3CF13EE0908473AAA6E0@phx.gbl>
References: <c5b438120703270530p505df831jd8817cce5f69042b@mail.gmail.com>
	<BAY130-F40B834B3CF13EE0908473AAA6E0@phx.gbl>
Message-ID: <c5b438120703270637o58443bd7w9ef83397717cc1c9@mail.gmail.com>

What platform are you running?  There should be an executable for
windows that puts them in the proper directory.  They can go anywhere
on your PYTHONPATH, but the right answer is in the directory
C:\Python24\Lib\site-packages
(or I presume Python25 if you are running the current version of python).

So, I have the folders
C:\Python24\Lib\site-packages\matplotlib
C:\Python24\Lib\site-packages\scipy
etc.

On linux, you usually want to unzip the tarball and there should be a
setup.py file in the top level.  Then just run
sudo python setup.py install
(or similar if you aren't on a debian based "sudo" distribution).

Let me know if you still have trouble.

Ryan

On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
>
> thanks a lot for ur reply i will download it but the thig is that i dont
> know were to put it them to import it from python.....
> thanks a lot Ryan
> nowar
>
> >From: "Ryan Krauss" <ryanlists at gmail.com>
> >Reply-To: SciPy Users List <scipy-user at scipy.org>
> >To: "SciPy Users List" <scipy-user at scipy.org>
> >Subject: Re: [SciPy-user] plotting with scipy
> >Date: Tue, 27 Mar 2007 07:30:48 -0500
> >
> >I think almost all users of scipy use matplotlib instead of scipy's
> >plotting capabilities.  It is really a great package especially if you
> >only need 2D plotting.  Check out matplotlib.sourceforge.net.
> >
> >Ryan
> >
> >On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > > hi,
> > >
> > > i am looking for the easy way to plot some coordinates with
> > > scipy.gplt.pyPlot this is how i try it but is not working as i would
> >like
> > > to....
> > > the new list is a list of pair coordinates something like this:
> > >
> > > [-6.654644,65.55555]
> > > [-6.645654,65.98798]
> > > [-6.987813,65.28951]
> > > [-6.645654,65.65987]
> > > [-6.634895,65.36985]
> > >
> > > >>>import scipy.plt
> > > >>>from scipy import gplt
> > > >>>draw=scipy.gplt.pyPlot.Plot()
> > > >>>draw.plot(newlist)
> > > >>>draw.title('little map')
> > > >>>draw.xtitle('longitude')
> > > >>>draw.ytitle('latitude')
> > >
> > > the thing is that i get in the x axes the number of list and then two
> >strait
> > > lines
> > >
> > >
> > > 65_________________________________
> > > 50
> > > 35
> > > 20
> > > 5
> > > -10-------------------------------------------------
> > > 0      1          2         3         4       5     ............
> > >
> > > this is wat it plots more or less
> > > can anybody help me.
> > > thanks in advance
> > > Nowar00
> > >
> > > _________________________________________________________________
> > > Descarga gratis la Barra de Herramientas de MSN
> > >
> >http://www.msn.es/usuario/busqueda/barra?XAPID=2031&DI=1055&SU=http%3A//www.hotmail.com&HL=LINKTAG1OPENINGTEXT_MSNBH
> > >
> > > _______________________________________________
> > > SciPy-user mailing list
> > > SciPy-user at scipy.org
> > > http://projects.scipy.org/mailman/listinfo/scipy-user
> > >
> >_______________________________________________
> >SciPy-user mailing list
> >SciPy-user at scipy.org
> >http://projects.scipy.org/mailman/listinfo/scipy-user
>
> _________________________________________________________________
> Acepta el reto MSN Premium: Correos m?s divertidos con fotos y textos
> incre?bles en MSN Premium. Desc?rgalo y pru?balo 2 meses gratis.
> http://join.msn.com?XAPID=1697&DI=1055&HL=Footer_mailsenviados_correosmasdivertidos
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From nowar00 at hotmail.com  Tue Mar 27 09:48:01 2007
From: nowar00 at hotmail.com (javi markez bigara)
Date: Tue, 27 Mar 2007 13:48:01 +0000
Subject: [SciPy-user] plotting with scipy
In-Reply-To: <c5b438120703270637o58443bd7w9ef83397717cc1c9@mail.gmail.com>
Message-ID: <BAY130-F40EF114A184B2B02F117C4AA6E0@phx.gbl>


this is the python i am using, i download the mathplotlib but i think is not 
working
i just save it in the Desktop.. and this is the forder i get:
matplotlib-0.90.0

Python 2.3.5 (#2, Sep  4 2005, 22:01:42)
[GCC 3.3.5 (Debian 1:3.3.5-13)] on linux2


>From: "Ryan Krauss" <ryanlists at gmail.com>
>Reply-To: SciPy Users List <scipy-user at scipy.org>
>To: "SciPy Users List" <scipy-user at scipy.org>
>Subject: Re: [SciPy-user] plotting with scipy
>Date: Tue, 27 Mar 2007 08:37:52 -0500
>
>What platform are you running?  There should be an executable for
>windows that puts them in the proper directory.  They can go anywhere
>on your PYTHONPATH, but the right answer is in the directory
>C:\Python24\Lib\site-packages
>(or I presume Python25 if you are running the current version of python).
>
>So, I have the folders
>C:\Python24\Lib\site-packages\matplotlib
>C:\Python24\Lib\site-packages\scipy
>etc.
>
>On linux, you usually want to unzip the tarball and there should be a
>setup.py file in the top level.  Then just run
>sudo python setup.py install
>(or similar if you aren't on a debian based "sudo" distribution).
>
>Let me know if you still have trouble.
>
>Ryan
>
>On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> >
> > thanks a lot for ur reply i will download it but the thig is that i dont
> > know were to put it them to import it from python.....
> > thanks a lot Ryan
> > nowar
> >
> > >From: "Ryan Krauss" <ryanlists at gmail.com>
> > >Reply-To: SciPy Users List <scipy-user at scipy.org>
> > >To: "SciPy Users List" <scipy-user at scipy.org>
> > >Subject: Re: [SciPy-user] plotting with scipy
> > >Date: Tue, 27 Mar 2007 07:30:48 -0500
> > >
> > >I think almost all users of scipy use matplotlib instead of scipy's
> > >plotting capabilities.  It is really a great package especially if you
> > >only need 2D plotting.  Check out matplotlib.sourceforge.net.
> > >
> > >Ryan
> > >
> > >On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > > > hi,
> > > >
> > > > i am looking for the easy way to plot some coordinates with
> > > > scipy.gplt.pyPlot this is how i try it but is not working as i would
> > >like
> > > > to....
> > > > the new list is a list of pair coordinates something like this:
> > > >
> > > > [-6.654644,65.55555]
> > > > [-6.645654,65.98798]
> > > > [-6.987813,65.28951]
> > > > [-6.645654,65.65987]
> > > > [-6.634895,65.36985]
> > > >
> > > > >>>import scipy.plt
> > > > >>>from scipy import gplt
> > > > >>>draw=scipy.gplt.pyPlot.Plot()
> > > > >>>draw.plot(newlist)
> > > > >>>draw.title('little map')
> > > > >>>draw.xtitle('longitude')
> > > > >>>draw.ytitle('latitude')
> > > >
> > > > the thing is that i get in the x axes the number of list and then 
>two
> > >strait
> > > > lines
> > > >
> > > >
> > > > 65_________________________________
> > > > 50
> > > > 35
> > > > 20
> > > > 5
> > > > -10-------------------------------------------------
> > > > 0      1          2         3         4       5     ............
> > > >
> > > > this is wat it plots more or less
> > > > can anybody help me.
> > > > thanks in advance
> > > > Nowar00
> > > >
> > > > _________________________________________________________________
> > > > Descarga gratis la Barra de Herramientas de MSN
> > > >
> > 
> >http://www.msn.es/usuario/busqueda/barra?XAPID=2031&DI=1055&SU=http%3A//www.hotmail.com&HL=LINKTAG1OPENINGTEXT_MSNBH
> > > >
> > > > _______________________________________________
> > > > SciPy-user mailing list
> > > > SciPy-user at scipy.org
> > > > http://projects.scipy.org/mailman/listinfo/scipy-user
> > > >
> > >_______________________________________________
> > >SciPy-user mailing list
> > >SciPy-user at scipy.org
> > >http://projects.scipy.org/mailman/listinfo/scipy-user
> >
> > _________________________________________________________________
> > Acepta el reto MSN Premium: Correos m?s divertidos con fotos y textos
> > incre?bles en MSN Premium. Desc?rgalo y pru?balo 2 meses gratis.
> > 
>http://join.msn.com?XAPID=1697&DI=1055&HL=Footer_mailsenviados_correosmasdivertidos
> >
> > _______________________________________________
> > SciPy-user mailing list
> > SciPy-user at scipy.org
> > http://projects.scipy.org/mailman/listinfo/scipy-user
> >
>_______________________________________________
>SciPy-user mailing list
>SciPy-user at scipy.org
>http://projects.scipy.org/mailman/listinfo/scipy-user

_________________________________________________________________
Grandes ?xitos, superh?roes, imitaciones, cine y TV... 
http://es.msn.kiwee.com/ Lo mejor para tu m?vil.



From nowar00 at hotmail.com  Tue Mar 27 09:54:57 2007
From: nowar00 at hotmail.com (javi markez bigara)
Date: Tue, 27 Mar 2007 13:54:57 +0000
Subject: [SciPy-user] plotting with scipy
In-Reply-To: <BAY130-F40EF114A184B2B02F117C4AA6E0@phx.gbl>
Message-ID: <BAY130-F36FC9E1C7E062A710D938AA6E0@phx.gbl>

this is where the folder is
/home2/diazdeim/matplotlib-0.90.0
i want to run the setup from the terminal but i dont know how to do it
thaks for helping


>From: "javi markez bigara" <nowar00 at hotmail.com>
>Reply-To: SciPy Users List <scipy-user at scipy.org>
>To: scipy-user at scipy.org
>Subject: Re: [SciPy-user] plotting with scipy
>Date: Tue, 27 Mar 2007 13:48:01 +0000
>
>
>this is the python i am using, i download the mathplotlib but i think is 
>not
>working
>i just save it in the Desktop.. and this is the forder i get:
>matplotlib-0.90.0
>
>Python 2.3.5 (#2, Sep  4 2005, 22:01:42)
>[GCC 3.3.5 (Debian 1:3.3.5-13)] on linux2
>
>
> >From: "Ryan Krauss" <ryanlists at gmail.com>
> >Reply-To: SciPy Users List <scipy-user at scipy.org>
> >To: "SciPy Users List" <scipy-user at scipy.org>
> >Subject: Re: [SciPy-user] plotting with scipy
> >Date: Tue, 27 Mar 2007 08:37:52 -0500
> >
> >What platform are you running?  There should be an executable for
> >windows that puts them in the proper directory.  They can go anywhere
> >on your PYTHONPATH, but the right answer is in the directory
> >C:\Python24\Lib\site-packages
> >(or I presume Python25 if you are running the current version of python).
> >
> >So, I have the folders
> >C:\Python24\Lib\site-packages\matplotlib
> >C:\Python24\Lib\site-packages\scipy
> >etc.
> >
> >On linux, you usually want to unzip the tarball and there should be a
> >setup.py file in the top level.  Then just run
> >sudo python setup.py install
> >(or similar if you aren't on a debian based "sudo" distribution).
> >
> >Let me know if you still have trouble.
> >
> >Ryan
> >
> >On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > >
> > > thanks a lot for ur reply i will download it but the thig is that i 
>dont
> > > know were to put it them to import it from python.....
> > > thanks a lot Ryan
> > > nowar
> > >
> > > >From: "Ryan Krauss" <ryanlists at gmail.com>
> > > >Reply-To: SciPy Users List <scipy-user at scipy.org>
> > > >To: "SciPy Users List" <scipy-user at scipy.org>
> > > >Subject: Re: [SciPy-user] plotting with scipy
> > > >Date: Tue, 27 Mar 2007 07:30:48 -0500
> > > >
> > > >I think almost all users of scipy use matplotlib instead of scipy's
> > > >plotting capabilities.  It is really a great package especially if 
>you
> > > >only need 2D plotting.  Check out matplotlib.sourceforge.net.
> > > >
> > > >Ryan
> > > >
> > > >On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > > > > hi,
> > > > >
> > > > > i am looking for the easy way to plot some coordinates with
> > > > > scipy.gplt.pyPlot this is how i try it but is not working as i 
>would
> > > >like
> > > > > to....
> > > > > the new list is a list of pair coordinates something like this:
> > > > >
> > > > > [-6.654644,65.55555]
> > > > > [-6.645654,65.98798]
> > > > > [-6.987813,65.28951]
> > > > > [-6.645654,65.65987]
> > > > > [-6.634895,65.36985]
> > > > >
> > > > > >>>import scipy.plt
> > > > > >>>from scipy import gplt
> > > > > >>>draw=scipy.gplt.pyPlot.Plot()
> > > > > >>>draw.plot(newlist)
> > > > > >>>draw.title('little map')
> > > > > >>>draw.xtitle('longitude')
> > > > > >>>draw.ytitle('latitude')
> > > > >
> > > > > the thing is that i get in the x axes the number of list and then
> >two
> > > >strait
> > > > > lines
> > > > >
> > > > >
> > > > > 65_________________________________
> > > > > 50
> > > > > 35
> > > > > 20
> > > > > 5
> > > > > -10-------------------------------------------------
> > > > > 0      1          2         3         4       5     ............
> > > > >
> > > > > this is wat it plots more or less
> > > > > can anybody help me.
> > > > > thanks in advance
> > > > > Nowar00
> > > > >
> > > > > _________________________________________________________________
> > > > > Descarga gratis la Barra de Herramientas de MSN
> > > > >
> > >
> > 
> >http://www.msn.es/usuario/busqueda/barra?XAPID=2031&DI=1055&SU=http%3A//www.hotmail.com&HL=LINKTAG1OPENINGTEXT_MSNBH
> > > > >
> > > > > _______________________________________________
> > > > > SciPy-user mailing list
> > > > > SciPy-user at scipy.org
> > > > > http://projects.scipy.org/mailman/listinfo/scipy-user
> > > > >
> > > >_______________________________________________
> > > >SciPy-user mailing list
> > > >SciPy-user at scipy.org
> > > >http://projects.scipy.org/mailman/listinfo/scipy-user
> > >
> > > _________________________________________________________________
> > > Acepta el reto MSN Premium: Correos m?s divertidos con fotos y textos
> > > incre?bles en MSN Premium. Desc?rgalo y pru?balo 2 meses gratis.
> > >
> >http://join.msn.com?XAPID=1697&DI=1055&HL=Footer_mailsenviados_correosmasdivertidos
> > >
> > > _______________________________________________
> > > SciPy-user mailing list
> > > SciPy-user at scipy.org
> > > http://projects.scipy.org/mailman/listinfo/scipy-user
> > >
> >_______________________________________________
> >SciPy-user mailing list
> >SciPy-user at scipy.org
> >http://projects.scipy.org/mailman/listinfo/scipy-user
>
>_________________________________________________________________
>Grandes ?xitos, superh?roes, imitaciones, cine y TV...
>http://es.msn.kiwee.com/ Lo mejor para tu m?vil.
>
>_______________________________________________
>SciPy-user mailing list
>SciPy-user at scipy.org
>http://projects.scipy.org/mailman/listinfo/scipy-user

_________________________________________________________________
Acepta el reto MSN Premium: Correos m?s divertidos con fotos y textos 
incre?bles en MSN Premium. Desc?rgalo y pru?balo 2 meses gratis. 
http://join.msn.com?XAPID=1697&DI=1055&HL=Footer_mailsenviados_correosmasdivertidos



From ryanlists at gmail.com  Tue Mar 27 09:57:18 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Tue, 27 Mar 2007 08:57:18 -0500
Subject: [SciPy-user] plotting with scipy
In-Reply-To: <BAY130-F36FC9E1C7E062A710D938AA6E0@phx.gbl>
References: <BAY130-F40EF114A184B2B02F117C4AA6E0@phx.gbl>
	<BAY130-F36FC9E1C7E062A710D938AA6E0@phx.gbl>
Message-ID: <c5b438120703270657r2d5050dbta9b49d6d5913d828@mail.gmail.com>

I am not running Debian and am not on a linux box right now. I can be
in a few minutes.  I think we should move this to the matplotlib users
list.  Someone there can help you install on Debian.

Ryan

On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> this is where the folder is
> /home2/diazdeim/matplotlib-0.90.0
> i want to run the setup from the terminal but i dont know how to do it
> thaks for helping
>
>
> >From: "javi markez bigara" <nowar00 at hotmail.com>
> >Reply-To: SciPy Users List <scipy-user at scipy.org>
> >To: scipy-user at scipy.org
> >Subject: Re: [SciPy-user] plotting with scipy
> >Date: Tue, 27 Mar 2007 13:48:01 +0000
> >
> >
> >this is the python i am using, i download the mathplotlib but i think is
> >not
> >working
> >i just save it in the Desktop.. and this is the forder i get:
> >matplotlib-0.90.0
> >
> >Python 2.3.5 (#2, Sep  4 2005, 22:01:42)
> >[GCC 3.3.5 (Debian 1:3.3.5-13)] on linux2
> >
> >
> > >From: "Ryan Krauss" <ryanlists at gmail.com>
> > >Reply-To: SciPy Users List <scipy-user at scipy.org>
> > >To: "SciPy Users List" <scipy-user at scipy.org>
> > >Subject: Re: [SciPy-user] plotting with scipy
> > >Date: Tue, 27 Mar 2007 08:37:52 -0500
> > >
> > >What platform are you running?  There should be an executable for
> > >windows that puts them in the proper directory.  They can go anywhere
> > >on your PYTHONPATH, but the right answer is in the directory
> > >C:\Python24\Lib\site-packages
> > >(or I presume Python25 if you are running the current version of python).
> > >
> > >So, I have the folders
> > >C:\Python24\Lib\site-packages\matplotlib
> > >C:\Python24\Lib\site-packages\scipy
> > >etc.
> > >
> > >On linux, you usually want to unzip the tarball and there should be a
> > >setup.py file in the top level.  Then just run
> > >sudo python setup.py install
> > >(or similar if you aren't on a debian based "sudo" distribution).
> > >
> > >Let me know if you still have trouble.
> > >
> > >Ryan
> > >
> > >On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > > >
> > > > thanks a lot for ur reply i will download it but the thig is that i
> >dont
> > > > know were to put it them to import it from python.....
> > > > thanks a lot Ryan
> > > > nowar
> > > >
> > > > >From: "Ryan Krauss" <ryanlists at gmail.com>
> > > > >Reply-To: SciPy Users List <scipy-user at scipy.org>
> > > > >To: "SciPy Users List" <scipy-user at scipy.org>
> > > > >Subject: Re: [SciPy-user] plotting with scipy
> > > > >Date: Tue, 27 Mar 2007 07:30:48 -0500
> > > > >
> > > > >I think almost all users of scipy use matplotlib instead of scipy's
> > > > >plotting capabilities.  It is really a great package especially if
> >you
> > > > >only need 2D plotting.  Check out matplotlib.sourceforge.net.
> > > > >
> > > > >Ryan
> > > > >
> > > > >On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > > > > > hi,
> > > > > >
> > > > > > i am looking for the easy way to plot some coordinates with
> > > > > > scipy.gplt.pyPlot this is how i try it but is not working as i
> >would
> > > > >like
> > > > > > to....
> > > > > > the new list is a list of pair coordinates something like this:
> > > > > >
> > > > > > [-6.654644,65.55555]
> > > > > > [-6.645654,65.98798]
> > > > > > [-6.987813,65.28951]
> > > > > > [-6.645654,65.65987]
> > > > > > [-6.634895,65.36985]
> > > > > >
> > > > > > >>>import scipy.plt
> > > > > > >>>from scipy import gplt
> > > > > > >>>draw=scipy.gplt.pyPlot.Plot()
> > > > > > >>>draw.plot(newlist)
> > > > > > >>>draw.title('little map')
> > > > > > >>>draw.xtitle('longitude')
> > > > > > >>>draw.ytitle('latitude')
> > > > > >
> > > > > > the thing is that i get in the x axes the number of list and then
> > >two
> > > > >strait
> > > > > > lines
> > > > > >
> > > > > >
> > > > > > 65_________________________________
> > > > > > 50
> > > > > > 35
> > > > > > 20
> > > > > > 5
> > > > > > -10-------------------------------------------------
> > > > > > 0      1          2         3         4       5     ............
> > > > > >
> > > > > > this is wat it plots more or less
> > > > > > can anybody help me.
> > > > > > thanks in advance
> > > > > > Nowar00
> > > > > >
> > > > > > _________________________________________________________________
> > > > > > Descarga gratis la Barra de Herramientas de MSN
> > > > > >
> > > >
> > >
> > >http://www.msn.es/usuario/busqueda/barra?XAPID=2031&DI=1055&SU=http%3A//www.hotmail.com&HL=LINKTAG1OPENINGTEXT_MSNBH
> > > > > >
> > > > > > _______________________________________________
> > > > > > SciPy-user mailing list
> > > > > > SciPy-user at scipy.org
> > > > > > http://projects.scipy.org/mailman/listinfo/scipy-user
> > > > > >
> > > > >_______________________________________________
> > > > >SciPy-user mailing list
> > > > >SciPy-user at scipy.org
> > > > >http://projects.scipy.org/mailman/listinfo/scipy-user
> > > >
> > > > _________________________________________________________________
> > > > Acepta el reto MSN Premium: Correos m?s divertidos con fotos y textos
> > > > incre?bles en MSN Premium. Desc?rgalo y pru?balo 2 meses gratis.
> > > >
> > >http://join.msn.com?XAPID=1697&DI=1055&HL=Footer_mailsenviados_correosmasdivertidos
> > > >
> > > > _______________________________________________
> > > > SciPy-user mailing list
> > > > SciPy-user at scipy.org
> > > > http://projects.scipy.org/mailman/listinfo/scipy-user
> > > >
> > >_______________________________________________
> > >SciPy-user mailing list
> > >SciPy-user at scipy.org
> > >http://projects.scipy.org/mailman/listinfo/scipy-user
> >
> >_________________________________________________________________
> >Grandes ?xitos, superh?roes, imitaciones, cine y TV...
> >http://es.msn.kiwee.com/ Lo mejor para tu m?vil.
> >
> >_______________________________________________
> >SciPy-user mailing list
> >SciPy-user at scipy.org
> >http://projects.scipy.org/mailman/listinfo/scipy-user
>
> _________________________________________________________________
> Acepta el reto MSN Premium: Correos m?s divertidos con fotos y textos
> incre?bles en MSN Premium. Desc?rgalo y pru?balo 2 meses gratis.
> http://join.msn.com?XAPID=1697&DI=1055&HL=Footer_mailsenviados_correosmasdivertidos
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From v-nijs at kellogg.northwestern.edu  Tue Mar 27 09:58:15 2007
From: v-nijs at kellogg.northwestern.edu (Vincent Nijs)
Date: Tue, 27 Mar 2007 08:58:15 -0500
Subject: [SciPy-user] Scipy and Multiple Regression ??
In-Reply-To: <4608C8A8.5060703@rub.de>
Message-ID: <C22E8D27.5384%v-nijs@kellogg.northwestern.edu>

Alex,

If you just want OLS and would like a bunch of fit-statistics to come with
it you might also want to check-out

http://www.scipy.org/Cookbook/OLS

Best,

Vincent


On 3/27/07 2:32 AM, "Axel Kowald" <Axel.Kowald at rub.de> wrote:

> Hi,
> 
> thanks for the quick reply. Yes, I want to do linear least-squares and
> it seems scipy.linalg.lstsq() is what I'm looking for.
> 
> Many thanks,
> 
>        axel
> 
>> What exactly do you mean by "multiple regression"? If you mean linear
>> least-squares with models of the form
>> 
>>   y = a0 + a1*x + a2*x**2 + ...
>> 
>> or
>> 
>>   y = a0 + a1*x1 + a2*x2 + ...
>> 
>> then scipy.linalg.lstsq() should work for you.
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
> 




From fredmfp at gmail.com  Tue Mar 27 10:06:30 2007
From: fredmfp at gmail.com (fred)
Date: Tue, 27 Mar 2007 16:06:30 +0200
Subject: [SciPy-user] plotting with scipy
In-Reply-To: <c5b438120703270657r2d5050dbta9b49d6d5913d828@mail.gmail.com>
References: <BAY130-F40EF114A184B2B02F117C4AA6E0@phx.gbl>	<BAY130-F36FC9E1C7E062A710D938AA6E0@phx.gbl>
	<c5b438120703270657r2d5050dbta9b49d6d5913d828@mail.gmail.com>
Message-ID: <460924E6.2070901@gmail.com>

Ryan Krauss a ?crit :
> I am not running Debian and am not on a linux box right now. I can be
> in a few minutes.  I think we should move this to the matplotlib users
> list.  Someone there can help you install on Debian.
>   
Sure :-)

Simply type

apt-get install python-matplotlib.

You should be on your own.

-- 
http://scipy.org/FredericPetit



From pgmdevlist at gmail.com  Tue Mar 27 10:06:38 2007
From: pgmdevlist at gmail.com (Pierre GM)
Date: Tue, 27 Mar 2007 10:06:38 -0400
Subject: [SciPy-user] plotting with scipy
In-Reply-To: <c5b438120703270657r2d5050dbta9b49d6d5913d828@mail.gmail.com>
References: <BAY130-F40EF114A184B2B02F117C4AA6E0@phx.gbl>
	<BAY130-F36FC9E1C7E062A710D938AA6E0@phx.gbl>
	<c5b438120703270657r2d5050dbta9b49d6d5913d828@mail.gmail.com>
Message-ID: <200703271006.38193.pgmdevlist@gmail.com>


> On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > this is where the folder is
> > /home2/diazdeim/matplotlib-0.90.0
> > i want to run the setup from the terminal but i dont know how to do it
> > thaks for helping

Er, 
cd /home2/diazdeim/matplotlib-0.90.0
python setup.py install

doesn't work for  you ?


From nowar00 at hotmail.com  Tue Mar 27 10:18:22 2007
From: nowar00 at hotmail.com (javi markez bigara)
Date: Tue, 27 Mar 2007 14:18:22 +0000
Subject: [SciPy-user] plotting with scipy
In-Reply-To: <200703271006.38193.pgmdevlist@gmail.com>
Message-ID: <BAY130-F32F41B2830B9D9EFF24C8BAA6E0@phx.gbl>

hey PIER thanks a lot now i think i have instaled it
thanks a lot man



>From: Pierre GM <pgmdevlist at gmail.com>
>Reply-To: SciPy Users List <scipy-user at scipy.org>
>To: SciPy Users List <scipy-user at scipy.org>
>Subject: Re: [SciPy-user] plotting with scipy
>Date: Tue, 27 Mar 2007 10:06:38 -0400
>
>
> > On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > > this is where the folder is
> > > /home2/diazdeim/matplotlib-0.90.0
> > > i want to run the setup from the terminal but i dont know how to do it
> > > thaks for helping
>
>Er,
>cd /home2/diazdeim/matplotlib-0.90.0
>python setup.py install
>
>doesn't work for  you ?
>_______________________________________________
>SciPy-user mailing list
>SciPy-user at scipy.org
>http://projects.scipy.org/mailman/listinfo/scipy-user

_________________________________________________________________
Un amor, una aventura, compa??a para un viaje. Reg?strate gratis en MSN Amor 
& Amistad. http://match.msn.es/match/mt.cfm?pg=channel&tcid=162349



From zufus at zufus.org  Tue Mar 27 16:40:17 2007
From: zufus at zufus.org (Marco Presi)
Date: Tue, 27 Mar 2007 21:40:17 +0100
Subject: [SciPy-user] plotting with scipy
In-Reply-To: <c5b438120703270657r2d5050dbta9b49d6d5913d828@mail.gmail.com>
References: <BAY130-F40EF114A184B2B02F117C4AA6E0@phx.gbl>
	<BAY130-F36FC9E1C7E062A710D938AA6E0@phx.gbl>
	<c5b438120703270657r2d5050dbta9b49d6d5913d828@mail.gmail.com>
Message-ID: <1175028017.4093.0.camel@mafia.sssup.it>

Il giorno mar, 27/03/2007 alle 08.57 -0500, Ryan Krauss ha scritto:
> I am not running Debian and am not on a linux box right now. I can be
> in a few minutes.  I think we should move this to the matplotlib users
> list.  Someone there can help you install on Debian.

#aptitude install python-matplotlib

that's it.

After installing, run

$ipython -pylab

Regards

Marco

> Ryan
> 
> On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > this is where the folder is
> > /home2/diazdeim/matplotlib-0.90.0
> > i want to run the setup from the terminal but i dont know how to do it
> > thaks for helping
> >
> >
> > >From: "javi markez bigara" <nowar00 at hotmail.com>
> > >Reply-To: SciPy Users List <scipy-user at scipy.org>
> > >To: scipy-user at scipy.org
> > >Subject: Re: [SciPy-user] plotting with scipy
> > >Date: Tue, 27 Mar 2007 13:48:01 +0000
> > >
> > >
> > >this is the python i am using, i download the mathplotlib but i think is
> > >not
> > >working
> > >i just save it in the Desktop.. and this is the forder i get:
> > >matplotlib-0.90.0
> > >
> > >Python 2.3.5 (#2, Sep  4 2005, 22:01:42)
> > >[GCC 3.3.5 (Debian 1:3.3.5-13)] on linux2
> > >
> > >
> > > >From: "Ryan Krauss" <ryanlists at gmail.com>
> > > >Reply-To: SciPy Users List <scipy-user at scipy.org>
> > > >To: "SciPy Users List" <scipy-user at scipy.org>
> > > >Subject: Re: [SciPy-user] plotting with scipy
> > > >Date: Tue, 27 Mar 2007 08:37:52 -0500
> > > >
> > > >What platform are you running?  There should be an executable for
> > > >windows that puts them in the proper directory.  They can go anywhere
> > > >on your PYTHONPATH, but the right answer is in the directory
> > > >C:\Python24\Lib\site-packages
> > > >(or I presume Python25 if you are running the current version of python).
> > > >
> > > >So, I have the folders
> > > >C:\Python24\Lib\site-packages\matplotlib
> > > >C:\Python24\Lib\site-packages\scipy
> > > >etc.
> > > >
> > > >On linux, you usually want to unzip the tarball and there should be a
> > > >setup.py file in the top level.  Then just run
> > > >sudo python setup.py install
> > > >(or similar if you aren't on a debian based "sudo" distribution).
> > > >
> > > >Let me know if you still have trouble.
> > > >
> > > >Ryan
> > > >
> > > >On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > > > >
> > > > > thanks a lot for ur reply i will download it but the thig is that i
> > >dont
> > > > > know were to put it them to import it from python.....
> > > > > thanks a lot Ryan
> > > > > nowar
> > > > >
> > > > > >From: "Ryan Krauss" <ryanlists at gmail.com>
> > > > > >Reply-To: SciPy Users List <scipy-user at scipy.org>
> > > > > >To: "SciPy Users List" <scipy-user at scipy.org>
> > > > > >Subject: Re: [SciPy-user] plotting with scipy
> > > > > >Date: Tue, 27 Mar 2007 07:30:48 -0500
> > > > > >
> > > > > >I think almost all users of scipy use matplotlib instead of scipy's
> > > > > >plotting capabilities.  It is really a great package especially if
> > >you
> > > > > >only need 2D plotting.  Check out matplotlib.sourceforge.net.
> > > > > >
> > > > > >Ryan
> > > > > >
> > > > > >On 3/27/07, javi markez bigara <nowar00 at hotmail.com> wrote:
> > > > > > > hi,
> > > > > > >
> > > > > > > i am looking for the easy way to plot some coordinates with
> > > > > > > scipy.gplt.pyPlot this is how i try it but is not working as i
> > >would
> > > > > >like
> > > > > > > to....
> > > > > > > the new list is a list of pair coordinates something like this:
> > > > > > >
> > > > > > > [-6.654644,65.55555]
> > > > > > > [-6.645654,65.98798]
> > > > > > > [-6.987813,65.28951]
> > > > > > > [-6.645654,65.65987]
> > > > > > > [-6.634895,65.36985]
> > > > > > >
> > > > > > > >>>import scipy.plt
> > > > > > > >>>from scipy import gplt
> > > > > > > >>>draw=scipy.gplt.pyPlot.Plot()
> > > > > > > >>>draw.plot(newlist)
> > > > > > > >>>draw.title('little map')
> > > > > > > >>>draw.xtitle('longitude')
> > > > > > > >>>draw.ytitle('latitude')
> > > > > > >
> > > > > > > the thing is that i get in the x axes the number of list and then
> > > >two
> > > > > >strait
> > > > > > > lines
> > > > > > >
> > > > > > >
> > > > > > > 65_________________________________
> > > > > > > 50
> > > > > > > 35
> > > > > > > 20
> > > > > > > 5
> > > > > > > -10-------------------------------------------------
> > > > > > > 0      1          2         3         4       5     ............
> > > > > > >
> > > > > > > this is wat it plots more or less
> > > > > > > can anybody help me.
> > > > > > > thanks in advance
> > > > > > > Nowar00
> > > > > > >
> > > > > > > _________________________________________________________________
> > > > > > > Descarga gratis la Barra de Herramientas de MSN
> > > > > > >
> > > > >
> > > >
> > > >http://www.msn.es/usuario/busqueda/barra?XAPID=2031&DI=1055&SU=http%3A//www.hotmail.com&HL=LINKTAG1OPENINGTEXT_MSNBH
> > > > > > >
> > > > > > > _______________________________________________
> > > > > > > SciPy-user mailing list
> > > > > > > SciPy-user at scipy.org
> > > > > > > http://projects.scipy.org/mailman/listinfo/scipy-user
> > > > > > >
> > > > > >_______________________________________________
> > > > > >SciPy-user mailing list
> > > > > >SciPy-user at scipy.org
> > > > > >http://projects.scipy.org/mailman/listinfo/scipy-user
> > > > >
> > > > > _________________________________________________________________
> > > > > Acepta el reto MSN Premium: Correos m?s divertidos con fotos y textos
> > > > > incre?bles en MSN Premium. Desc?rgalo y pru?balo 2 meses gratis.
> > > > >
> > > >http://join.msn.com?XAPID=1697&DI=1055&HL=Footer_mailsenviados_correosmasdivertidos
> > > > >
> > > > > _______________________________________________
> > > > > SciPy-user mailing list
> > > > > SciPy-user at scipy.org
> > > > > http://projects.scipy.org/mailman/listinfo/scipy-user
> > > > >
> > > >_______________________________________________
> > > >SciPy-user mailing list
> > > >SciPy-user at scipy.org
> > > >http://projects.scipy.org/mailman/listinfo/scipy-user
> > >
> > >_________________________________________________________________
> > >Grandes ?xitos, superh?roes, imitaciones, cine y TV...
> > >http://es.msn.kiwee.com/ Lo mejor para tu m?vil.
> > >
> > >_______________________________________________
> > >SciPy-user mailing list
> > >SciPy-user at scipy.org
> > >http://projects.scipy.org/mailman/listinfo/scipy-user
> >
> > _________________________________________________________________
> > Acepta el reto MSN Premium: Correos m?s divertidos con fotos y textos
> > incre?bles en MSN Premium. Desc?rgalo y pru?balo 2 meses gratis.
> > http://join.msn.com?XAPID=1697&DI=1055&HL=Footer_mailsenviados_correosmasdivertidos
> >
> > _______________________________________________
> > SciPy-user mailing list
> > SciPy-user at scipy.org
> > http://projects.scipy.org/mailman/listinfo/scipy-user
> >
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user



From emilia12 at mail.bg  Tue Mar 27 17:28:08 2007
From: emilia12 at mail.bg (emilia12 at mail.bg)
Date: Wed, 28 Mar 2007 00:28:08 +0300
Subject: [SciPy-user] [Cookbook] FiltFilt does not work
In-Reply-To: <20070326204851.GB15905@localhost.localdomain>
References: <1174932755.baa896d35359e@mail.bg>
	<20070326204851.GB15905@localhost.localdomain>
Message-ID: <1175030888.7f1e1d87f62d8@mail.bg>

hello Antoine,

you are right, the last 20-30 points are too far. I agree
that the picture do not match the result from the code.
What a shame ... IMO this is not the sci way to speak in
filter praise ;-)

best regards,
e.


????? ?? ????? ?? Antoine Sirinelli
<scipy-user at monte-stello.com>:

> On Mon, Mar 26, 2007 at 09:12:35PM +0300,
> emilia12 at mail.bg wrote:
> > and my results are very close to curves on the picture
> from
> > the example
> > the difference is only in the noise (the signal have a
> small
> > random part, which is unique for each calculation)
>
> I agree that the random part have to be different but as
> you can see on
> the figure attached, the last points show a sligtly big
> difference. The
> filtered signal goes away from the noisy signal.
>
> Antoine
>
>




-----------------------------

SCENA - ???????????? ????????? ???????? ?? ??????? ??????????? ? ??????????.
http://www.bgscena.com/



From emin.shopper at gmail.com  Tue Mar 27 17:30:08 2007
From: emin.shopper at gmail.com (Emin.shopper Martinian.shopper)
Date: Tue, 27 Mar 2007 17:30:08 -0400
Subject: [SciPy-user] fmin stopping on something not a local optimum
Message-ID: <32e43bb70703271430u5d98d411y86eda0776395607d@mail.gmail.com>

Dear Experts,

I am getting some strange behavior with scipy.optimize.fmin. It seems to
"converge" to what it thinks is an optimal solution that is not even a local
optimum. Consequently, I end up having to do things like

answers = [initialGuess]
for i in range(5):
   answers.append( scipy.optimize.fmin(func,x0=answers[-1]) )

In each iteration of the loop, fmin prints information saying it terminated
successfully (i.e., it is not hitting the maxiter or maxfun constraints
since I have set these very high) yet the "Current function value" keeps
improving.

Should I be setting the parameters for fmin in a special way to tell it not
to stop too early? I've tried fiddling with xtol and ftol without much
success.

Is one supposed to call fmin repeatedly like this?

Thanks,
-Emin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070327/70fb031c/attachment.html>

From pgmdevlist at gmail.com  Tue Mar 27 18:27:39 2007
From: pgmdevlist at gmail.com (Pierre GM)
Date: Tue, 27 Mar 2007 18:27:39 -0400
Subject: [SciPy-user] Scipy examples
Message-ID: <200703271827.39154.pgmdevlist@gmail.com>

All,
Is there an equivalent of the Numpy_Example wiki page for scipy, listing 
functions per package ? If not, how difficult is it to get one ?


From robert.kern at gmail.com  Tue Mar 27 18:30:10 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 27 Mar 2007 17:30:10 -0500
Subject: [SciPy-user] Scipy examples
In-Reply-To: <200703271827.39154.pgmdevlist@gmail.com>
References: <200703271827.39154.pgmdevlist@gmail.com>
Message-ID: <46099AF2.5060605@gmail.com>

Pierre GM wrote:
> All,
> Is there an equivalent of the Numpy_Example wiki page for scipy, listing 
> functions per package ? If not, how difficult is it to get one ?

Here is the start of such a page:

  http://www.scipy.org/scipy_Example_List

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From pgmdevlist at gmail.com  Tue Mar 27 18:35:00 2007
From: pgmdevlist at gmail.com (Pierre GM)
Date: Tue, 27 Mar 2007 18:35:00 -0400
Subject: [SciPy-user] Scipy examples
In-Reply-To: <46099AF2.5060605@gmail.com>
References: <200703271827.39154.pgmdevlist@gmail.com>
	<46099AF2.5060605@gmail.com>
Message-ID: <200703271835.00984.pgmdevlist@gmail.com>

On Tuesday 27 March 2007 18:30:10 Robert Kern wrote:
> Pierre GM wrote:
> > All,
> > Is there an equivalent of the Numpy_Example wiki page for scipy, listing
> > functions per package ? If not, how difficult is it to get one ?
>
> Here is the start of such a page:
>
>   http://www.scipy.org/scipy_Example_List

OK, that's a start. Thanks a lot!


From akumar at iitm.ac.in  Tue Mar 27 21:30:36 2007
From: akumar at iitm.ac.in (Kumar Appaiah)
Date: Wed, 28 Mar 2007 07:00:36 +0530
Subject: [SciPy-user] A communications module
Message-ID: <20070328013036.GA7095@localhost>

Dear SciPy users,

I have been using SciPy for quite some time now, although at a basic
level. It seems to be very nice for doing the stuff, and it has been a
long time since I have done anything much in GNU Octave, my other
favourite.

There is one thing that I have noticed in SciPy which makes it
attractive for everyone, that is it's compatibility with
Matlab. However, in this respect, the communications module seems to
be missing. I was wondering whether we could put some snippets to do
the basic stuff just the way Matlab does (PSK, QAM, AWGN channel etc.)
and get started with a ``scipy.comm'' module.

I wanted to ping the list because I wanted to know of anyone else
shares this view, and I would love to contribute code to such a
module, but I would like to wait for comments, to get to know the
"right way" to start this off. I mean, identifying the first things to
write, how to modularize the stuff etc.

Thanks.

Kumar
-- 
Kumar Appaiah,
462, Jamuna Hostel,
Indian Institute of Technology Madras,
Chennai - 600 036



From david.huard at gmail.com  Wed Mar 28 09:26:50 2007
From: david.huard at gmail.com (David Huard)
Date: Wed, 28 Mar 2007 09:26:50 -0400
Subject: [SciPy-user] fmin stopping on something not a local optimum
In-Reply-To: <32e43bb70703271430u5d98d411y86eda0776395607d@mail.gmail.com>
References: <32e43bb70703271430u5d98d411y86eda0776395607d@mail.gmail.com>
Message-ID: <91cf711d0703280626n11519ce5re26fa83d783ba928@mail.gmail.com>

Hi Emin,
just a thought: fmin searches for a minimum, not an optimum in the general
sense. Could that be the problem ?

If not, the problem could indeed be with ftol being set too high. What kind
of function are you optimizing ? Maybe try using brute force optimization
just to get an idea of what's going on. I'm using fmin extensively and the
only times it gives me trouble, it ends up the source of the problem is
sitting in front of the screen.

David

2007/3/27, Emin.shopper Martinian.shopper <emin.shopper at gmail.com>:
>
> Dear Experts,
>
> I am getting some strange behavior with scipy.optimize.fmin. It seems to
> "converge" to what it thinks is an optimal solution that is not even a local
> optimum. Consequently, I end up having to do things like
>
> answers = [initialGuess]
> for i in range(5):
>    answers.append( scipy.optimize.fmin(func,x0=answers[-1]) )
>
> In each iteration of the loop, fmin prints information saying it
> terminated successfully (i.e., it is not hitting the maxiter or maxfun
> constraints since I have set these very high) yet the "Current function
> value" keeps improving.
>
> Should I be setting the parameters for fmin in a special way to tell it
> not to stop too early? I've tried fiddling with xtol and ftol without much
> success.
>
> Is one supposed to call fmin repeatedly like this?
>
> Thanks,
> -Emin
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070328/9ff02a3d/attachment.html>

From rfl26 at cam.ac.uk  Wed Mar 28 10:00:48 2007
From: rfl26 at cam.ac.uk (Fred Ludlow)
Date: Wed, 28 Mar 2007 15:00:48 +0100
Subject: [SciPy-user] fmin stopping on something not a local optimum
In-Reply-To: <32e43bb70703271430u5d98d411y86eda0776395607d@mail.gmail.com>
References: <32e43bb70703271430u5d98d411y86eda0776395607d@mail.gmail.com>
Message-ID: <460A7510.4010601@cam.ac.uk>

Emin.shopper Martinian.shopper wrote:
> Dear Experts,
> 
> I am getting some strange behavior with scipy.optimize.fmin. It seems to 
> "converge" to what it thinks is an optimal solution that is not even a 
> local optimum. Consequently, I end up having to do things like
> 
> answers = [initialGuess]
> for i in range(5):
>    answers.append( scipy.optimize.fmin(func,x0=answers[-1]) )
> 
> In each iteration of the loop, fmin prints information saying it 
> terminated successfully (i.e., it is not hitting the maxiter or maxfun 
> constraints since I have set these very high) yet the "Current function 
> value" keeps improving.
> 
> Should I be setting the parameters for fmin in a special way to tell it 
> not to stop too early? I've tried fiddling with xtol and ftol without 
> much success.
> 
> Is one supposed to call fmin repeatedly like this?

Hi Emin,

I don't think that's how it's supposed to work. This is a bit of a 
guess, and I'm not an expert, but I'm sure there'll be someone helpful 
reading this to correct my mistakes... :)

What kind of function are you using as "func"? Looking at the fmin 
source, the first points it evaluates are the x0 you supply, then it 
varies each parameter in turn to +5% of it's current value (or just 0.05 
if it starts at zero), to generate the intial simplex. So, if you're 
objective function is not very smooth (for example if it's generated by 
another iterative procedure with fairly coarse convergence requirements) 
it might immediately find itself in a very local minimum that's an 
artifact of an imprecise objective function.

On the other hand, if your function is arithmetic and you don't lose 
much precision in it I don't suppose this is relevant.

Fred


From giorgio.luciano at chimica.unige.it  Wed Mar 28 10:58:42 2007
From: giorgio.luciano at chimica.unige.it (Giorgio Luciano)
Date: Wed, 28 Mar 2007 16:58:42 +0200
Subject: [SciPy-user] matlab, idle, interactivity and teaching
Message-ID: <460A82A2.90602@chimica.unige.it>

Hello to all,
I've thread that apperead some time ago on this list about matlab and 
teaching.
I've discovered python recently  and translated part of the routine I 
use in python (www.chemometrics.it).
Some of my collegue asked me if I could show them how to use python. For 
matlab user I guess the first problem is to setup everything, but I just 
fixed it preparing a directory with all the package I need and a 
matplotlibrc file for interactive mode + a shortcut for idle -n use.
The second problem is that people now  wants some bells and whistles of 
matlab that I have to admit sometime can be very helpful for saving 
time. The bells and whistles are about the workspace.
It's difficult to cut and paste from gnumeric/excel (I generally use txt 
file but it's no so immediate) and also there is no "visual" workspace. 
I cannot succeed also in saving workspace (I know there is a function so 
iosave.mat but I didn't manage easily hot to use it)
For overpass this problems I've tried to use QME-DEV which is in early 
stage of development (alpha) but promise well.
What people like of python/matplot/scipy
-its free ;)
-they like a lot the plotting style and capabilities (they find the png 
and svg file very clear and accurate)
-they like IDLE as editor (ehy it's has the same color of matlab ;) ! )

So my question is . Do you have a similar experience ?
How do you help people in moving the first step ?
do you use (and also does it exist) a more friendly environment than 
IDLE except from QME-DEV.

I know that this question may look silly, but in my opinion also how 
much is user friendly a software is very important for getting new users.
Cheers to all
Giorgio




From peridot.faceted at gmail.com  Wed Mar 28 12:00:41 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Wed, 28 Mar 2007 12:00:41 -0400
Subject: [SciPy-user] fmin stopping on something not a local optimum
In-Reply-To: <32e43bb70703271430u5d98d411y86eda0776395607d@mail.gmail.com>
References: <32e43bb70703271430u5d98d411y86eda0776395607d@mail.gmail.com>
Message-ID: <ce557a360703280900q4d61450an19e745e5c5acb29e@mail.gmail.com>

On 27/03/07, Emin.shopper Martinian.shopper <emin.shopper at gmail.com> wrote:
> Dear Experts,

Well, this isn't addressed to me, but I do use fmin; my objective
function is obtained by numerical integration, but the only real
problem I've had was when it started returning NaNs (numerical
overflows in scipy.stats.chi2.pdf; is there no log-pdf function?).

> I am getting some strange behavior with scipy.optimize.fmin. It seems to
> "converge" to what it thinks is an optimal solution that is not even a local
> optimum. Consequently, I end up having to do things like
>
> answers = [initialGuess]
> for i in range(5):
>    answers.append( scipy.optimize.fmin(func,x0=answers[-1]) )
>
> In each iteration of the loop, fmin prints information saying it terminated
> successfully (i.e., it is not hitting the maxiter or maxfun constraints
> since I have set these very high) yet the "Current function value" keeps
> improving.
>
> Should I be setting the parameters for fmin in a special way to tell it not
> to stop too early? I've tried fiddling with xtol and ftol without much
> success.
>
> Is one supposed to call fmin repeatedly like this?

No. It does have a procedure for deciding when to stop, which can be
fooled. When I was getting NaNs, it signalled error by running to
completion and returning my input guesses, unimproved; the error
handling could be better.

ftol is an absolute tolerance - is your function actually changing on
the scale of ftol? are your x values varying on the scale of xtol?
does your goal function produce values more accurate than ftol? (I
know you said you'd fiddled with them.)

You might look at the code - IIRC it's not some unintelligible FORTRAN
code, it's just python.

Anne


From fccoelho at gmail.com  Wed Mar 28 13:57:06 2007
From: fccoelho at gmail.com (Flavio Coelho)
Date: Wed, 28 Mar 2007 14:57:06 -0300
Subject: [SciPy-user] returning an array from weave inline
Message-ID: <d9af7a880703281057y55f8e063u239753fa8a774346@mail.gmail.com>

I get a compilation error when  I try to return an array from weave.inline

here is my test code ( a simple matrix multplication)

from scipy import weave
from scipy.weave import converters

def Dot(a1,a2):
    """
    multiplica??o de matrizes em C
    """
    #print a1.shape,a2.shape
    d1,d2,d3 = a1.shape[0],a2.shape[1], a1.shape[1]
    a3 = zeros((d1,d2))
    code = """
    {
    int i, j, k;
    for( i = 0; i < d1; i++)
           for( j = 0; j < d2; j++)
               for( k = 0; k < d3; k++)
                 a3(i,j) +=  a1(i,k)*a2(k,j);
    }
    return_val = a3;
    """
    return weave.inline
(code,['a1','a2','a3','d1','d2','d3'],type_converters=converters.blitz
,compiler='gcc')

this is the compilation error I get::

<weave: compiling>
/home/flavio/.python24_compiled/sc_80afed445639e2098f517fccd0b436f93.cpp: In
function 'PyObject* compiled_func(PyObject*, PyObject*)':
/home/flavio/.python24_compiled/sc_80afed445639e2098f517fccd0b436f93.cpp:738:
error: no match for 'operator=' in 'return_val = a3'
/usr/lib/python2.4/site-packages/scipy/weave/scxx/object.h:179: note:
candidates are: py::object& py::object::operator=(const py::object&)
/home/flavio/.python24_compiled/sc_80afed445639e2098f517fccd0b436f93.cpp: In
function 'PyObject* compiled_func(PyObject*, PyObject*)':
/home/flavio/.python24_compiled/sc_80afed445639e2098f517fccd0b436f93.cpp:738:
error: no match for 'operator=' in 'return_val = a3'
/usr/lib/python2.4/site-packages/scipy/weave/scxx/object.h:179: note:
candidates are: py::object& py::object::operator=(const py::object&)
Traceback (most recent call last):
  File "DengueRJdin_01.py", line 314, in ?
    mainMeld(K,L)
  File "DengueRJdin_01.py", line 221, in mainMeld
    phi,q1theta = Rep(K, tmax,**para)
  File "DengueRJdin_01.py", line 22, in Rep
    res,inc = simEpid(tmax,q1theta[0][i],q1theta[1][i],q1theta[2][i],**args)
  File "DengueRJdin_01.py", line 143, in simEpid
    yant[:,j] = Dot(D,l)
  File "DengueRJdin_01.py", line 120, in Dot
    return weave.inline
(code,['a1','a2','a3','d1','d2','d3'],type_converters=converters.blitz
,compiler='gcc')
  File "/usr/lib/python2.4/site-packages/scipy/weave/inline_tools.py", line
338, in inline
    auto_downcast = auto_downcast,
  File "/usr/lib/python2.4/site-packages/scipy/weave/inline_tools.py", line
447, in compile_function
    verbose=verbose, **kw)
  File "/usr/lib/python2.4/site-packages/scipy/weave/ext_tools.py", line
365, in compile
    verbose = verbose, **kw)
  File "/usr/lib/python2.4/site-packages/scipy/weave/build_tools.py", line
269, in build_extension
    setup(name = module_name, ext_modules = [ext],verbose=verb)
  File "/usr/lib/python2.4/site-packages/numpy-1.0.1-py2.4-linux-i686.egg/numpy/distutils/core.py",
line 174, in setup
    return old_setup(**new_attr)
  File "/usr/lib/python2.4/distutils/core.py", line 166, in setup
    raise SystemExit, "error: " + str(msg)
scipy.weave.build_tools.CompileError: error: Command "i686-pc-linux-gnu-g++
-pthread -fno-strict-aliasing -DNDEBUG -fPIC
-I/usr/lib/python2.4/site-packages/scipy/weave
-I/usr/lib/python2.4/site-packages/scipy/weave/scxx
-I/usr/lib/python2.4/site-packages/scipy/weave/blitz
-I/usr/lib/python2.4/site-packages/numpy-1.0.1-py2.4-linux-i686.egg/numpy/core/include
-I/usr/include/python2.4 -c
/home/flavio/.python24_compiled/sc_80afed445639e2098f517fccd0b436f93.cpp -o
/tmp/flavio/python24_intermediate/compiler_060e323e3e036dd7c46bf6c968ae89ac/home/flavio/.python24_compiled/sc_80afed445639e2098f517fccd0b436f93.o"
failed with exit status 1

any help will be greatly appreciated.


-- 
Fl?vio Code?o Coelho
registered Linux user # 386432
get counted at http://counter.li.org
---------------------------
"software gets slower faster than hardware gets faster"
Niklaus Wirth's law
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070328/e6a9cbe3/attachment.html>

From fperez.net at gmail.com  Wed Mar 28 14:05:45 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Wed, 28 Mar 2007 12:05:45 -0600
Subject: [SciPy-user] returning an array from weave inline
In-Reply-To: <d9af7a880703281057y55f8e063u239753fa8a774346@mail.gmail.com>
References: <d9af7a880703281057y55f8e063u239753fa8a774346@mail.gmail.com>
Message-ID: <db6b5ecc0703281105p6229955dxc97eae2c999ca884@mail.gmail.com>

On 3/28/07, Flavio Coelho <fccoelho at gmail.com> wrote:
> I get a compilation error when  I try to return an array from weave.inline
>
> here is my test code ( a simple matrix multplication)
>
> from scipy import weave
> from scipy.weave import converters
>
> def Dot(a1,a2):
>     """
>     multiplica??o de matrizes em C
>     """
>     #print a1.shape,a2.shape
>     d1,d2,d3 = a1.shape[0],a2.shape[1], a1.shape[1]
>     a3 = zeros((d1,d2))
>     code = """
>     {
>     int i, j, k;
>     for( i = 0; i < d1; i++)
>            for( j = 0; j < d2; j++)
>                for( k = 0; k < d3; k++)
>                  a3(i,j) +=  a1(i,k)*a2(k,j);
>     }
>     return_val = a3;
>     """
>     return
> weave.inline(code,['a1','a2','a3','d1','d2','d3'],type_converters=converters.blitz,compiler='gcc')

Don't return anything.  Your code has already filled in a3, so there's
no need for you to return.

Cheers,

f


From s.mientki at ru.nl  Wed Mar 28 15:17:18 2007
From: s.mientki at ru.nl (Stef Mientki)
Date: Wed, 28 Mar 2007 21:17:18 +0200
Subject: [SciPy-user] matlab, idle, interactivity and teaching
In-Reply-To: <460A82A2.90602@chimica.unige.it>
References: <460A82A2.90602@chimica.unige.it>
Message-ID: <460ABF3E.7050809@ru.nl>

hello Giorgio,

Giorgio Luciano wrote:
> Hello to all,
> I've thread that apperead some time ago on this list about matlab and 
> teaching.
>   
I'm not directly involved in teaching,
but my advise was asked if Labview would be a good replacement better
alternative for MatLab,
in education of medical students and for medical research.
When I discovered that Labview was certainly not a good replacement,
I first tried SciLab, but soon discovered Python, which with some tools,
looks very promising.
> I've discovered python recently  and translated part of the routine I 
> use in python (www.chemometrics.it).
> Some of my collegue asked me if I could show them how to use python. For 
> matlab user I guess the first problem is to setup everything, but I just 
> fixed it preparing a directory with all the package I need and a 
> matplotlibrc file for interactive mode + a shortcut for idle -n use.
>   
Which operating system are you using ?
Did you use Enthought edition ?
On windows, with just a few clicks, I had installed Enthought edition
and PyScripter,
and could start working, almost exactly as I was used in MatLab.
> The second problem is that people now  wants some bells and whistles of 
> matlab that I have to admit sometime can be very helpful for saving 
> time. The bells and whistles are about the workspace.
> It's difficult to cut and paste from gnumeric/excel (I generally use txt 
> file but it's no so immediate) and also there is no "visual" workspace. 
> I cannot succeed also in saving workspace (I know there is a function so 
> iosave.mat but I didn't manage easily hot to use it)
>   
>From my experience, MatLab workspace is somewhat better, but still 
very limited, if you've large data sets.
In the past I've several dedicated workspaces for MatLab to overcome
these limitations.
Pyscripter gives you almost the same workspace as MatLab,
and I wouldn't be surprised if you ask the creater of PyScripter,
to make something like the MatLab workspace, he'll do it.
> For overpass this problems I've tried to use QME-DEV which is in early 
> stage of development (alpha) but promise well.
> What people like of python/matplot/scipy
> -its free ;)
> -they like a lot the plotting style and capabilities (they find the png 
> and svg file very clear and accurate)
> -they like IDLE as editor (ehy it's has the same color of matlab ;) ! )
>   
and probably can print in color, which Matlab can't ;-)
> So my question is . Do you have a similar experience ?
> How do you help people in moving the first step ?
> do you use (and also does it exist) a more friendly environment than 
> IDLE except from QME-DEV.
>   
Yes, I think we've similar experiences here, and I even go one step
further, the MatLab workspace is also too limited.
I've looked at QME-DEV, and although it looks very promising,
> I know that this question may look silly, but in my opinion also how 
> much is user friendly a software is very important for getting new users.
> Cheers to all
> Giorgio
>
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>   

-- 
cheers,
Stef Mientki
http://pic.flappie.nl




From fccoelho at gmail.com  Wed Mar 28 18:18:49 2007
From: fccoelho at gmail.com (Flavio Coelho)
Date: Wed, 28 Mar 2007 19:18:49 -0300
Subject: [SciPy-user] returning an array from weave inline
In-Reply-To: <db6b5ecc0703281105p6229955dxc97eae2c999ca884@mail.gmail.com>
References: <d9af7a880703281057y55f8e063u239753fa8a774346@mail.gmail.com>
	<db6b5ecc0703281105p6229955dxc97eae2c999ca884@mail.gmail.com>
Message-ID: <d9af7a880703281518o338d5165je0ae8b5e1580899@mail.gmail.com>

Thanks Fernando,

I thought variables converted to C were copies, so that there could not be
an in place operation on a Python variable.

Fl?vio



On 3/28/07, Fernando Perez <fperez.net at gmail.com> wrote:
>
> On 3/28/07, Flavio Coelho <fccoelho at gmail.com> wrote:
> > I get a compilation error when  I try to return an array from
> weave.inline
> >
> > here is my test code ( a simple matrix multplication)
> >
> > from scipy import weave
> > from scipy.weave import converters
> >
> > def Dot(a1,a2):
> >     """
> >     multiplica??o de matrizes em C
> >     """
> >     #print a1.shape,a2.shape
> >     d1,d2,d3 = a1.shape[0],a2.shape[1], a1.shape[1]
> >     a3 = zeros((d1,d2))
> >     code = """
> >     {
> >     int i, j, k;
> >     for( i = 0; i < d1; i++)
> >            for( j = 0; j < d2; j++)
> >                for( k = 0; k < d3; k++)
> >                  a3(i,j) +=  a1(i,k)*a2(k,j);
> >     }
> >     return_val = a3;
> >     """
> >     return
> > weave.inline(code,['a1','a2','a3','d1','d2','d3'],type_converters=
> converters.blitz,compiler='gcc')
>
> Don't return anything.  Your code has already filled in a3, so there's
> no need for you to return.
>
> Cheers,
>
> f
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>



-- 
Fl?vio Code?o Coelho
registered Linux user # 386432
get counted at http://counter.li.org
---------------------------
"software gets slower faster than hardware gets faster"
Niklaus Wirth's law
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070328/e0cf02b5/attachment.html>

From giorgio.luciano at chimica.unige.it  Thu Mar 29 03:22:25 2007
From: giorgio.luciano at chimica.unige.it (Giorgio Luciano)
Date: Thu, 29 Mar 2007 09:22:25 +0200
Subject: [SciPy-user] matlab, idle, interactivity and teaching
In-Reply-To: <460ABF3E.7050809@ru.nl>
References: <460A82A2.90602@chimica.unige.it> <460ABF3E.7050809@ru.nl>
Message-ID: <460B6931.8030109@chimica.unige.it>

Thanks for your reply,
I currently does not use Enthought (I'm waiting for a more uptodate 
package) but I surely give a try to Pyscripter and eventually conctact 
the author.

Thanks for sharing
Giorgio



From olivetti at itc.it  Thu Mar 29 03:45:24 2007
From: olivetti at itc.it (Emanuele Olivetti)
Date: Thu, 29 Mar 2007 09:45:24 +0200
Subject: [SciPy-user] returning an array from weave inline
In-Reply-To: <d9af7a880703281518o338d5165je0ae8b5e1580899@mail.gmail.com>
References: <d9af7a880703281057y55f8e063u239753fa8a774346@mail.gmail.com>	<db6b5ecc0703281105p6229955dxc97eae2c999ca884@mail.gmail.com>
	<d9af7a880703281518o338d5165je0ae8b5e1580899@mail.gmail.com>
Message-ID: <460B6E94.8010708@itc.it>

Flavio Coelho wrote:
> Thanks Fernando,
> 
> I thought variables converted to C were copies, so that there could not be
> an in place operation on a Python variable.

They are copies like it is in the C way of thinking: if you pass a
basic type (e.g. an int) to the inlined C code and modify it then
when you are back to python there is no change in the initial variable
you passed since just the copy was modified. If you pass something more complex
(e.g. a numpy array) then you are just giving a copy of the it's reference
(or pointer) to that function. With that reference (or copy of) you
can access the the _original_ array and modify it. So when you are back to
python your array _is_ changed. This mechanism allows better efficiency:
what happens if you have a huge 2Gb array and pass it to the inline code?
It's not desirable to make a whole copy and allocate other 2Gb or RAM...

Anyway python itself works like that. E.g.:
----
def f(a,b):
     a=a+1
     b.append(1)
     return

a=1
b=[3,2]
print a,b
f(a,b)
print a,b
----

Hope this helps,

Emanuele

------------------
ITC -> dall'1 marzo 2007 Fondazione Bruno Kessler
ITC -> since 1 March 2007 Fondazione Bruno Kessler
------------------


From raphael.langella at steria.com  Thu Mar 29 08:03:45 2007
From: raphael.langella at steria.com (raphael langella)
Date: Thu, 29 Mar 2007 14:03:45 +0200
Subject: [SciPy-user] building numpy/scipy on Solaris
Message-ID: <13a1ee25e.e25e13a1e@steria.com>

I just realized I've got much more recent versions of libsunperf and C
and fortran compilers. I've got Sun Studio 10 & 11 installed. So I tried
building numpy with both of them and each time I get this error when
importing numpy :

>>> import numpy
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/tmp/lib/python2.5/site-packages/numpy/__init__.py", line 40, in
<module>
    import linalg
  File "/tmp/lib/python2.5/site-packages/numpy/linalg/__init__.py", line
4, in <module>
    from linalg import *
  File "/tmp/lib/python2.5/site-packages/numpy/linalg/linalg.py", line
25, in <module>
    from numpy.linalg import lapack_lite
ImportError: ld.so.1: python: fatal: relocation error: file
/tmp/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so: symbol
__mt_MasterFunction_rtc_: referenced symbol not found

I tried the CPPFLAGS='-DNO_APPEND_FORTRAN' option, but it didn't change
anything. Any idea?

-------------- next part --------------
_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user

From raphael.langella at steria.com  Thu Mar 29 08:09:20 2007
From: raphael.langella at steria.com (raphael langella)
Date: Thu, 29 Mar 2007 14:09:20 +0200
Subject: [SciPy-user] building numpy/scipy on Solaris
Message-ID: <e8d5fe4a.fe4ae8d5@steria.com>

I just realized I've got much more recent versions of libsunperf and C
and fortran compilers. I've got Sun Studio 10 & 11 installed. So I tried
building numpy with both of them and each time I get this error when
importing numpy :

>>> import numpy
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/tmp/lib/python2.5/site-packages/numpy/__init__.py", line 40, in
<module>
    import linalg
  File "/tmp/lib/python2.5/site-packages/numpy/linalg/__init__.py", line
4, in <module>
    from linalg import *
  File "/tmp/lib/python2.5/site-packages/numpy/linalg/linalg.py", line
25, in <module>
    from numpy.linalg import lapack_lite
ImportError: ld.so.1: python: fatal: relocation error: file
/tmp/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so: symbol
__mt_MasterFunction_rtc_: referenced symbol not found

I tried the CPPFLAGS='-DNO_APPEND_FORTRAN' option, but it didn't change
anything. Any idea?


---- Messages d?origine ----
De: raphael langella <raphael.langella at steria.com>
Date: mardi, mars 20, 2007 12:08 pm
Objet: [SciPy-user] building numpy/scipy on Solaris

> I'm trying to build numpy and scipy on Solaris 8.
> The BLAS FAQ on netlib.org suggests using optimized BLAS librairies
> provided by computer vendor, like the SUN Performance Library. This
> library is supposed to provide enhanced and optimized version of BLAS
> and LAPACK. I happen to have Forte 7 installed, so I first tried to
> build against this library (libsunperf.a).
> I tried several versions and different compilation options, but I 
> alwaysget undefined symbols. Is this supported? As anyone ever 
> succeeded in
> using this library to compile scipy and numpy?
> I also did some unsuccessful tests with Atlas. But, I think I'm now 
> veryclose to have it working with the standard BLAS/LAPACK (from
> netlib.org). I have to use GNU ld (unresolved symbols when using Sun's
> ld). Numpy builds and tests without any error. I've got a few 
> errors and
> failures with scipy (see attachment for the complete test report). Are
> these errors critical? I don't understand where they come from and how
> to correct them.
> Thanks for your help.
> 
> Rapha?l Langella
> 
-------------- next part --------------
_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user

From cookedm at physics.mcmaster.ca  Thu Mar 29 11:13:16 2007
From: cookedm at physics.mcmaster.ca (David M. Cooke)
Date: Thu, 29 Mar 2007 11:13:16 -0400
Subject: [SciPy-user] building numpy/scipy on Solaris
In-Reply-To: <13a1ee25e.e25e13a1e@steria.com>
References: <13a1ee25e.e25e13a1e@steria.com>
Message-ID: <460BD78C.8090800@physics.mcmaster.ca>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

raphael langella wrote:
> I just realized I've got much more recent versions of libsunperf and C
> and fortran compilers. I've got Sun Studio 10 & 11 installed. So I tried
> building numpy with both of them and each time I get this error when
> importing numpy :
> 
>>>> import numpy
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File "/tmp/lib/python2.5/site-packages/numpy/__init__.py", line 40, in
> <module>
>     import linalg
>   File "/tmp/lib/python2.5/site-packages/numpy/linalg/__init__.py", line
> 4, in <module>
>     from linalg import *
>   File "/tmp/lib/python2.5/site-packages/numpy/linalg/linalg.py", line
> 25, in <module>
>     from numpy.linalg import lapack_lite
> ImportError: ld.so.1: python: fatal: relocation error: file
> /tmp/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so: symbol
> __mt_MasterFunction_rtc_: referenced symbol not found
> 
> I tried the CPPFLAGS='-DNO_APPEND_FORTRAN' option, but it didn't change
> anything. Any idea?

Some library is not getting linked in. Best I can come up with is its
something to do with threads (mt == multithread? rtc == real-time
clock?). Maybe add '-mt' to the CFLAGS? That's equivalent to
- -D_REENTRANT -lthread.

- --
|>|\/|<
/------------------------------------------------------------------\
|David M. Cooke              http://arbutus.physics.mcmaster.ca/dmc/
|cookedm at physics.mcmaster.ca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGC9eM+kNzddXW8YwRAtonAKC1vK7o0APKkXlpPgQSH57AlZhAFgCg5eW1
2GTO3w6H6gxP6y80okigldc=
=Rab3
-----END PGP SIGNATURE-----


From david.huard at gmail.com  Thu Mar 29 21:07:54 2007
From: david.huard at gmail.com (David Huard)
Date: Thu, 29 Mar 2007 20:07:54 -0500
Subject: [SciPy-user] Numpy distutils failing to report ndarrayobject.h
In-Reply-To: <46040D99.6020509@ee.byu.edu>
References: <4603FF06.2090206@cse.ucsc.edu> <46040D99.6020509@ee.byu.edu>
Message-ID: <91cf711d0703291807w35e8730do343842413a38edc0@mail.gmail.com>

I checked and it's rather that ndarrayobject.h simply isn't in any folder.
If I remember correctly, on this installation numpy came installed with
enthought binaries (python24). Installing the numpy binaries from
scipy.orgsolves the problem.

David

2007/3/23, Travis Oliphant <oliphant at ee.byu.edu>:
>
> Anand Patil wrote:
>
> >Hi all,
> >
> >The following in a Pyrex extension module:
> >
> >cdef extern from "numpy/ndarrayobject.h":
> >    void* PyArray_DATA(object obj)
> >
> >produces the following in C:
> >
> >#include "numpy/ndarrayobject.h".
> >
> >The numpy distutils apparently fail to compile the extension module
> >under win2k, Python 2.4:
> >
> >PyMC2/PyrexLazyFunction.c:15:33: numpy/ndarrayobject.h: No such file or
> directory
> >PyMC2/PyrexLazyFunction.c: In function
> >`__pyx_f_17PyrexLazyFunction_12LazyFunction_get_array_data':
> >
> >
> >but it compiles fine on my machine, running OSX 10.4 and Python 2.5.
> >Unfortunately I can't experiment much because I don't have a Windows
> >machine available. Can anyone help me out with this?
> >
> >
>
> The include directory is not in the list of directories to search for
> the header files and so it is not being found.  I'm not sure what would
> cause this.  It's possible that the windows system doesn't have NumPy
> installed (or else the headers are not installed).  Or, it's possible
> that numpy.distutils is not adding the NumPy directory to the list of
> directories to search for header files.
>
> I'm not sure why it's not adding the location of the NumPy headers to
> the compile line.
>
> -Travis
>
>
>
> >Thanks,
> >Anand
> >_______________________________________________
> >SciPy-user mailing list
> >SciPy-user at scipy.org
> >http://projects.scipy.org/mailman/listinfo/scipy-user
> >
> >
> >
> >
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070329/b76483bb/attachment.html>

From hai at lpt.rwth-aachen.de  Fri Mar 30 10:05:38 2007
From: hai at lpt.rwth-aachen.de (Ri Hai)
Date: Fri, 30 Mar 2007 16:05:38 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
Message-ID: <1DA101C8C553404E8367810F73E3926A73E919@squirrel.lpt.rwth-aachen.de>

Hell folks,

 

When I use fsolve, python crashes! Now I am testing fsolve with the code
from the mailing list:
 
#-----------
from scipy.optimize import fsolve
 
 
def func2(x):
        out = [x[0]+2.*x[1]+2.*x[2]-1.]
        out.append(x[0]+x[2] - 2.*x[1])
        out.append(x[3]+x[4]-1.)
        out.append(1./5-2./5*x[2]+2./5*x[0]*x[3]-x[0])
        out.append(1./5-x[1]/5.-x[2]/5.+2/5.*x[1]*x[4]+x[1]/5.-x[1])
        #out.append(1/5-2/5*x[1]+2*x[2]/5+x[2]*x[3]/5)
        return out
 
x02 = fsolve(func2, [0.25,0.25,0.5,0.5,0.5])
 
#----------------
 
The code is taken form the mail from john in July 2006 and he said it
did work. However my python crashed again! I really don't know how to
solve this problem.  
 
I use python 2.5, scipy 0.5.2, numpy 1.01 and windows XP.
 
Has someone the same problem? Can somebody help me? Please...
 
Hai 
    
 

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070330/bc3cb6e6/attachment.html>

From hasslerjc at comcast.net  Fri Mar 30 10:25:09 2007
From: hasslerjc at comcast.net (John Hassler)
Date: Fri, 30 Mar 2007 10:25:09 -0400
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <1DA101C8C553404E8367810F73E3926A73E919@squirrel.lpt.rwth-aachen.de>
References: <1DA101C8C553404E8367810F73E3926A73E919@squirrel.lpt.rwth-aachen.de>
Message-ID: <460D1DC5.80305@comcast.net>

An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070330/cf85a494/attachment.html>

From nwagner at iam.uni-stuttgart.de  Fri Mar 30 10:25:29 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Fri, 30 Mar 2007 16:25:29 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <1DA101C8C553404E8367810F73E3926A73E919@squirrel.lpt.rwth-aachen.de>
References: <1DA101C8C553404E8367810F73E3926A73E919@squirrel.lpt.rwth-aachen.de>
Message-ID: <460D1DD9.101@iam.uni-stuttgart.de>

Ri Hai wrote:
>
> Hell folks,
>
> When I use fsolve, python crashes! Now I am testing fsolve with the code from the mailing list:
>  
> #-----------
> from scipy.optimize import fsolve
>  
>  
> def func2(x):
>         out = [x[0]+2.*x[1]+2.*x[2]-1.]
>         out.append(x[0]+x[2] - 2.*x[1])
>         out.append(x[3]+x[4]-1.)
>         out.append(1./5-2./5*x[2]+2./5*x[0]*x[3]-x[0])
>         out.append(1./5-x[1]/5.-x[2]/5.+2/5.*x[1]*x[4]+x[1]/5.-x[1])
>         #out.append(1/5-2/5*x[1]+2*x[2]/5+x[2]*x[3]/5)
>         return out
>  
> x02 = fsolve(func2, [0.25,0.25,0.5,0.5,0.5])
>  
> #----------------
> / /
> The code is taken form the mail from john in July 2006 and he said it did work. However my python crashed again! I really don?t know how to solve this problem.  
>  
> I use python 2.5, scipy 0.5.2, numpy 1.01 and windows XP.
>  
> Has someone the same problem? Can somebody help me? Please?
>  
> Hai 
>     
>  
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   

No problem here.
>>> numpy.__version__
'1.0.2.dev3617'
>>> scipy.__version__
'0.5.3.dev2892'

>>> x02
array([ 0.125 , 0.1875, 0.25 , 0.5 , 0.5 ])
>>> func2(x02)
[0.0, 0.0, 0.0, 3.4924590996965321e-11, -1.7462309376270468e-11]

What do you mean by crash ?


Nils



From hai at lpt.rwth-aachen.de  Fri Mar 30 10:48:05 2007
From: hai at lpt.rwth-aachen.de (Ri Hai)
Date: Fri, 30 Mar 2007 16:48:05 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <460D1DD9.101@iam.uni-stuttgart.de>
Message-ID: <1DA101C8C553404E8367810F73E3926A73E91A@squirrel.lpt.rwth-aachen.de>




-----Urspr?ngliche Nachricht-----
Von: scipy-user-bounces at scipy.org [mailto:scipy-user-bounces at scipy.org] Im Auftrag von Nils Wagner
Gesendet: Freitag, 30. M?rz 2007 16:25
An: SciPy Users List
Betreff: Re: [SciPy-user] how can I get fsolve() work?!

Ri Hai wrote:
>
> Hell folks,
>
> When I use fsolve, python crashes! Now I am testing fsolve with the code from the mailing list:
>  
> #-----------
> from scipy.optimize import fsolve
>  
>  
> def func2(x):
>         out = [x[0]+2.*x[1]+2.*x[2]-1.]
>         out.append(x[0]+x[2] - 2.*x[1])
>         out.append(x[3]+x[4]-1.)
>         out.append(1./5-2./5*x[2]+2./5*x[0]*x[3]-x[0])
>         out.append(1./5-x[1]/5.-x[2]/5.+2/5.*x[1]*x[4]+x[1]/5.-x[1])
>         #out.append(1/5-2/5*x[1]+2*x[2]/5+x[2]*x[3]/5)
>         return out
>  
> x02 = fsolve(func2, [0.25,0.25,0.5,0.5,0.5])
>  
> #----------------
> / /
> The code is taken form the mail from john in July 2006 and he said it did work. However my python crashed again! I really don't know how to solve this problem.  
>  
> I use python 2.5, scipy 0.5.2, numpy 1.01 and windows XP.
>  
> Has someone the same problem? Can somebody help me? Please...
>  
> Hai 
>     
>  
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   

No problem here.
>>> numpy.__version__
'1.0.2.dev3617'
>>> scipy.__version__
'0.5.3.dev2892'

>>> x02
array([ 0.125 , 0.1875, 0.25 , 0.5 , 0.5 ])
>>> func2(x02)
[0.0, 0.0, 0.0, 3.4924590996965321e-11, -1.7462309376270468e-11]

What do you mean by crash ?


Nils

---------------------------------
Hello Nils,


Using fsolve, I get the following error message (in German and I have translated it): 

"pythonw.exe ascertained a problem and python will be closed now."

So I could not even know what is wrong. 

I have tried with your commands. You can see the information below.    

>>> numpy.__version__
'1.0.1'
>>> scipy.__version__
'0.5.2'

Is there anything wrong with the version I have taken?


Hai





From nwagner at iam.uni-stuttgart.de  Fri Mar 30 10:55:20 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Fri, 30 Mar 2007 16:55:20 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <1DA101C8C553404E8367810F73E3926A73E91A@squirrel.lpt.rwth-aachen.de>
References: <1DA101C8C553404E8367810F73E3926A73E91A@squirrel.lpt.rwth-aachen.de>
Message-ID: <460D24D8.8020706@iam.uni-stuttgart.de>

Ri Hai wrote:
>
> -----Urspr?ngliche Nachricht-----
> Von: scipy-user-bounces at scipy.org [mailto:scipy-user-bounces at scipy.org] Im Auftrag von Nils Wagner
> Gesendet: Freitag, 30. M?rz 2007 16:25
> An: SciPy Users List
> Betreff: Re: [SciPy-user] how can I get fsolve() work?!
>
> Ri Hai wrote:
>   
>> Hell folks,
>>
>> When I use fsolve, python crashes! Now I am testing fsolve with the code from the mailing list:
>>  
>> #-----------
>> from scipy.optimize import fsolve
>>  
>>  
>> def func2(x):
>>         out = [x[0]+2.*x[1]+2.*x[2]-1.]
>>         out.append(x[0]+x[2] - 2.*x[1])
>>         out.append(x[3]+x[4]-1.)
>>         out.append(1./5-2./5*x[2]+2./5*x[0]*x[3]-x[0])
>>         out.append(1./5-x[1]/5.-x[2]/5.+2/5.*x[1]*x[4]+x[1]/5.-x[1])
>>         #out.append(1/5-2/5*x[1]+2*x[2]/5+x[2]*x[3]/5)
>>         return out
>>  
>> x02 = fsolve(func2, [0.25,0.25,0.5,0.5,0.5])
>>  
>> #----------------
>> / /
>> The code is taken form the mail from john in July 2006 and he said it did work. However my python crashed again! I really don't know how to solve this problem.  
>>  
>> I use python 2.5, scipy 0.5.2, numpy 1.01 and windows XP.
>>  
>> Has someone the same problem? Can somebody help me? Please...
>>  
>> Hai 
>>     
>>  
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> SciPy-user mailing list
>> SciPy-user at scipy.org
>> http://projects.scipy.org/mailman/listinfo/scipy-user
>>   
>>     
>
> No problem here.
>   
>>>> numpy.__version__
>>>>         
> '1.0.2.dev3617'
>   
>>>> scipy.__version__
>>>>         
> '0.5.3.dev2892'
>
>   
>>>> x02
>>>>         
> array([ 0.125 , 0.1875, 0.25 , 0.5 , 0.5 ])
>   
>>>> func2(x02)
>>>>         
> [0.0, 0.0, 0.0, 3.4924590996965321e-11, -1.7462309376270468e-11]
>
> What do you mean by crash ?
>
>
> Nils
>
> ---------------------------------
> Hello Nils,
>
>
> Using fsolve, I get the following error message (in German and I have translated it): 
>
> "pythonw.exe ascertained a problem and python will be closed now."
>
> So I could not even know what is wrong. 
>
> I have tried with your commands. You can see the information below.    
>
>   
>>>> numpy.__version__
>>>>         
> '1.0.1'
>   
>>>> scipy.__version__
>>>>         
> '0.5.2'
>
> Is there anything wrong with the version I have taken?
>
>   
I don't think so. It seems to be a Windows specific problem.

Nils

> Hai
>
>
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   

 


From hai at lpt.rwth-aachen.de  Fri Mar 30 11:14:07 2007
From: hai at lpt.rwth-aachen.de (Ri Hai)
Date: Fri, 30 Mar 2007 17:14:07 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <460D1DC5.80305@comcast.net>
Message-ID: <1DA101C8C553404E8367810F73E3926A73E91B@squirrel.lpt.rwth-aachen.de>

Hallo John,

 

I think there must be something wrong with my installation, but I don't know where. Please read the following message.   

 

>>> scipy.test()

 

Warning: FAILURE importing tests for <module 'scipy.linsolve.umfpack.umfpack' from '...y\\linsolve\\umfpack\\umfpack.pyc'>

C:\Python25\lib\site-packages\scipy\linsolve\umfpack\tests\test_umfpack.py:17: AttributeError: 'module' object has no attribute 'umfpack' (in <module>)

 

  Found 4 tests for scipy.io.array_import

  Found 1 tests for scipy.cluster.vq

  Found 128 tests for scipy.linalg.fblas

  Found 397 tests for scipy.ndimage

  Found 10 tests for scipy.integrate.quadpack

  Found 98 tests for scipy.stats.stats

  Found 53 tests for scipy.linalg.decomp

  Found 3 tests for scipy.integrate.quadrature

  Found 96 tests for scipy.sparse.sparse

  Found 20 tests for scipy.fftpack.pseudo_diffs

  Found 6 tests for scipy.optimize.optimize

  Found 6 tests for scipy.interpolate

  Found 12 tests for scipy.io.mmio

  Found 10 tests for scipy.stats.morestats

  Found 4 tests for scipy.linalg.lapack

  Found 4 tests for scipy.io.recaster

 

Warning: FAILURE importing tests for <module 'scipy.linsolve.umfpack' from '...\\linsolve\\umfpack\\__init__.pyc'>

C:\Python25\lib\site-packages\scipy\linsolve\umfpack\tests\test_umfpack.py:17: AttributeError: 'module' object has no attribute 'umfpack' (in <module>)

 

  Found 4 tests for scipy.optimize.zeros

  Found 6 tests for scipy.interpolate.fitpack

  Found 28 tests for scipy.io.mio

  Found 41 tests for scipy.linalg.basic

  Found 2 tests for scipy.maxentropy.maxentropy

  Found 358 tests for scipy.special.basic

  Found 128 tests for scipy.lib.blas.fblas

  Found 7 tests for scipy.linalg.matfuncs

  Found 42 tests for scipy.lib.lapack

  Found 18 tests for scipy.fftpack.basic

  Found 1 tests for scipy.optimize.cobyla

  Found 16 tests for scipy.lib.blas

  Found 1 tests for scipy.integrate

  Found 14 tests for scipy.linalg.blas

  Found 70 tests for scipy.stats.distributions

  Found 4 tests for scipy.fftpack.helper

  Found 4 tests for scipy.signal.signaltools

  Found 0 tests for __main__

 

Don't worry about a warning regarding the number of bytes read.

 

........................................................................................................................................................................................................................................................................................................................................................................................, at this point, python was closed.

 

Do you know why?    

 

 

Hai 

 

 

 

-----Urspr?ngliche Nachricht-----
Von: scipy-user-bounces at scipy.org [mailto:scipy-user-bounces at scipy.org] Im Auftrag von John Hassler
Gesendet: Freitag, 30. M?rz 2007 16:25
An: SciPy Users List
Betreff: Re: [SciPy-user] how can I get fsolve() work?!

 

I copy-pasted from your post, and it still works for me.  Does your scipy.test() work without errors?
john

Ri Hai wrote: 

Hell folks,

 

When I use fsolve, python crashes! Now I am testing fsolve with the code from the mailing list:
 
#-----------
from scipy.optimize import fsolve
 
 
def func2(x):
        out = [x[0]+2.*x[1]+2.*x[2]-1.]
        out.append(x[0]+x[2] - 2.*x[1])
        out.append(x[3]+x[4]-1.)
        out.append(1./5-2./5*x[2]+2./5*x[0]*x[3]-x[0])
        out.append(1./5-x[1]/5.-x[2]/5.+2/5.*x[1]*x[4]+x[1]/5.-x[1])
        #out.append(1/5-2/5*x[1]+2*x[2]/5+x[2]*x[3]/5)
        return out
 
x02 = fsolve(func2, [0.25,0.25,0.5,0.5,0.5])
 
#----------------
 
The code is taken form the mail from john in July 2006 and he said it did work. However my python crashed again! I really don't know how to solve this problem.  
 
I use python 2.5, scipy 0.5.2, numpy 1.01 and windows XP.
 
Has someone the same problem? Can somebody help me? Please...
 
Hai 
    
 

 

 

 



________________________________



 
_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user
  
 



________________________________



 
No virus found in this incoming message.
Checked by AVG Free Edition.
Version: 7.5.446 / Virus Database: 268.18.22/739 - Release Date: 3/29/2007 1:36 PM
  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070330/4676be76/attachment.html>

From elcorto at gmx.net  Fri Mar 30 11:25:14 2007
From: elcorto at gmx.net (Steve Schmerler)
Date: Fri, 30 Mar 2007 17:25:14 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <460D24D8.8020706@iam.uni-stuttgart.de>
References: <1DA101C8C553404E8367810F73E3926A73E91A@squirrel.lpt.rwth-aachen.de>
	<460D24D8.8020706@iam.uni-stuttgart.de>
Message-ID: <460D2BDA.4050201@gmx.net>

Nils Wagner wrote:
> Ri Hai wrote:
>> -----Urspr?ngliche Nachricht-----
>> Von: scipy-user-bounces at scipy.org [mailto:scipy-user-bounces at scipy.org] Im Auftrag von Nils Wagner
>> Gesendet: Freitag, 30. M?rz 2007 16:25
>> An: SciPy Users List
>> Betreff: Re: [SciPy-user] how can I get fsolve() work?!
>>
>> Ri Hai wrote:
>>   
>>> Hell folks,
>>>
>>> When I use fsolve, python crashes! Now I am testing fsolve with the code from the mailing list:
>>>  
>>> #-----------
>>> from scipy.optimize import fsolve
>>>  
>>>  
>>> def func2(x):
>>>         out = [x[0]+2.*x[1]+2.*x[2]-1.]
>>>         out.append(x[0]+x[2] - 2.*x[1])
>>>         out.append(x[3]+x[4]-1.)
>>>         out.append(1./5-2./5*x[2]+2./5*x[0]*x[3]-x[0])
>>>         out.append(1./5-x[1]/5.-x[2]/5.+2/5.*x[1]*x[4]+x[1]/5.-x[1])
>>>         #out.append(1/5-2/5*x[1]+2*x[2]/5+x[2]*x[3]/5)
>>>         return out
>>>  
>>> x02 = fsolve(func2, [0.25,0.25,0.5,0.5,0.5])
>>>  
>>> #----------------
>>> / /
>>> The code is taken form the mail from john in July 2006 and he said it did work. However my python crashed again! I really don't know how to solve this problem.  
>>>  
>>> I use python 2.5, scipy 0.5.2, numpy 1.01 and windows XP.
>>>  
>>> Has someone the same problem? Can somebody help me? Please...
>>>  
>>> Hai 
>>>     
>>>  
>>>
>>> ------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> SciPy-user mailing list
>>> SciPy-user at scipy.org
>>> http://projects.scipy.org/mailman/listinfo/scipy-user
>>>   
>>>     
>> No problem here.
>>   
>>>>> numpy.__version__
>>>>>         
>> '1.0.2.dev3617'
>>   
>>>>> scipy.__version__
>>>>>         
>> '0.5.3.dev2892'
>>
>>   
>>>>> x02
>>>>>         
>> array([ 0.125 , 0.1875, 0.25 , 0.5 , 0.5 ])
>>   
>>>>> func2(x02)
>>>>>         
>> [0.0, 0.0, 0.0, 3.4924590996965321e-11, -1.7462309376270468e-11]
>>
>> What do you mean by crash ?
>>
>>
>> Nils
>>
>> ---------------------------------
>> Hello Nils,
>>
>>
>> Using fsolve, I get the following error message (in German and I have translated it): 
>>
>> "pythonw.exe ascertained a problem and python will be closed now."
>>
>> So I could not even know what is wrong. 
>>
>> I have tried with your commands. You can see the information below.    
>>
>>   
>>>>> numpy.__version__
>>>>>         
>> '1.0.1'
>>   
>>>>> scipy.__version__
>>>>>         
>> '0.5.2'
>>
>> Is there anything wrong with the version I have taken?
>>
>>   
> I don't think so. It seems to be a Windows specific problem.
> 
> Nils
> 

This may be the case.

Some days ago I had to convert some code to work also under Windows 
(Python 2.5, scipy 0.5.2, numpy 1.0.1).

The code "crashed" with a message box saying the same thing as on Ri's 
machine (although it was "python.exe", not "pythonw.exe" I think). Part 
of the code uses scipy's odeint. Clicking on "Details" I found that the 
crash occured when Python tried to load _odepack.pyd (same as 
_odepack.so on *nix). Just for the record, with Python 2.4.4 under Win 
it gave me some DLL load error.

Testing it on other Win boxes (also XP running) the code executed 
without error. So I blamed a crappy Win installation and didn't 
investigate it further (I didn't try using functions that call other 
*.pyd libs like _zeros.pyd called by the brentq, brenth, bisect, for 
example).

hth
-- 
cheers,
steve

Random number generation is the art of producing pure gibberish as 
quickly as possible.


From elcorto at gmx.net  Fri Mar 30 11:33:44 2007
From: elcorto at gmx.net (Steve Schmerler)
Date: Fri, 30 Mar 2007 17:33:44 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <1DA101C8C553404E8367810F73E3926A73E91B@squirrel.lpt.rwth-aachen.de>
References: <1DA101C8C553404E8367810F73E3926A73E91B@squirrel.lpt.rwth-aachen.de>
Message-ID: <460D2DD8.8080903@gmx.net>

Ri Hai wrote:

[snip]

> 
> Warning: FAILURE importing tests for <module 'scipy.linsolve.umfpack' 
> from '...\\linsolve\\umfpack\\__init__.pyc'>
> 
> C:\Python25\lib\site-packages\scipy\linsolve\umfpack\tests\test_umfpack.py:17: 
> AttributeError: 'module' object has no attribute 'umfpack' (in <module>)
> 

[snip]


The UMFPACK maessage just tells you that your installation has no module 
scipy.linsolve.umfpack (the build of this was disabled when scipy was 
compliled) which shouldn't be something to worry about as long as you 
don't want to use it.

> 
> Don't worry about a warning regarding the number of bytes read.
> 
>  
> 
> ........................................................................................................................................................................................................................................................................................................................................................................................, 
> at this point, python was closed.
> 

Does this mean the test didn't finish? Did you see something like


#----------------------------------------------------------------------#

 >>> scipy.test()

[snip]

..zcopy:n=3
.............zscal:n=4
....zswap:n=4
..zswap:n=3
...Result may be inaccurate, approximate err = 6.99076835435e-09
...Result may be inaccurate, approximate err = 7.27595761418e-12
.............................................................Residual: 
1.05006950608e-07
.
****************************************************************
WARNING: cblas module is empty
-----------
See scipy/INSTALL.txt for troubleshooting.
Notes:
* If atlas library is not found by numpy/distutils/system_info.py,
   then scipy uses fblas instead of cblas.
****************************************************************

......................
----------------------------------------------------------------------
Ran 1602 tests in 3.848s

OK
<unittest.TextTestRunner object at 0xa6133d6c>
 >>>

#----------------------------------------------------------------------#

-- 
cheers,
steve

Random number generation is the art of producing pure gibberish as 
quickly as possible.


From hai at lpt.rwth-aachen.de  Fri Mar 30 11:45:45 2007
From: hai at lpt.rwth-aachen.de (Ri Hai)
Date: Fri, 30 Mar 2007 17:45:45 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <460D2DD8.8080903@gmx.net>
Message-ID: <1DA101C8C553404E8367810F73E3926A73E91C@squirrel.lpt.rwth-aachen.de>

Hello Steve,

you are right, the test didn't finish and I didn't see any text that you added.

Now I am reading the information about installation in Scipy homepage. Maybe I have done something wrong. If you have any idea, please let me know. Thank you!

Hai

   

-----Urspr?ngliche Nachricht-----
Von: scipy-user-bounces at scipy.org [mailto:scipy-user-bounces at scipy.org] Im Auftrag von Steve Schmerler
Gesendet: Freitag, 30. M?rz 2007 17:34
An: SciPy Users List
Betreff: Re: [SciPy-user] how can I get fsolve() work?!

Ri Hai wrote:

[snip]

> 
> Warning: FAILURE importing tests for <module 'scipy.linsolve.umfpack' 
> from '...\\linsolve\\umfpack\\__init__.pyc'>
> 
> C:\Python25\lib\site-packages\scipy\linsolve\umfpack\tests\test_umfpack.py:17: 
> AttributeError: 'module' object has no attribute 'umfpack' (in <module>)
> 

[snip]


The UMFPACK maessage just tells you that your installation has no module 
scipy.linsolve.umfpack (the build of this was disabled when scipy was 
compliled) which shouldn't be something to worry about as long as you 
don't want to use it.

> 
> Don't worry about a warning regarding the number of bytes read.
> 
>  
> 
> ........................................................................................................................................................................................................................................................................................................................................................................................, 
> at this point, python was closed.
> 

Does this mean the test didn't finish? Did you see something like


#----------------------------------------------------------------------#

 >>> scipy.test()

[snip]

..zcopy:n=3
.............zscal:n=4
....zswap:n=4
..zswap:n=3
...Result may be inaccurate, approximate err = 6.99076835435e-09
...Result may be inaccurate, approximate err = 7.27595761418e-12
.............................................................Residual: 
1.05006950608e-07
.
****************************************************************
WARNING: cblas module is empty
-----------
See scipy/INSTALL.txt for troubleshooting.
Notes:
* If atlas library is not found by numpy/distutils/system_info.py,
   then scipy uses fblas instead of cblas.
****************************************************************

......................
----------------------------------------------------------------------
Ran 1602 tests in 3.848s

OK
<unittest.TextTestRunner object at 0xa6133d6c>
 >>>

#----------------------------------------------------------------------#

-- 
cheers,
steve

Random number generation is the art of producing pure gibberish as 
quickly as possible.
_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user




From elcorto at gmx.net  Fri Mar 30 12:11:34 2007
From: elcorto at gmx.net (Steve Schmerler)
Date: Fri, 30 Mar 2007 18:11:34 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <1DA101C8C553404E8367810F73E3926A73E91C@squirrel.lpt.rwth-aachen.de>
References: <1DA101C8C553404E8367810F73E3926A73E91C@squirrel.lpt.rwth-aachen.de>
Message-ID: <460D36B6.7040205@gmx.net>

Ri Hai wrote:
> Hello Steve,
> 
> you are right, the test didn't finish and I didn't see any text that you added.

With the test, does it show the same error message box as it does if you
run your fsolve() example? If so, can you click on "details" (I think
upper right corner of the box) and see if it's related to some kind of
*.pyd file?

> 
> Now I am reading the information about installation in Scipy homepage. Maybe I have done something wrong. If you have any idea, please let me know. Thank you!
> 

On Win, it's usually just installing Python (python-2.5.msi or
something), numpy  (numpy-1.0.1.win32-py2.5.exe) and scipy
(scipy-0.5.2.win32-py2.5.exe). If you plan to build things from source,
that's another story (haven't done that on Win).

Anyway, if it is the same problem that I had (loading a *.pyd lib) I
suspect that it's not a scipy-specific issue but rather a Windows
problem (but I'm not an expert here...). If you have access to another
Win box, I suggest that you try running you example there before doing
anything else.

-- 
cheers,
steve

I love deadlines. I like the whooshing sound they make as they fly by.
-- Douglas Adams



From fccoelho at gmail.com  Fri Mar 30 12:18:07 2007
From: fccoelho at gmail.com (Flavio Coelho)
Date: Fri, 30 Mar 2007 13:18:07 -0300
Subject: [SciPy-user] returning an array from weave inline
In-Reply-To: <460B6E94.8010708@itc.it>
References: <d9af7a880703281057y55f8e063u239753fa8a774346@mail.gmail.com>
	<db6b5ecc0703281105p6229955dxc97eae2c999ca884@mail.gmail.com>
	<d9af7a880703281518o338d5165je0ae8b5e1580899@mail.gmail.com>
	<460B6E94.8010708@itc.it>
Message-ID: <d9af7a880703300918x2241e1d9o4931b416c87620ef@mail.gmail.com>

Yes Emanuele, that makes complete sense.

thanks,

Fl?vio

On 3/29/07, Emanuele Olivetti <olivetti at itc.it> wrote:
>
> Flavio Coelho wrote:
> > Thanks Fernando,
> >
> > I thought variables converted to C were copies, so that there could not
> be
> > an in place operation on a Python variable.
>
> They are copies like it is in the C way of thinking: if you pass a
> basic type (e.g. an int) to the inlined C code and modify it then
> when you are back to python there is no change in the initial variable
> you passed since just the copy was modified. If you pass something more
> complex
> (e.g. a numpy array) then you are just giving a copy of the it's reference
> (or pointer) to that function. With that reference (or copy of) you
> can access the the _original_ array and modify it. So when you are back to
> python your array _is_ changed. This mechanism allows better efficiency:
> what happens if you have a huge 2Gb array and pass it to the inline code?
> It's not desirable to make a whole copy and allocate other 2Gb or RAM...
>
> Anyway python itself works like that. E.g.:
> ----
> def f(a,b):
>      a=a+1
>      b.append(1)
>      return
>
> a=1
> b=[3,2]
> print a,b
> f(a,b)
> print a,b
> ----
>
> Hope this helps,
>
> Emanuele
>
> ------------------
> ITC -> dall'1 marzo 2007 Fondazione Bruno Kessler
> ITC -> since 1 March 2007 Fondazione Bruno Kessler
> ------------------
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>



-- 
Fl?vio Code?o Coelho
registered Linux user # 386432
get counted at http://counter.li.org
---------------------------
"software gets slower faster than hardware gets faster"
Niklaus Wirth's law
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070330/b9dbab08/attachment.html>

From hai at lpt.rwth-aachen.de  Fri Mar 30 12:26:37 2007
From: hai at lpt.rwth-aachen.de (Ri Hai)
Date: Fri, 30 Mar 2007 18:26:37 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <460D36B6.7040205@gmx.net>
Message-ID: <1DA101C8C553404E8367810F73E3926A73E91D@squirrel.lpt.rwth-aachen.de>


Hello Steve,

> With the test, does it show the same error message box as it does if you
> run your fsolve() example? 

Yes.


> If so, can you click on "details" (I think upper right corner of the box) > and see if it's related to some kind of *.pyd file?

Yes!

> Anyway, if it is the same problem that I had (loading a *.pyd lib) I
> suspect that it's not a scipy-specific issue but rather a Windows
> problem (but I'm not an expert here...).

I think you are right!!

> If you have access to another Win box, I suggest that you try running you > example there before doing anything else.

What do you mean by Win box? 


Hai



-----Urspr?ngliche Nachricht-----
Von: scipy-user-bounces at scipy.org [mailto:scipy-user-bounces at scipy.org] Im Auftrag von Steve Schmerler
Gesendet: Freitag, 30. M?rz 2007 18:12
An: SciPy Users List
Betreff: Re: [SciPy-user] how can I get fsolve() work?!

Ri Hai wrote:
> Hello Steve,
> 
> you are right, the test didn't finish and I didn't see any text that you added.

With the test, does it show the same error message box as it does if you
run your fsolve() example? If so, can you click on "details" (I think
upper right corner of the box) and see if it's related to some kind of
*.pyd file?

> 
> Now I am reading the information about installation in Scipy homepage. Maybe I have done something wrong. If you have any idea, please let me know. Thank you!
> 

On Win, it's usually just installing Python (python-2.5.msi or
something), numpy  (numpy-1.0.1.win32-py2.5.exe) and scipy
(scipy-0.5.2.win32-py2.5.exe). If you plan to build things from source,
that's another story (haven't done that on Win).

Anyway, if it is the same problem that I had (loading a *.pyd lib) I
suspect that it's not a scipy-specific issue but rather a Windows
problem (but I'm not an expert here...). If you have access to another
Win box, I suggest that you try running you example there before doing
anything else.

-- 
cheers,
steve

I love deadlines. I like the whooshing sound they make as they fly by.
-- Douglas Adams

_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user




From elcorto at gmx.net  Fri Mar 30 13:05:32 2007
From: elcorto at gmx.net (Steve Schmerler)
Date: Fri, 30 Mar 2007 19:05:32 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <1DA101C8C553404E8367810F73E3926A73E91D@squirrel.lpt.rwth-aachen.de>
References: <1DA101C8C553404E8367810F73E3926A73E91D@squirrel.lpt.rwth-aachen.de>
Message-ID: <460D435C.30703@gmx.net>

Ri Hai wrote:

>> If you have access to another Win box, I suggest that you try running you > example there before doing anything else.
> 
> What do you mean by Win box? 
> 

A computer running Windows :)


-- 
cheers,
steve

I love deadlines. I like the whooshing sound they make as they fly by.
-- Douglas Adams



From eric at deeplycloudy.com  Fri Mar 30 13:27:55 2007
From: eric at deeplycloudy.com (Eric Bruning)
Date: Fri, 30 Mar 2007 12:27:55 -0500
Subject: [SciPy-user] voronoi diagrams and the delaunay module
Message-ID: <be09c002f137de276b58e0fb4dc7f8f9@deeplycloudy.com>

I'd like to retrieve the Voronoi diagram for a set of points. 
Internally, the delaunay module (in the sandbox) appears calculate the 
voronoi diagram, but I don't think it's exposed in python.

What would need to be done do to enable voronoi output?

Eventually, I'm hoping to pass the polygons of the voronoi diagram to a 
polycollection in matplotlib...

Eric



From jtravs at gmail.com  Fri Mar 30 13:44:49 2007
From: jtravs at gmail.com (John Travers)
Date: Fri, 30 Mar 2007 18:44:49 +0100
Subject: [SciPy-user] collocation code for boundary value ODEs
Message-ID: <3a1077e70703301044j1f5252d6xce20126139daddcb@mail.gmail.com>

Hi all,
I'm wondering if there is any python code in scipy or elsewhere which
implements a collocation method for solving boundary value ODEs?
Something similar to or based on COLNEW or COLSYS from netlib.
I'd be very happy if there is. If not, then I'll write my own and post
it back here.
Thanks for any help,
John Travers


From robert.kern at gmail.com  Fri Mar 30 14:15:13 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 30 Mar 2007 13:15:13 -0500
Subject: [SciPy-user] voronoi diagrams and the delaunay module
In-Reply-To: <be09c002f137de276b58e0fb4dc7f8f9@deeplycloudy.com>
References: <be09c002f137de276b58e0fb4dc7f8f9@deeplycloudy.com>
Message-ID: <460D53B1.8010200@gmail.com>

Eric Bruning wrote:
> I'd like to retrieve the Voronoi diagram for a set of points. 
> Internally, the delaunay module (in the sandbox) appears calculate the 
> voronoi diagram, but I don't think it's exposed in python.
> 
> What would need to be done do to enable voronoi output?
> 
> Eventually, I'm hoping to pass the polygons of the voronoi diagram to a 
> polycollection in matplotlib...

All of the information that you need and that the algorithm can provide is
already given. Just connect the circumcenter of each triangle to the
circumcenters of the triangle's neighbors. I left out doing much more because I
have no idea what concrete formats consumers might need. Voronoi diagrams are
somewhat abstract beasts; the circumcenters and the connections between them
(and a point at infinity) are all that is really defined for them. Doing
something concrete with them requires more information than a general tool
really has available. For example, if you are going to draw the lines out to
infinity, how should I represent them?

Forming them into polygons is a bit tricky, but not too bad. For each point in
the interior of the triangulation (i.e., points not on the convex hull), walk
around the triangles that impinge on it counterclockwise and connect their
circumcenters. To handle the edges of the Voronoi diagram that go out to
infinity, you need to know the bounding box of your display. Walk around the
edges of the convex hull, find the circumcenter of the triangle with the edge
and connect it to a point outside the display bounding box along the line from
the circumcenter perpendicular to the convex hull edge. Now walk around the
points of the convex hull, and make a Voronoi polygon for each of them using the
appropriate "points at infinity" that you made.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


From emmanuelj at aims.ac.za  Fri Mar 30 17:35:50 2007
From: emmanuelj at aims.ac.za (Emmanuel Ohieku Jonah)
Date: Fri, 30 Mar 2007 23:35:50 +0200
Subject: [SciPy-user] weave.blitz
Message-ID: <460D82B6.7020003@aims.ac.za>

Hi All,

I need a simple code to illustrate how weave.blitz works. A working
example might help.

Thanks.

-- 
Emmanuel O. Jonah
African Institute for Mathematical Sciences
No. 6 Melrose road
Muizenberg, 7945
Cape Town
South Africa.
Mobile Phone: +27 78 324485
email: emmanuelj at aims.ac.za
       emijones2001 at yahoo.com



From rhc28 at cornell.edu  Fri Mar 30 18:24:10 2007
From: rhc28 at cornell.edu (Rob Clewley)
Date: Fri, 30 Mar 2007 18:24:10 -0400
Subject: [SciPy-user] collocation code for boundary value ODEs
In-Reply-To: <3a1077e70703301044j1f5252d6xce20126139daddcb@mail.gmail.com>
References: <3a1077e70703301044j1f5252d6xce20126139daddcb@mail.gmail.com>
Message-ID: <a749952d0703301524p5250da76n4a0a1aebc1a12a7c@mail.gmail.com>

John,

Sorry it won't be very soon but members of my group and I are actively
working on an interface to AUTO's collocation routines that will allow
us to solve BVPs as part of PyDSTool (and in a much more user-friendly
way than the native AUTO setup). If you would like to contribute to
that work it would move it along a lot quicker, but of course that
would require you to learn and use the PyDSTool environment for your
computations!

Otherwise, to my knowledge there hasn't been any noise made about new
BVP solvers involving Scipy since the last time your question was
asked here a few months ago.

-Rob


On 30/03/07, John Travers <jtravs at gmail.com> wrote:
> Hi all,
> I'm wondering if there is any python code in scipy or elsewhere which
> implements a collocation method for solving boundary value ODEs?
> Something similar to or based on COLNEW or COLSYS from netlib.
> I'd be very happy if there is. If not, then I'll write my own and post
> it back here.
> Thanks for any help,
> John Travers
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From ckkart at hoc.net  Fri Mar 30 20:48:41 2007
From: ckkart at hoc.net (Christian K)
Date: Sat, 31 Mar 2007 09:48:41 +0900
Subject: [SciPy-user] weave.blitz
In-Reply-To: <460D82B6.7020003@aims.ac.za>
References: <460D82B6.7020003@aims.ac.za>
Message-ID: <eukb50$vbk$1@sea.gmane.org>

Emmanuel Ohieku Jonah wrote:
> Hi All,
> 
> I need a simple code to illustrate how weave.blitz works. A working
> example might help.

I first read this: http://www.scipy.org/PerformancePython

the weave docs are at:
http://projects.scipy.org/scipy/scipy/browser/trunk/Lib/weave/doc/tutorial.html?format=raw

Christian



From hai at lpt.rwth-aachen.de  Sat Mar 31 06:45:11 2007
From: hai at lpt.rwth-aachen.de (Ri Hai)
Date: Sat, 31 Mar 2007 12:45:11 +0200
Subject: [SciPy-user] how can I get fsolve() work?!
In-Reply-To: <460D435C.30703@gmx.net>
Message-ID: <1DA101C8C553404E8367810F73E3926A73E91E@squirrel.lpt.rwth-aachen.de>


Oh! :-)
Thanks, I will try it. 

Ri

Ri Hai wrote:

>> If you have access to another Win box, I suggest that you try running
you > example there before doing anything else.
> 
> What do you mean by Win box? 
> 

A computer running Windows :)


-- 
cheers,
steve

I love deadlines. I like the whooshing sound they make as they fly by.
-- Douglas Adams

_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.org
http://projects.scipy.org/mailman/listinfo/scipy-user




From jtravs at gmail.com  Sat Mar 31 08:00:06 2007
From: jtravs at gmail.com (John Travers)
Date: Sat, 31 Mar 2007 13:00:06 +0100
Subject: [SciPy-user] collocation code for boundary value ODEs
In-Reply-To: <a749952d0703301524p5250da76n4a0a1aebc1a12a7c@mail.gmail.com>
References: <3a1077e70703301044j1f5252d6xce20126139daddcb@mail.gmail.com>
	<a749952d0703301524p5250da76n4a0a1aebc1a12a7c@mail.gmail.com>
Message-ID: <3a1077e70703310500w71da3b61of994a2379fda55a6@mail.gmail.com>

Hi Rob, all,
Thanks for the info. PyDSTool looks good, I didn't know about this
package. However, I have found a quicker solution to my problem, a
direct wrapper of COLNEW:

http://www.elisanet.fi/ptvirtan/software/bvp/index.html

by  Pauli Virtanen.

Thanks again,
John


From rhc28 at cornell.edu  Sat Mar 31 11:26:10 2007
From: rhc28 at cornell.edu (Rob Clewley)
Date: Sat, 31 Mar 2007 11:26:10 -0400
Subject: [SciPy-user] collocation code for boundary value ODEs
In-Reply-To: <3a1077e70703310500w71da3b61of994a2379fda55a6@mail.gmail.com>
References: <3a1077e70703301044j1f5252d6xce20126139daddcb@mail.gmail.com>
	<a749952d0703301524p5250da76n4a0a1aebc1a12a7c@mail.gmail.com>
	<3a1077e70703310500w71da3b61of994a2379fda55a6@mail.gmail.com>
Message-ID: <a749952d0703310826m2a32dd3dre92eceaeb1c7f11a@mail.gmail.com>

This package might be very useful to us, too. Thanks for sharing!
Rob


From cclarke at chrisdev.com  Sat Mar 31 12:53:51 2007
From: cclarke at chrisdev.com (Christopher Clarke)
Date: Sat, 31 Mar 2007 12:53:51 -0400
Subject: [SciPy-user] Spyce and Wingware
Message-ID: <C5FF1C6B-9705-43DF-9DA0-AEB91E0EBB35@chrisdev.com>

Hi All
Anybody manage to configure Wing and spyce??
Regards
Chris


From silva at crans.org  Sat Mar 31 13:57:34 2007
From: silva at crans.org (Fabrice Silva)
Date: Sat, 31 Mar 2007 17:57:34 +0000 (UTC)
Subject: [SciPy-user] invfreq
Message-ID: <pan.2007.03.31.17.57.34@crans.org>

Hello to all,
I wonder there is no 'invfreq' function in scipy.signal like in octave 
(http://www.koders.com/matlab/fidF8CAA512B66C33B66B3C078DE89255A3BBA49C55.aspx)

Is there another implementation or an equivalent function?