From openopt at ukr.net  Sun Jul  1 07:16:35 2007
From: openopt at ukr.net (dmitrey)
Date: Sun, 01 Jul 2007 14:16:35 +0300
Subject: [SciPy-user] once again about zeros() and ones()
Message-ID: <46878D13.10505@ukr.net>

Hi all,
There already was some discussion about zeros() and ones() usage, but I 
still can't understand some things.
1) currently ones(4,5, **kwargs) just yields error. Why it can't be 
translated to ones((4,5), **kwargs) ?
See my example of Ones() below.
##############################
from numpy import ones
def Ones(*args, **kwargs):
    if type(args[0]) in (type(()), type([])):
        return ones(*args, **kwargs)
    else: return ones(tuple(args), **kwargs)
#############################
if __name__ == '__main__':
    print Ones((2,2))
    print 2*Ones([2,2])
    print 3*Ones(2,2)   
    print 4*Ones(2,2,2, dtype=float)
    print 5*Ones((2,2,2), dtype='float')
    print 6*Ones([2,2,2], dtype='int')
    print 7*Ones(3,3, dtype=int, order = 'C')
    print 8*Ones((3,3), dtype='int')
    print 9*Ones(3, dtype='int')
    print 10*Ones((3,), dtype='int')
#############################

As one of the former MATLAB users, I'm very disappointed in using 
zeros((...)) instead of zeros(...), because I'm one of big number of 
those who should type the zeros() many times per day.
Also, same to ones(), tile(), reshape().
2) what does "order" param means? (in 
zeros((d1,...,dn),dtype=float,order='C') )
Currently help(zeros) just says
zeros(...)
    zeros((d1,...,dn),dtype=float,order='C')
    Return a new array of shape (d1,...,dn) and type typecode with all
    it's entries initialized to zero.
So nothing about "order" is known

Maybe you have already answered it somewhere, but I guess it's better to 
update ones() and zeros() documentation  than answer the question again 
and again in mailing lists.

Regards, D.



From fredmfp at gmail.com  Sun Jul  1 07:36:21 2007
From: fredmfp at gmail.com (fred)
Date: Sun, 01 Jul 2007 13:36:21 +0200
Subject: [SciPy-user] once again about zeros() and ones()
In-Reply-To: <46878D13.10505@ukr.net>
References: <46878D13.10505@ukr.net>
Message-ID: <468791B5.1000906@gmail.com>

dmitrey a ?crit :
> 2) what does "order" param means? (in 
> zeros((d1,...,dn),dtype=float,order='C') )
>   
I guess it is related to how data are stored : 'C order' or 'Fortran order'
which are not the same, IIRC.

Try with order='C' and order='F'.

My 2 cts.

-- 
http://scipy.org/FredericPetit



From matthieu.brucher at gmail.com  Sun Jul  1 08:45:12 2007
From: matthieu.brucher at gmail.com (Matthieu Brucher)
Date: Sun, 1 Jul 2007 14:45:12 +0200
Subject: [SciPy-user] once again about zeros() and ones()
In-Reply-To: <46878D13.10505@ukr.net>
References: <46878D13.10505@ukr.net>
Message-ID: <e76aa17f0707010545y18744281k922badf24566d82f@mail.gmail.com>

2007/7/1, dmitrey <openopt at ukr.net>:
>
> Hi all,
> There already was some discussion about zeros() and ones() usage, but I
> still can't understand some things.
> 1) currently ones(4,5, **kwargs) just yields error. Why it can't be
> translated to ones((4,5), **kwargs) ?
> See my example of Ones() below.
> ##############################
> from numpy import ones
> def Ones(*args, **kwargs):
>     if type(args[0]) in (type(()), type([])):
>         return ones(*args, **kwargs)
>     else: return ones(tuple(args), **kwargs)
> #############################
> if __name__ == '__main__':
>     print Ones((2,2))
>     print 2*Ones([2,2])
>     print 3*Ones(2,2)
>     print 4*Ones(2,2,2, dtype=float)
>     print 5*Ones((2,2,2), dtype='float')
>     print 6*Ones([2,2,2], dtype='int')
>     print 7*Ones(3,3, dtype=int, order = 'C')
>     print 8*Ones((3,3), dtype='int')
>     print 9*Ones(3, dtype='int')
>     print 10*Ones((3,), dtype='int')
> #############################
>

The first argument is a shape, the second is the type of the data. Python
has no way of knowing that the second argument is part of the shape and not
the type. This way, the arguments are defined correctly, although you could
make a wrapper *args, **kwargs and forcing people to use explicitely
dtype=float, which they won't do, thus leading to a lot of trouble and
errors.

Matthieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070701/8e61657a/attachment.html>

From openopt at ukr.net  Sun Jul  1 09:10:43 2007
From: openopt at ukr.net (dmitrey)
Date: Sun, 01 Jul 2007 16:10:43 +0300
Subject: [SciPy-user] once again about zeros() and ones()
In-Reply-To: <e76aa17f0707010545y18744281k922badf24566d82f@mail.gmail.com>
References: <46878D13.10505@ukr.net>
	<e76aa17f0707010545y18744281k922badf24566d82f@mail.gmail.com>
Message-ID: <4687A7D3.6030605@ukr.net>

I can't understand, which example of calling current numpy.zeros(...) is 
incompatible with mine, i.e. which one will yield other results?
D.

Matthieu Brucher wrote:
>
>
> 2007/7/1, dmitrey <openopt at ukr.net <mailto:openopt at ukr.net>>:
>
>     Hi all,
>     There already was some discussion about zeros() and ones() usage,
>     but I
>     still can't understand some things.
>     1) currently ones(4,5, **kwargs) just yields error. Why it can't be
>     translated to ones((4,5), **kwargs) ?
>     See my example of Ones() below.
>     ##############################
>     from numpy import ones
>     def Ones(*args, **kwargs):
>         if type(args[0]) in (type(()), type([])):
>             return ones(*args, **kwargs)
>         else: return ones(tuple(args), **kwargs)
>     #############################
>     if __name__ == '__main__':
>         print Ones((2,2))
>         print 2*Ones([2,2])
>         print 3*Ones(2,2)
>         print 4*Ones(2,2,2, dtype=float)
>         print 5*Ones((2,2,2), dtype='float')
>         print 6*Ones([2,2,2], dtype='int')
>         print 7*Ones(3,3, dtype=int, order = 'C')
>         print 8*Ones((3,3), dtype='int')
>         print 9*Ones(3, dtype='int')
>         print 10*Ones((3,), dtype='int')
>     #############################
>
>
> The first argument is a shape, the second is the type of the data. 
> Python has no way of knowing that the second argument is part of the 
> shape and not the type. This way, the arguments are defined correctly, 
> although you could make a wrapper *args, **kwargs and forcing people 
> to use explicitely dtype=float, which they won't do, thus leading to a 
> lot of trouble and errors.
>
> Matthieu
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   



From wjdandreta at att.net  Sun Jul  1 09:22:51 2007
From: wjdandreta at att.net (Bill Dandreta)
Date: Sun, 01 Jul 2007 09:22:51 -0400
Subject: [SciPy-user] problems with scipy in gentoo
In-Reply-To: <11f4deca0706301729g123b6200xc707b1ba240b40d2@mail.gmail.com>
References: <11f4deca0706301729g123b6200xc707b1ba240b40d2@mail.gmail.com>
Message-ID: <4687AAAB.4050400@att.net>

Angel Yanguas-Gil wrote:
> the scipy package in gentoo is currently masked and cannot be
> installed. The problem is not in scipy but in one of the dependencies
> which, at least in my case, is blas. Does anybody know if it is
> possible to install scipy with an alternative library?
>   
Below are notes I wrote to myself to get scipy and matplotlib to install 
on gentoo:

Put these lines in /etc/portage/package.keywords

app-admin/eselect-blas ~amd64
app-admin/eselect-lapack ~amd64
dev-python/matplotlib ~amd64
dev-python/numpy ~amd64
dev-python/python-dateutil ~amd64
dev-python/pytz ~amd64
sci-libs/scipy amd64
---

emerge -av blas-reference eselect eselect-blas virtual/blas

#scipy requires lapack and numpy, matplotlib requires numpy.
#numpy does not require lapack but can use it.
#lapack-reference requires a fortran 77 compiler.
#gcc-4.X has gfortran, a fortran 95 compiler.
#Need to use gcc-3.x

emerge -av =sys-devel/gcc-3*

#switch compilers:
gcc-config -l
gcc-config x86_64-pc-linux-gnu-3.4.6
source /etc/profile

emerge -av virtual/lapack lapack-reference eselect-lapack

#switch compilers back:
gcc-config x86_64-pc-linux-gnu-4.1.2
source /etc/profile

emerge -av numpy matplotlib scipy

I have not gotten eselect-blas or eselect-lapack to work yet (maybe 
that's why they are masked <g>) but with blas-reference virtual/blas 
lapack-reference and virtual/lapack installed, portage is smart enough 
to use them when installing scipy. I've used blas-atlas and lapack-atlas 
in the past but since they take so long to compile and I don't do much 
linear algebra, the reference ones are OK for me.

Bill
-- 

Bill

wjdandreta at att.net

Gentoo Linux 2.6.20-gentoo-r8

Reclaim Your Inbox with http://www.mozilla.org/products/thunderbird/

All things cometh to he who waiteth as long as he who waiteth worketh like hell while he waiteth.



From openopt at ukr.net  Sun Jul  1 09:46:17 2007
From: openopt at ukr.net (dmitrey)
Date: Sun, 01 Jul 2007 16:46:17 +0300
Subject: [SciPy-user] sparse stuff in numpy/scipy
Message-ID: <4687B029.9070609@ukr.net>

Hi all,
I had heard there are some sparse matrix abilities in scipy (btw don't 
you think it's better to store those one in numpy?).
However, neither dir(scipy), nor dir(numpy), nor 
http://www.scipy.org/Topical_Software doesn't tell me what should I use.
Can scipy team just decide, which way sparse stuff will be implemented? 
Either it will be own numpy/scipy package, or some 3rd-party package 
will be connected, or maybe one of well-known Fortran/C sparse libraries 
will be connected? I didn't demand any excellent benchmark results, I 
need just something working + having status of common scipy/numpy users 
standard (like matplotlib has), that will be maintained - may be not 
very intensively, but at least somehow.
And, of course, it's highly preferable to have OSI-approved license 
without copyleft.

Regards, D.



From matthieu.brucher at gmail.com  Sun Jul  1 11:13:17 2007
From: matthieu.brucher at gmail.com (Matthieu Brucher)
Date: Sun, 1 Jul 2007 17:13:17 +0200
Subject: [SciPy-user] once again about zeros() and ones()
In-Reply-To: <4687A7D3.6030605@ukr.net>
References: <46878D13.10505@ukr.net>
	<e76aa17f0707010545y18744281k922badf24566d82f@mail.gmail.com>
	<4687A7D3.6030605@ukr.net>
Message-ID: <e76aa17f0707010813r5e286916h3feaffeccbbd8159@mail.gmail.com>

numpy.zeros((2, 3, 4), float) for instance.

Matthieu

2007/7/1, dmitrey <openopt at ukr.net>:
>
> I can't understand, which example of calling current numpy.zeros(...) is
> incompatible with mine, i.e. which one will yield other results?
> D.
>
> Matthieu Brucher wrote:
> >
> >
> > 2007/7/1, dmitrey <openopt at ukr.net <mailto:openopt at ukr.net>>:
> >
> >     Hi all,
> >     There already was some discussion about zeros() and ones() usage,
> >     but I
> >     still can't understand some things.
> >     1) currently ones(4,5, **kwargs) just yields error. Why it can't be
> >     translated to ones((4,5), **kwargs) ?
> >     See my example of Ones() below.
> >     ##############################
> >     from numpy import ones
> >     def Ones(*args, **kwargs):
> >         if type(args[0]) in (type(()), type([])):
> >             return ones(*args, **kwargs)
> >         else: return ones(tuple(args), **kwargs)
> >     #############################
> >     if __name__ == '__main__':
> >         print Ones((2,2))
> >         print 2*Ones([2,2])
> >         print 3*Ones(2,2)
> >         print 4*Ones(2,2,2, dtype=float)
> >         print 5*Ones((2,2,2), dtype='float')
> >         print 6*Ones([2,2,2], dtype='int')
> >         print 7*Ones(3,3, dtype=int, order = 'C')
> >         print 8*Ones((3,3), dtype='int')
> >         print 9*Ones(3, dtype='int')
> >         print 10*Ones((3,), dtype='int')
> >     #############################
> >
> >
> > The first argument is a shape, the second is the type of the data.
> > Python has no way of knowing that the second argument is part of the
> > shape and not the type. This way, the arguments are defined correctly,
> > although you could make a wrapper *args, **kwargs and forcing people
> > to use explicitely dtype=float, which they won't do, thus leading to a
> > lot of trouble and errors.
> >
> > Matthieu
> >
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > 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/20070701/5cc6168c/attachment.html>

From openopt at ukr.net  Sun Jul  1 13:12:12 2007
From: openopt at ukr.net (dmitrey)
Date: Sun, 01 Jul 2007 20:12:12 +0300
Subject: [SciPy-user] once again about zeros() and ones()
In-Reply-To: <e76aa17f0707010813r5e286916h3feaffeccbbd8159@mail.gmail.com>
References: <46878D13.10505@ukr.net>	<e76aa17f0707010545y18744281k922badf24566d82f@mail.gmail.com>	<4687A7D3.6030605@ukr.net>
	<e76aa17f0707010813r5e286916h3feaffeccbbd8159@mail.gmail.com>
Message-ID: <4687E06C.7040108@ukr.net>


Matthieu Brucher wrote:
> numpy.zeros((2, 3, 4), float) for instance.
>
> Matthieu
Ok, please try the updated file.
Maybe, it could be simplified.
D.

from numpy import ones
def Ones(*args, **kwargs):
    if type(args[0]) in (type(()), type([])):
        return ones(*args, **kwargs)
    else:
        i, args2 = 1, [args[0]]
        while i<len(args) and type(args[i])==type(15):  
args2.append(args[i]);  i+=1
        if len(args[i:]) ==0: return ones(args2, **kwargs)
        elif len(args[i:]) ==1: return ones(args2, args[i], **kwargs)
        elif len(args[i:]) ==2: return ones(args2, args[i], args[i+1], 
**kwargs)
        else: print 'ERROR!'

if __name__ == '__main__':
    print Ones((2,2))
    print 2*Ones([2,2])
    print 3*Ones(2,2)   
    print 4*Ones(2,2,2, dtype=float)
    print 5*Ones((2,2,2), dtype='float')
    print 6*Ones([2,2,2], dtype='int')
    print 7*Ones(3,3, dtype=int, order = 'C')
    print 8*Ones((3,3), dtype='int')
    print 9*Ones(3, dtype='int')
    print 10*Ones((3,), dtype='int')

    print 11*ones((2, 3, 4), float)
    print 12*Ones([2, 3, 4], 'float')
    print 13*Ones(2, 3, 4, float)
    print 14*Ones(2, 3, 4, int, order = 'C')
    print 15*Ones(2, 3, 4, float, 'F')





From tritemio at gmail.com  Sun Jul  1 14:18:02 2007
From: tritemio at gmail.com (Antonino Ingargiola)
Date: Sun, 1 Jul 2007 20:18:02 +0200
Subject: [SciPy-user] sparse stuff in numpy/scipy
In-Reply-To: <4687B029.9070609@ukr.net>
References: <4687B029.9070609@ukr.net>
Message-ID: <5486cca80707011118n4beff38ewff50cf0d8cb42730@mail.gmail.com>

2007/7/1, dmitrey <openopt at ukr.net>:
> Hi all,
> I had heard there are some sparse matrix abilities in scipy (btw don't
> you think it's better to store those one in numpy?).
> However, neither dir(scipy), nor dir(numpy), nor
> http://www.scipy.org/Topical_Software doesn't tell me what should I use.

Have you tried "import scipy.sparse"?

Maybe one have to know something before beginning to make requests ;)


Best Regards,

    ~ Antonio


From wbaxter at gmail.com  Sun Jul  1 15:01:54 2007
From: wbaxter at gmail.com (Bill Baxter)
Date: Mon, 2 Jul 2007 04:01:54 +0900
Subject: [SciPy-user] once again about zeros() and ones()
In-Reply-To: <4687E06C.7040108@ukr.net>
References: <46878D13.10505@ukr.net>
	<e76aa17f0707010545y18744281k922badf24566d82f@mail.gmail.com>
	<4687A7D3.6030605@ukr.net>
	<e76aa17f0707010813r5e286916h3feaffeccbbd8159@mail.gmail.com>
	<4687E06C.7040108@ukr.net>
Message-ID: <e86a5fd00707011201g6a5d4b91je91819a50cc8e06a@mail.gmail.com>

The main argument against allowing zeros(1,2,3) and zeros([1,2,3]) is that
it complicates the implementation too much, and goes against the grain of
the language.
I argued to allow this sort of thing a while back also, and also posted my
own versions of your wrapper function, but now I agree that it isn't really
worth it just to avoid typing a few extra parentheses.

Problem 1:  complexity
The code does become more complex, and though the added code is not rocket
science, it's still not really trivial to verify by a quick glance if the
code is correct or not (as you experienced first hand).  It also adds to the
testing burden.  Every allowable call format really needs to have a unit
test and this increases the number of call formats.

Problem 2: transparency
Argument lists for functions that do this kind of thing all become a very
uninformative (*args,**kwds).  You can document what the possible call
formats are, but it's still not as convenient as being able to tell what the
function expects right from the argument list.  Moreover it becomes the
burden of the docstring to make sure this information is accurate, and it's
too easy for docstrings to get out of date with the code.  It's harder for
the parameter names themselves to become inaccurate.

Problem 3: consistency
To maintain consistency across the numpy API you really need to do this for
ALL functions that take a shape tuple.  Just adding it to ones() and zeros()
and a few other "common" commands will have users constantly asking
themselves "is this a common command? do I need parentheses here?"  The way
it is, all you have to remember is that if it's a shape, it'll be a tuple
argument.

Of course you can add your wrapper function to your own customization module
and import that instead of numpy.  I added mine to my local setup, but in
the end pulling in an extra module dependency just isn't really worth the
added benefit.

I do think it is handy for interactive use, though.  So I do load that
module in my PYTHONSTARTUP.

--bb

On 7/2/07, dmitrey <openopt at ukr.net> wrote:
>
>
> Matthieu Brucher wrote:
> > numpy.zeros((2, 3, 4), float) for instance.
> >
> > Matthieu
> Ok, please try the updated file.
> Maybe, it could be simplified.
> D.
>
> from numpy import ones
> def Ones(*args, **kwargs):
>     if type(args[0]) in (type(()), type([])):
>         return ones(*args, **kwargs)
>     else:
>         i, args2 = 1, [args[0]]
>         while i<len(args) and type(args[i])==type(15):
> args2.append(args[i]);  i+=1
>         if len(args[i:]) ==0: return ones(args2, **kwargs)
>         elif len(args[i:]) ==1: return ones(args2, args[i], **kwargs)
>         elif len(args[i:]) ==2: return ones(args2, args[i], args[i+1],
> **kwargs)
>         else: print 'ERROR!'
>
> if __name__ == '__main__':
>     print Ones((2,2))
>     print 2*Ones([2,2])
>     print 3*Ones(2,2)
>     print 4*Ones(2,2,2, dtype=float)
>     print 5*Ones((2,2,2), dtype='float')
>     print 6*Ones([2,2,2], dtype='int')
>     print 7*Ones(3,3, dtype=int, order = 'C')
>     print 8*Ones((3,3), dtype='int')
>     print 9*Ones(3, dtype='int')
>     print 10*Ones((3,), dtype='int')
>
>     print 11*ones((2, 3, 4), float)
>     print 12*Ones([2, 3, 4], 'float')
>     print 13*Ones(2, 3, 4, float)
>     print 14*Ones(2, 3, 4, int, order = 'C')
>     print 15*Ones(2, 3, 4, float, 'F')



FWIW, here's my current muti-arg ones().  Not sure if it's any better/worse
than yours, but it's different at least.


def ones(shape, *varg, **kwarg):
    """ones(shape, dtype=<type 'int32scalar'>, order='C')

    ones(shape, dtype=int_) returns an array of the given
    dimensions which is initialized to all ones.
    """

    if not kwarg.has_key('dtype'): kwarg['dtype']=numpy.float64
    if hasattr(shape,'__getitem__'):
        return numpy.ones(shape,*varg,**kwarg)

    i = 0
    for x in varg:
        if type(x)==types.IntType:
            i+=1
        else:
            break

    tshape = (shape,)
    if i>0:
        tshape += tuple( varg[0:i] )
    args = varg[i:]
    return numpy.ones(tshape, *args, **kwarg)

--bb
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070702/32622e0b/attachment.html>

From openopt at ukr.net  Sun Jul  1 15:28:16 2007
From: openopt at ukr.net (dmitrey)
Date: Sun, 01 Jul 2007 22:28:16 +0300
Subject: [SciPy-user] sparse stuff in numpy/scipy
In-Reply-To: <5486cca80707011118n4beff38ewff50cf0d8cb42730@mail.gmail.com>
References: <4687B029.9070609@ukr.net>
	<5486cca80707011118n4beff38ewff50cf0d8cb42730@mail.gmail.com>
Message-ID: <46880050.9080905@ukr.net>

Ok, thx, I hadn't noticed that one in dir(scipy)
But I still think it's better include to numpy.
D.

Antonino Ingargiola wrote:
> 2007/7/1, dmitrey <openopt at ukr.net>:
>   
>> Hi all,
>> I had heard there are some sparse matrix abilities in scipy (btw don't
>> you think it's better to store those one in numpy?).
>> However, neither dir(scipy), nor dir(numpy), nor
>> http://www.scipy.org/Topical_Software doesn't tell me what should I use.
>>     
>
> Have you tried "import scipy.sparse"?
>
> Maybe one have to know something before beginning to make requests ;)
>
>
> Best Regards,
>
>     ~ Antonio
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>
>   



From matthieu.brucher at gmail.com  Sun Jul  1 15:40:07 2007
From: matthieu.brucher at gmail.com (Matthieu Brucher)
Date: Sun, 1 Jul 2007 21:40:07 +0200
Subject: [SciPy-user] sparse stuff in numpy/scipy
In-Reply-To: <46880050.9080905@ukr.net>
References: <4687B029.9070609@ukr.net>
	<5486cca80707011118n4beff38ewff50cf0d8cb42730@mail.gmail.com>
	<46880050.9080905@ukr.net>
Message-ID: <e76aa17f0707011240r7ecffe49mee9af76b0b5d285f@mail.gmail.com>

Hi,

You'd better use pysparse, available on http://pysparse.sf.net
It's the up-to-date package :)

Matthieu

2007/7/1, dmitrey <openopt at ukr.net>:
>
> Ok, thx, I hadn't noticed that one in dir(scipy)
> But I still think it's better include to numpy.
> D.
>
> Antonino Ingargiola wrote:
> > 2007/7/1, dmitrey <openopt at ukr.net>:
> >
> >> Hi all,
> >> I had heard there are some sparse matrix abilities in scipy (btw don't
> >> you think it's better to store those one in numpy?).
> >> However, neither dir(scipy), nor dir(numpy), nor
> >> http://www.scipy.org/Topical_Software doesn't tell me what should I
> use.
> >>
> >
> > Have you tried "import scipy.sparse"?
> >
> > Maybe one have to know something before beginning to make requests ;)
> >
> >
> > Best Regards,
> >
> >     ~ Antonio
> > _______________________________________________
> > 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/20070701/b9386c0c/attachment.html>

From wbaxter at gmail.com  Sun Jul  1 15:46:29 2007
From: wbaxter at gmail.com (Bill Baxter)
Date: Mon, 2 Jul 2007 04:46:29 +0900
Subject: [SciPy-user] sparse stuff in numpy/scipy
In-Reply-To: <5486cca80707011118n4beff38ewff50cf0d8cb42730@mail.gmail.com>
References: <4687B029.9070609@ukr.net>
	<5486cca80707011118n4beff38ewff50cf0d8cb42730@mail.gmail.com>
Message-ID: <e86a5fd00707011246p1ce75ffbv7203fef82fb9eb28@mail.gmail.com>

On 7/2/07, Antonino Ingargiola <tritemio at gmail.com> wrote:
>
> 2007/7/1, dmitrey <openopt at ukr.net>:
> > Hi all,
> > I had heard there are some sparse matrix abilities in scipy (btw don't
> > you think it's better to store those one in numpy?).
> > However, neither dir(scipy), nor dir(numpy), nor
> > http://www.scipy.org/Topical_Software doesn't tell me what should I use.
>
> Have you tried "import scipy.sparse"?


Maybe one have to know something before beginning to make requests ;)


Dmitrey, you didn't miss it in the dir(). It's not there.  Some packages
like sparse and linalg that are considered 'extra' aren't loaded by default
to cut down on the load times for those that aren't using those packages.
Unfortunately these packages don't show up at all under the scipy namespace
until you explicitly import them.  It is mentioned in the docstring for
scipy itself.  I think that's the only way to know it's there.

Python gurus -- Is there some way that scipy can avoid hiding on-demand
packages like that?

It would be nice if there were some way to load a light-weight shims that
would show up in the dir() and maybe just have a docstring telling you you
need to import it explicity to actually load it.

No idea if python can do that sort of thing, but it would be nice.  Right
now google

As for scipy.sparse itself, it seems to be in need of some love.  I tried it
a while back and it seemed that many standard ndarray functions that could
be implemented simply weren't.  And you can't do any sort of fancy indexing,
which is often a key part of assembling the system matrix for sparse linear
systems (like those that arise from FEM or finite differences).

I think it would be great for Numpy to gain solid sparse support.  It makes
sense for that to be in the core of a library for numerical computing.  But
the devs seem to think of numpy more as a strided memory access library than
a numerical computing library.  And sparse matrices are not strided memory.
Hence not Numpy.

--bb
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070702/8ef4f14e/attachment.html>

From edschofield at gmail.com  Sun Jul  1 16:09:32 2007
From: edschofield at gmail.com (Ed Schofield)
Date: Sun, 1 Jul 2007 21:09:32 +0100
Subject: [SciPy-user] sparse stuff in numpy/scipy
In-Reply-To: <e76aa17f0707011240r7ecffe49mee9af76b0b5d285f@mail.gmail.com>
References: <4687B029.9070609@ukr.net>
	<5486cca80707011118n4beff38ewff50cf0d8cb42730@mail.gmail.com>
	<46880050.9080905@ukr.net>
	<e76aa17f0707011240r7ecffe49mee9af76b0b5d285f@mail.gmail.com>
Message-ID: <1b5a37350707011309k720c65dev8d559e7ed4aa21b6@mail.gmail.com>

On 7/1/07, Matthieu Brucher <matthieu.brucher at gmail.com> wrote:
> Hi,
>
> You'd better use pysparse, available on http://pysparse.sf.net
> It's the up-to-date package :)

Scipy.sparse should be fine for modest needs, and I think it's easier
to use than pysparse. It's scipy.sandbox.pysparse that's out of date
...

-- Ed


From edschofield at gmail.com  Sun Jul  1 16:23:34 2007
From: edschofield at gmail.com (Ed Schofield)
Date: Sun, 1 Jul 2007 21:23:34 +0100
Subject: [SciPy-user] sparse stuff in numpy/scipy
In-Reply-To: <e86a5fd00707011246p1ce75ffbv7203fef82fb9eb28@mail.gmail.com>
References: <4687B029.9070609@ukr.net>
	<5486cca80707011118n4beff38ewff50cf0d8cb42730@mail.gmail.com>
	<e86a5fd00707011246p1ce75ffbv7203fef82fb9eb28@mail.gmail.com>
Message-ID: <1b5a37350707011323s3b533fbqc02dd0525f688560@mail.gmail.com>

On 7/1/07, Bill Baxter <wbaxter at gmail.com> wrote:
>
> On 7/2/07, Antonino Ingargiola <tritemio at gmail.com> wrote:
> > 2007/7/1, dmitrey <openopt at ukr.net>:
> > > Hi all,
> > > I had heard there are some sparse matrix abilities in scipy (btw don't
> > > you think it's better to store those one in numpy?).
> > > However, neither dir(scipy), nor dir(numpy), nor
> > > http://www.scipy.org/Topical_Software doesn't tell me
> what should I use.
> >
> > Have you tried "import scipy.sparse"?
> ...
> As for scipy.sparse itself, it seems to be in need of some love.  I tried it
> a while back and it seemed that many standard ndarray functions that could
> be implemented simply weren't.  And you can't do any sort of fancy indexing,
> which is often a key part of assembling the system matrix for sparse linear
> systems (like those that arise from FEM or finite differences).

I added some fancy indexing support to lil_matrix and cs{rc}_matrix a
couple of years ago. The code is a nightmare. I couldn't think of a
better way to implement fancy indexing for sparse matrices, and still
can't. In some of the remaining cases my conclusion was that fancy
indexing couldn't be done efficiently at all, and I left these
unimplemented (with, hopefully, some useful error messages) rather
than implementing a slow, nasty algorithm.

But ideas and patches are welcome! ;)

-- Ed


From strawman at astraw.com  Sun Jul  1 16:59:57 2007
From: strawman at astraw.com (Andrew Straw)
Date: Sun, 01 Jul 2007 13:59:57 -0700
Subject: [SciPy-user] sparse stuff in numpy/scipy
In-Reply-To: <e86a5fd00707011246p1ce75ffbv7203fef82fb9eb28@mail.gmail.com>
References: <4687B029.9070609@ukr.net>	<5486cca80707011118n4beff38ewff50cf0d8cb42730@mail.gmail.com>
	<e86a5fd00707011246p1ce75ffbv7203fef82fb9eb28@mail.gmail.com>
Message-ID: <468815CD.6040606@astraw.com>

Bill Baxter wrote:

> Python gurus -- Is there some way that scipy can avoid hiding on-demand
> packages like that? 

My recollection is that Pearu(?) had implemented a module auto-loader
that essentially did what you're suggesting. It was ripped out in the
name of simplification, which IMO is a good thing -- it was non-trivial
code, and it had non-obvious effects. For example, I seem to remember
tab completion in IPython failing in the 1st attempt then working the
2nd attempt.

So, if scipy sub-packages aren't in the scipy docstring, I think that
should be fixed. But having a transparent auto-loader is probably
something best left behind.

(Many of the same arguments for and against ones taking shape without
parentheses could be made for this case, too.)

-Andrew


From barrywark at gmail.com  Sun Jul  1 20:50:40 2007
From: barrywark at gmail.com (Barry Wark)
Date: Sun, 1 Jul 2007 17:50:40 -0700
Subject: [SciPy-user] Compile fails for r3123 on OS X 10.4 (Intel)
In-Reply-To: <cd7634ce0706291033ga50ed5bj1e0d7537b09bb346@mail.gmail.com>
References: <cd7634ce0706291033ga50ed5bj1e0d7537b09bb346@mail.gmail.com>
Message-ID: <cd7634ce0707011750q208bb52dg710b1e8e34f35c73@mail.gmail.com>

I've attached the output to ticket #449 on the scipy trac, which
appears to have reported the same problem.

On 6/29/07, Barry Wark <barrywark at gmail.com> wrote:
> I've been unable to compile scipy on OS X 10.4 (intel) from the recent
> trunk. Scipy built on this machine as of r2708. The output from
>
> python setup.py build
>
> is attached.
>
> I have fftw3 installed via macports (in /opt/local), and it appears
> that the build finds it properly, but the build fails with an error:
>
> building extension "scipy.fftpack._fftpack" sources
>   target build/src.macosx-10.3-fat-2.5/_fftpackmodule.c does not exist:
>    Assuming _fftpackmodule.c was generated with "build_src --inplace" command.
> error: '_fftpackmodule.c' missing
>
> I would appreciate any advice or suggestions!
>
> Thanks,
> Barry
>
>


From robert.vergnes at yahoo.fr  Mon Jul  2 04:58:05 2007
From: robert.vergnes at yahoo.fr (Robert VERGNES)
Date: Mon, 2 Jul 2007 10:58:05 +0200 (CEST)
Subject: [SciPy-user] scipy import warning - how to remove
Message-ID: <20070702085805.99798.qmail@web27407.mail.ukl.yahoo.com>

Hello,

each time I do a 
from scipy import * 

I get a warning. How to remove it ?

This has a bad result on py2exe which seems to believe it is an error and stop starting my application.

Any ideas ?
(how to force not to importe scipytest ?)

Thanx

Robert
Platform - Windows - Scipy 0.5.2


from scipy import *

C:\Python25\lib\site-packages\scipy\misc\__init__.py:25: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\linalg\__init__.py:32: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\ndimage\__init__.py:40: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\sparse\__init__.py:9: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\io\__init__.py:20: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\lib\__init__.py:5: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\linsolve\umfpack\__init__.py:7: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\linsolve\__init__.py:13: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\interpolate\__init__.py:15: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\optimize\__init__.py:17: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\special\__init__.py:22: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\stats\__init__.py:15: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\fftpack\__init__.py:21: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\integrate\__init__.py:16: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\signal\__init__.py:17: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test
C:\Python25\lib\site-packages\scipy\maxentropy\__init__.py:12: DeprecationWarning: ScipyTest is now called NumpyTest; please update your code
  test = ScipyTest().test


       
---------------------------------
 Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070702/35fe4586/attachment.html>

From david at ar.media.kyoto-u.ac.jp  Mon Jul  2 04:50:04 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Mon, 02 Jul 2007 17:50:04 +0900
Subject: [SciPy-user] scipy import warning - how to remove
In-Reply-To: <20070702085805.99798.qmail@web27407.mail.ukl.yahoo.com>
References: <20070702085805.99798.qmail@web27407.mail.ukl.yahoo.com>
Message-ID: <4688BC3C.2080506@ar.media.kyoto-u.ac.jp>

Robert VERGNES wrote:
> Hello,
>
> each time I do a
> from scipy import *
>
> I get a warning. How to remove it ?
>
If you can afford it, I would say the easiest is to use svn scipy, where 
those warnings have been "solved"

David


From fredmfp at gmail.com  Mon Jul  2 07:36:20 2007
From: fredmfp at gmail.com (fred)
Date: Mon, 02 Jul 2007 13:36:20 +0200
Subject: [SciPy-user] ndimage.convolve and NaN...
Message-ID: <4688E334.1070303@gmail.com>

Hi,

I want to apply a convolution on a 2D array data, which has a few NaN.

Works quite fine, but "holes" (ie NaN) are widened by the convolution.

By the way, as edges are processed correctly, I wonder if it could 
possible to process the holes
like the edges, ie without widening them.

Any clue are welcome.

TIA.


Cheers,

-- 
http://scipy.org/FredericPetit



From anirudhvij at gmail.com  Mon Jul  2 08:47:53 2007
From: anirudhvij at gmail.com (anirudh vij)
Date: Mon, 2 Jul 2007 18:17:53 +0530
Subject: [SciPy-user] ndimage.convolve and NaN...
In-Reply-To: <4688E334.1070303@gmail.com>
References: <4688E334.1070303@gmail.com>
Message-ID: <1aa72c450707020547h13e634cegb1f5b551623ad88f@mail.gmail.com>

hi fred,
this is'nt a solution,but you might try replacing NaN by 0.That would ensure
that they have no effect on the surrounding pixels.
Edges work fine because of padding.
NaN's will always widen "holes".You might work around this,but there's no
solution.

On 7/2/07, fred <fredmfp at gmail.com> wrote:
>
> Hi,
>
> I want to apply a convolution on a 2D array data, which has a few NaN.
>
> Works quite fine, but "holes" (ie NaN) are widened by the convolution.
>
> By the way, as edges are processed correctly, I wonder if it could
> possible to process the holes
> like the edges, ie without widening them.
>
> Any clue are welcome.
>
> TIA.
>
>
> Cheers,
>
> --
> http://scipy.org/FredericPetit
>
> _______________________________________________
> 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/20070702/7e1af504/attachment.html>

From fredmfp at gmail.com  Mon Jul  2 09:20:54 2007
From: fredmfp at gmail.com (fred)
Date: Mon, 02 Jul 2007 15:20:54 +0200
Subject: [SciPy-user] ndimage.convolve and NaN...
In-Reply-To: <1aa72c450707020547h13e634cegb1f5b551623ad88f@mail.gmail.com>
References: <4688E334.1070303@gmail.com>
	<1aa72c450707020547h13e634cegb1f5b551623ad88f@mail.gmail.com>
Message-ID: <4688FBB6.5050900@gmail.com>

anirudh vij a ?crit :
> hi fred,
> this is'nt a solution,but you might try replacing NaN by 0.That would 
> ensure that they have no effect on the surrounding pixels.
Hmm, 0 are _real_ values.
> Edges work fine because of padding.
Yes, but can't we think replace NaN by padding ?

-- 
http://scipy.org/FredericPetit



From daniel.wheeler at nist.gov  Mon Jul  2 10:27:27 2007
From: daniel.wheeler at nist.gov (Daniel Wheeler)
Date: Mon, 2 Jul 2007 14:27:27 +0000
Subject: [SciPy-user] Deleting sandbox.pysparse
In-Reply-To: <46853223.1040607@gmail.com>
References: <1b5a37350706290911y1a1932b2had40fded7379c9a6@mail.gmail.com>
	<46853223.1040607@gmail.com>
Message-ID: <01315CA4-64FA-4337-84F8-6B2EB87A672A@nist.gov>


On Jun 29, 2007, at 4:24 PM, Dominique Orban wrote:

> Ed Schofield wrote:
>> Hi everyone,
>>
>> The pysparse snapshot in the sandbox is old and unmaintained, and I'd
>> like to remove it from the SVN tree. Is anyone using it?
>
> There were some messages on this list today from people using  
> Pysparse. Perhaps
> the maintainer of Pysparse (Daniel Wheeler I believe) should be  
> contacted to see
> if he would agree to maintain Pysparse as part of SciPy. I think it  
> would be a
> great feature.

As far as my project is concerned, I would prefer that pysparse  
remains at sourceforge for the time being.
Our python code (fipy) relies heavily on pysparse and I believe that  
it is easier for users to install
pysparse as a small package rather than making scipy a requirement.

Having said that, I guess it wouldn't be too difficult to move  
pysparse into the scipy repository and still
build separate snapshots of pysparse and upload then to sourceforge  
for users who only want pysparse.

However, I believe that having two incompatible sparse matrix and  
linalg modules within scipy just causes confusion.
Ripping pysparse code may be useful to help improve scipy's solver  
and linalg modules, but merely maintaining it
in the sandbox separate from the scipy solvers (as it has been) just  
adds to the confusion.

Cheers

--
Daniel Wheeler




From aisaac at american.edu  Mon Jul  2 11:14:04 2007
From: aisaac at american.edu (Alan G Isaac)
Date: Mon, 2 Jul 2007 11:14:04 -0400
Subject: [SciPy-user] Deleting sandbox.pysparse
In-Reply-To: <01315CA4-64FA-4337-84F8-6B2EB87A672A@nist.gov>
References: <1b5a37350706290911y1a1932b2had40fded7379c9a6@mail.gmail.com><46853223.1040607@gmail.com><01315CA4-64FA-4337-84F8-6B2EB87A672A@nist.gov>
Message-ID: <Mahogany-0.67.0-1656-20070702-111404.00@american.edu>

On Mon, 2 Jul 2007, Daniel Wheeler apparently wrote: 
> As far as my project is concerned, I would prefer that 
> pysparse remains at sourceforge for the time being.  Our 
> python code (fipy) relies heavily on pysparse and 
> I believe that it is easier for users to install pysparse 
> as a small package rather than making scipy a requirement. 

> Having said that, I guess it wouldn't be too difficult to 
> move pysparse into the scipy repository and still build 
> separate snapshots of pysparse and upload then to 
> sourceforge for users who only want pysparse. 


If it were a SciKit, users could simply retrieve the package 
from SVN and install it as usual, without I think adding any 
SciPy dependency.  (I am not trying to say this is the right 
way to go.)

Cheers,
Alan Isaac





From eike.welk at gmx.net  Mon Jul  2 12:22:12 2007
From: eike.welk at gmx.net (Eike Welk)
Date: Mon, 02 Jul 2007 18:22:12 +0200
Subject: [SciPy-user] ndimage.convolve and NaN...
In-Reply-To: <4688FBB6.5050900@gmail.com>
References: <4688E334.1070303@gmail.com>
	<1aa72c450707020547h13e634cegb1f5b551623ad88f@mail.gmail.com>
	<4688FBB6.5050900@gmail.com>
Message-ID: <200707021822.13019.eike.welk@gmx.net>

On Monday 02 July 2007 15:20, fred wrote:
> anirudh vij a ?crit :
> > hi fred,
> > this is'nt a solution,but you might try replacing NaN by 0.That
> > would ensure that they have no effect on the surrounding pixels.
>
> Hmm, 0 are _real_ values.
>
> > Edges work fine because of padding.
>
> Yes, but can't we think replace NaN by padding ?

Padding with zero is one of the possible padding modes. You could just 
choose this and say that the results are consistent this way. From a 
quick look at:
http://www.scipy.org/SciPyPackages/Ndimage
mode="constant", cval=0   
should do it.

Maybe you could use morphological operators to get an effect 
comparable to the nearest mode:
1. replace nan with 0
2. use dilation or closing to copy the neighboring values into the 
holes.
3. use the results of the morphological operators only where the holes 
were.

Regards,
Eike.


From robert.kern at gmail.com  Mon Jul  2 12:23:45 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Mon, 02 Jul 2007 11:23:45 -0500
Subject: [SciPy-user] ndimage.convolve and NaN...
In-Reply-To: <4688FBB6.5050900@gmail.com>
References: <4688E334.1070303@gmail.com>	<1aa72c450707020547h13e634cegb1f5b551623ad88f@mail.gmail.com>
	<4688FBB6.5050900@gmail.com>
Message-ID: <46892691.7080304@gmail.com>

fred wrote:
> anirudh vij a ?crit :
>> hi fred,
>> this is'nt a solution,but you might try replacing NaN by 0.That would 
>> ensure that they have no effect on the surrounding pixels.
> Hmm, 0 are _real_ values.

Yes, and in the context of a convolution, that's fine.

>> Edges work fine because of padding.
> Yes, but can't we think replace NaN by padding ?

That's what replacing them with 0 does.

-- 
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 fredmfp at gmail.com  Mon Jul  2 13:13:47 2007
From: fredmfp at gmail.com (fred)
Date: Mon, 02 Jul 2007 19:13:47 +0200
Subject: [SciPy-user] calling scipy methods from fortran code...
Message-ID: <4689324B.2040008@gmail.com>

Hi,

I have browsed all my docs, but did not find anything relevant to my issue.

How can I call scipy methods from a fortran code ?
I want to call thereafter my fortran module in my python code.

The reason is that I have to call a lot of linalg.solve() inside loops, but:
- loops are "forbidden" in python code;
- I don't want to write a standalone fortran code to solve the matrices;
- I know a little about calling fortran module from python using f2py.
But if you have a better idea...

Any pointer, clue, docs are welcome.


TIA.


Cheers,

-- 
http://scipy.org/FredericPetit



From gnata at obs.univ-lyon1.fr  Mon Jul  2 13:39:58 2007
From: gnata at obs.univ-lyon1.fr (Gnata Xavier)
Date: Mon, 02 Jul 2007 19:39:58 +0200
Subject: [SciPy-user] calling scipy methods from fortran code...
In-Reply-To: <4689324B.2040008@gmail.com>
References: <4689324B.2040008@gmail.com>
Message-ID: <4689386E.1070308@obs.univ-lyon1.fr>

Hi,
"- loops are "forbidden" in python code; "
Well ok loop are slow in python but if each llinalg.solve() takes a 
significant amount of time, you may just want to use a python loop.
Anyway, calling scipy methods from fortran looks funny because scipy is 
based on fortran routines. Maybe you should have a direct look at lapack.

Xavier
> Hi,
>
> I have browsed all my docs, but did not find anything relevant to my issue.
>
> How can I call scipy methods from a fortran code ?
> I want to call thereafter my fortran module in my python code.
>
> The reason is that I have to call a lot of linalg.solve() inside loops, but:
> - loops are "forbidden" in python code;
> - I don't want to write a standalone fortran code to solve the matrices;
> - I know a little about calling fortran module from python using f2py.
> But if you have a better idea...
>
> Any pointer, clue, docs are welcome.
>
>
> TIA.
>
>
> Cheers,
>
>   



From fredmfp at gmail.com  Mon Jul  2 13:59:09 2007
From: fredmfp at gmail.com (fred)
Date: Mon, 02 Jul 2007 19:59:09 +0200
Subject: [SciPy-user] calling scipy methods from fortran code...
In-Reply-To: <4689386E.1070308@obs.univ-lyon1.fr>
References: <4689324B.2040008@gmail.com> <4689386E.1070308@obs.univ-lyon1.fr>
Message-ID: <46893CED.5050302@gmail.com>

Gnata Xavier a ?crit :
> Hi,
> "- loops are "forbidden" in python code; "
> Well ok loop are slow in python but if each llinalg.solve() takes a 
> significant amount of time, you may just want to use a python loop.
> Anyway, calling scipy methods from fortran looks funny because scipy is 
> based on fortran routines. Maybe you should have a direct look at lapack.
>   
Ok, you're right.
But I was looking for something easy & ready-to-use, so I thought about 
scipy calls.

Anyway, I prefer to use intel MKL and its symmetric undefinite matrix 
solver,
but this is another story...

Cheers,

-- 
http://scipy.org/FredericPetit



From peridot.faceted at gmail.com  Mon Jul  2 14:33:15 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Mon, 2 Jul 2007 14:33:15 -0400
Subject: [SciPy-user] ndimage.convolve and NaN...
In-Reply-To: <4688E334.1070303@gmail.com>
References: <4688E334.1070303@gmail.com>
Message-ID: <ce557a360707021133p3fab8ca3s4dc0b251c2147744@mail.gmail.com>

On 02/07/07, fred <fredmfp at gmail.com> wrote:

> I want to apply a convolution on a 2D array data, which has a few NaN.
>
> Works quite fine, but "holes" (ie NaN) are widened by the convolution.

This is actually unavoidable. Think of it this way: NaN means "I have
no idea what value goes here". If you replaced them by 10^20, the
convolution would spread this value over the nearby pixels, so if you
don't know what value goes in the NaN, you don't know what value goes
in the nearby pixels either.

That said, some implementations of convolution widens the NaNs more
than is mathematically necessary - if you pad the convolution kernel
with zeros, that shouldn't change the result but will (because 0*NaN
is NaN). The extreme example of this is when you do the convolution
with a single FFT, where your whole array turns into NaNs.

> By the way, as edges are processed correctly, I wonder if it could
> possible to process the holes
> like the edges, ie without widening them.

There are technical solutions to this problem, but do think about
definitions: mathematically, what do you want to happen? For
definiteness, let's suppose your kernel is a Gaussian blur, and you've
got a sizable hole full of NaNs. You can treat them as zeros, which
will avoid their values being added to the neighbours, but their
neighbours will become darker: the blur will distribute their value
over nearby pixels, but the hole will not distribute any value into
them. If you want the pixels that are unaffected by this darkening,
well, that's pretty much exactly the ones that are not turned into
NaNs by the current procedure.

Anne


From peridot.faceted at gmail.com  Mon Jul  2 14:43:11 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Mon, 2 Jul 2007 14:43:11 -0400
Subject: [SciPy-user] sparse stuff in numpy/scipy
In-Reply-To: <46880050.9080905@ukr.net>
References: <4687B029.9070609@ukr.net>
	<5486cca80707011118n4beff38ewff50cf0d8cb42730@mail.gmail.com>
	<46880050.9080905@ukr.net>
Message-ID: <ce557a360707021143g207ea899l1e5f2f937a92e695@mail.gmail.com>

On 01/07/07, dmitrey <openopt at ukr.net> wrote:
> Ok, thx, I hadn't noticed that one in dir(scipy)
> But I still think it's better include to numpy.

Use help(scipy). More generally, for any scipy subpackage, use
help(scipy.subpackage) to see what it has and what it does. If
anything is missing from these docstrings, file a bug.

The fact of dir() not listing all subpackages is more or less a python
fact of life, not just a scipy thing.

Anne


From fie.pye at atlas.cz  Mon Jul  2 18:01:56 2007
From: fie.pye at atlas.cz (fie.pye at atlas.cz)
Date: Mon, 02 Jul 2007 22:01:56 GMT
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
Message-ID: <8767d3c1ef714e7abe38efc2a99fd879@664f5b6eea4d473ba43a5e9884940967>

An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070702/07cb752a/attachment.html>

From dominique.orban at gmail.com  Mon Jul  2 18:01:44 2007
From: dominique.orban at gmail.com (Dominique Orban)
Date: Mon, 02 Jul 2007 18:01:44 -0400
Subject: [SciPy-user] calling scipy methods from fortran code...
In-Reply-To: <46893CED.5050302@gmail.com>
References: <4689324B.2040008@gmail.com> <4689386E.1070308@obs.univ-lyon1.fr>
	<46893CED.5050302@gmail.com>
Message-ID: <468975C8.3080306@gmail.com>

fred wrote:
> Gnata Xavier a ?crit :
>> Hi,
>> "- loops are "forbidden" in python code; "
>> Well ok loop are slow in python but if each llinalg.solve() takes a 
>> significant amount of time, you may just want to use a python loop.
>> Anyway, calling scipy methods from fortran looks funny because scipy is 
>> based on fortran routines. Maybe you should have a direct look at lapack.
>>   
> Ok, you're right.
> But I was looking for something easy & ready-to-use, so I thought about 
> scipy calls.
> 
> Anyway, I prefer to use intel MKL and its symmetric undefinite matrix 
> solver,
> but this is another story...

Fred,

The de-facto standard for solving symmetric indefinite linear systems is a 
package called MA27 from the Harwell Subroutine Library. This one is available 
for free for non-commercial and research projects 
(http://hsl.rl.ac.uk/archive/hslarchive.html).

I interfaced MA27 to Python in my NLPy package (http://nlpy.sf.net) and wrote a 
nice wraparound Python class. There is an example on the NLPy website (in the 
'Examples' section). The interface is with the PySparse sparse matrix package. 
You won't have to call Fortran from Python; I already wrote the interface.

Typically, for my optimization applications, I call MA27 many times inside a 
outer loop. Make sure to use NumPy for all array operations and things should go 
well. I'm interested in any comments you might have on the package.

Dominique


From fie.pye at atlas.cz  Mon Jul  2 18:37:00 2007
From: fie.pye at atlas.cz (fie.pye at atlas.cz)
Date: Mon, 02 Jul 2007 22:37:00 GMT
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
Message-ID: <cdf887ffa55c4a31a17d9a5a9e2eb47b@ff0adc736c274db493250a3e564ae499>

Hello.

I would like to install NumPy and SciPy modules. I compiled and install BLAS, LAPACK and ATLAS libraries. After installation and importing NumPy I obtained message:

>>> import numpy
Traceback (most recent call last):
File "", line 1, in 
File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/__init__.py", line 43, in 
import linalg
File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in 
from linalg import *
File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in 
from numpy.linalg import lapack_lite
ImportError: liblapack.so: cannot open shared object file: No such file or directory
It seams to me that BLAS and LAPACK installation is OK.

PC: 2x Dual-core Opteron 275
OS: CentOS 5.0, kernel 2.6.18-8.1.6.el5
compiler: gcc version 4.2.0, gfortran
python2.5.1
numpy-1.0.3-2
scipy-0.5.2


#BLAS installation Make.inc file:
SHELL = /bin/sh
PLAT = _LINUX
FORTRAN = gfortran
OPTS = -O3 -fPIC
DRVOPTS = $(OPTS)
NOOPT = -fPIC
LOADER = gfortran
LOADOPTS =
ARCH = ar
ARCHFLAGS= cr
RANLIB = ranlib
BLASLIB = blas$(PLAT).a

# Shared libraries
gfortran -shared -Wl,-soname,libblas.so -o libblas.so *.o


# LAPACK installation Make.inc
PLAT = _LINUX
FORTRAN = gfortran
OPTS = -funroll-all-loops -O3 -fpic
DRVOPTS = $(OPTS)
NOOPT =
LOADER = gfortran
LOADOPTS =
TIMER = INT_CPU_TIME
ARCH = ar
ARCHFLAGS= cr
RANLIB = ranlib
BLASLIB = /usr/local/lib64/blas/libblas.a
LAPACKLIB = lapack$(PLAT).a
TMGLIB = tmglib$(PLAT).a
EIGSRCLIB = eigsrc$(PLAT).a
LINSRCLIB = linsrc$(PLAT).a

# Shared libraries
gfortran -shared -Wl,-soname,liblapack.so -o liblapack.so *.o


# NumPy installation
# After ATLAS installation I put all libraries in
# /usr/local/lib64/atlas/
BLAS=/usr/local/lib64/atlas/libfblas.a
LAPACK=/usr/local/lib64/atlas/libflapack.a
ATLAS=/usr/local/lib64/atlas/libatlas.a
LD_LIBRARY_PATH=/usr/lib/:/usr/local/lib/:/usr/lib64/:/usr/local/lib64/:/usr/local/lib64/atlas/
export BLAS LAPACK ATLAS LD_LIBRARY_PATH
python2.5.1 setup.py build
python2.5.1 setup.py install


Thank you for help.
Fie Pye 

From fredmfp at gmail.com  Mon Jul  2 19:02:10 2007
From: fredmfp at gmail.com (fred)
Date: Tue, 03 Jul 2007 01:02:10 +0200
Subject: [SciPy-user] calling scipy methods from fortran code...
In-Reply-To: <468975C8.3080306@gmail.com>
References: <4689324B.2040008@gmail.com>
	<4689386E.1070308@obs.univ-lyon1.fr>	<46893CED.5050302@gmail.com>
	<468975C8.3080306@gmail.com>
Message-ID: <468983F2.6020502@gmail.com>

Dominique Orban a ?crit :
> Fred,
>
> The de-facto standard for solving symmetric indefinite linear systems is a 
> package called MA27 from the Harwell Subroutine Library. This one is available 
> for free for non-commercial and research projects 
> (http://hsl.rl.ac.uk/archive/hslarchive.html).
>
> I interfaced MA27 to Python in my NLPy package (http://nlpy.sf.net) and wrote a 
> nice wraparound Python class. There is an example on the NLPy website (in the 
> 'Examples' section). The interface is with the PySparse sparse matrix package. 
> You won't have to call Fortran from Python; I already wrote the interface.
>
> Typically, for my optimization applications, I call MA27 many times inside a 
> outer loop. Make sure to use NumPy for all array operations and things should go 
> well. I'm interested in any comments you might have on the package.
>   
Hey, sounds great ! :-)))

I will get it a try asap... tomorrow (or better today :-).


Thanks.

-- 
http://scipy.org/FredericPetit



From david at ar.media.kyoto-u.ac.jp  Mon Jul  2 22:10:12 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Tue, 03 Jul 2007 11:10:12 +0900
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
In-Reply-To: <cdf887ffa55c4a31a17d9a5a9e2eb47b@ff0adc736c274db493250a3e564ae499>
References: <cdf887ffa55c4a31a17d9a5a9e2eb47b@ff0adc736c274db493250a3e564ae499>
Message-ID: <4689B004.3060900@ar.media.kyoto-u.ac.jp>

fie.pye at atlas.cz wrote:
> Hello.
>
> I would like to install NumPy and SciPy modules. I compiled and install BLAS, LAPACK and ATLAS libraries. After installation and importing NumPy I obtained message:
>
>>>> import numpy
> Traceback (most recent call last):
> File "", line 1, in 
> File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/__init__.py", line 43, in 
> import linalg
> File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in 
> from linalg import *
> File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in 
> from numpy.linalg import lapack_lite
> ImportError: liblapack.so: cannot open shared object file: No such file or directory
> It seams to me that BLAS and LAPACK installation is OK.
Well, it may not be :) If you do (once you exported LD_LIBRARY_PATH):

ldd 
/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so

What do you get ? The most useful information, actually, is to give us 
the log of python setup.py config &> log when building numpy and scipy.
>
> # LAPACK installation Make.inc
> PLAT = _LINUX
> FORTRAN = gfortran
> OPTS = -funroll-all-loops -O3 -fpic
> DRVOPTS = $(OPTS)
> NOOPT =
Here you forgot NOOPT = -fpic. Then, how did you compile atlas ? You 
should try first without atlas at all, I think: compile blas, lapack, 
and check that python setup.py config picks up the libraries.

As a note, there are source rpms available for blas/lapack/atlas/scipy: 
the binary rpms do not support CENTOS, but I think building the rpm on 
CENTOS should not be too difficult. Everything is taken care of in the 
rpms: options, compatible fortran ABI, full pic version of liblapack to 
build complete LAPACK based on ATLAS, etc...

http://software.opensuse.org/download/home:/ashigabou/ (pick up source 
rpm for FC 6 or 7, this should be the more similar to centos).

If this does not work, I would like to hear about it,

David


From fredmfp at gmail.com  Tue Jul  3 04:35:45 2007
From: fredmfp at gmail.com (fred)
Date: Tue, 03 Jul 2007 10:35:45 +0200
Subject: [SciPy-user] calling scipy methods from fortran code...
In-Reply-To: <468975C8.3080306@gmail.com>
References: <4689324B.2040008@gmail.com>
	<4689386E.1070308@obs.univ-lyon1.fr>	<46893CED.5050302@gmail.com>
	<468975C8.3080306@gmail.com>
Message-ID: <468A0A61.6020303@gmail.com>

Dominique Orban a ?crit :
> Fred,
>
> The de-facto standard for solving symmetric indefinite linear systems is a 
> package called MA27 from the Harwell Subroutine Library. This one is available 
> for free for non-commercial and research projects 
> (http://hsl.rl.ac.uk/archive/hslarchive.html).
>
>   
Dominique,

I read on the nlpy page that MA27 is for sparse matrices.
My matrices are not sparse, but dense.
Should I still use it ?


Cheers,

-- 
http://scipy.org/FredericPetit



From nwagner at iam.uni-stuttgart.de  Tue Jul  3 05:20:33 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Tue, 03 Jul 2007 11:20:33 +0200
Subject: [SciPy-user] calling scipy methods from fortran code...
In-Reply-To: <468A0A61.6020303@gmail.com>
References: <4689324B.2040008@gmail.com>	<4689386E.1070308@obs.univ-lyon1.fr>	<46893CED.5050302@gmail.com>	<468975C8.3080306@gmail.com>
	<468A0A61.6020303@gmail.com>
Message-ID: <468A14E1.4090509@iam.uni-stuttgart.de>

fred wrote:
> Dominique Orban a ?crit :
>   
>> Fred,
>>
>> The de-facto standard for solving symmetric indefinite linear systems is a 
>> package called MA27 from the Harwell Subroutine Library. This one is available 
>> for free for non-commercial and research projects 
>> (http://hsl.rl.ac.uk/archive/hslarchive.html).
>>
>>   
>>     
> Dominique,
>
> I read on the nlpy page that MA27 is for sparse matrices.
> My matrices are not sparse, but dense.
> Should I still use it ?
>
>
> Cheers,
>
>   
In that case I suggest CVXOPT which comes with a LAPACK interface (LDL^T
factorization is included as opposed to scipy).
See
>>> help (cvxopt.lapack.sytrf)
>>> help (cvxopt.lapack.sytrs)

Nils
 


From fredmfp at gmail.com  Tue Jul  3 05:35:39 2007
From: fredmfp at gmail.com (fred)
Date: Tue, 03 Jul 2007 11:35:39 +0200
Subject: [SciPy-user] calling scipy methods from fortran code...
In-Reply-To: <468A14E1.4090509@iam.uni-stuttgart.de>
References: <4689324B.2040008@gmail.com>	<4689386E.1070308@obs.univ-lyon1.fr>	<46893CED.5050302@gmail.com>	<468975C8.3080306@gmail.com>	<468A0A61.6020303@gmail.com>
	<468A14E1.4090509@iam.uni-stuttgart.de>
Message-ID: <468A186B.3080207@gmail.com>

Nils Wagner a ?crit :
> fred wrote:
>   
>> Dominique Orban a ?crit :
>>   
>>     
>>> Fred,
>>>
>>> The de-facto standard for solving symmetric indefinite linear systems is a 
>>> package called MA27 from the Harwell Subroutine Library. This one is available 
>>> for free for non-commercial and research projects 
>>> (http://hsl.rl.ac.uk/archive/hslarchive.html).
>>>
>>>   
>>>     
>>>       
>> Dominique,
>>
>> I read on the nlpy page that MA27 is for sparse matrices.
>> My matrices are not sparse, but dense.
>> Should I still use it ?
>>
>>
>> Cheers,
>>
>>   
>>     
> In that case I suggest CVXOPT which comes with a LAPACK interface (LDL^T
> factorization is included as opposed to scipy).
> See
>   
>>>> help (cvxopt.lapack.sytrf)
>>>> help (cvxopt.lapack.sytrs)
>>>>         
Thanks.

-- 
http://scipy.org/FredericPetit



From tritemio at gmail.com  Tue Jul  3 06:02:42 2007
From: tritemio at gmail.com (Antonino Ingargiola)
Date: Tue, 3 Jul 2007 12:02:42 +0200
Subject: [SciPy-user] sparse stuff in numpy/scipy
In-Reply-To: <ce557a360707021143g207ea899l1e5f2f937a92e695@mail.gmail.com>
References: <4687B029.9070609@ukr.net>
	<5486cca80707011118n4beff38ewff50cf0d8cb42730@mail.gmail.com>
	<46880050.9080905@ukr.net>
	<ce557a360707021143g207ea899l1e5f2f937a92e695@mail.gmail.com>
Message-ID: <5486cca80707030302s79e4cf04tc935b77a39dd1556@mail.gmail.com>

2007/7/2, Anne Archibald <peridot.faceted at gmail.com>:
> On 01/07/07, dmitrey <openopt at ukr.net> wrote:
> > Ok, thx, I hadn't noticed that one in dir(scipy)
> > But I still think it's better include to numpy.
>
> Use help(scipy). More generally, for any scipy subpackage, use
> help(scipy.subpackage) to see what it has and what it does. If
> anything is missing from these docstrings, file a bug.
>
> The fact of dir() not listing all subpackages is more or less a python
> fact of life, not just a scipy thing.

I agree. I usually find very handy to look at the documentation using
pydoc. It creates html pages from the docstring on-fly (when queried),
its really amazing.

Here you can find a quick explanation how to use it:

http://pyplotsuite.sourceforge.net/NumericalPythonHowto.html


Regards,

    ~ Antonio


From dominique.orban at gmail.com  Tue Jul  3 11:14:07 2007
From: dominique.orban at gmail.com (Dominique Orban)
Date: Tue, 03 Jul 2007 11:14:07 -0400
Subject: [SciPy-user] calling scipy methods from fortran code...
In-Reply-To: <468A0A61.6020303@gmail.com>
References: <4689324B.2040008@gmail.com>	<4689386E.1070308@obs.univ-lyon1.fr>	<46893CED.5050302@gmail.com>	<468975C8.3080306@gmail.com>
	<468A0A61.6020303@gmail.com>
Message-ID: <468A67BF.7090804@gmail.com>



fred wrote:
> Dominique Orban a ?crit :
>> Fred,
>>
>> The de-facto standard for solving symmetric indefinite linear systems is a 
>> package called MA27 from the Harwell Subroutine Library. This one is available 
>> for free for non-commercial and research projects 
>> (http://hsl.rl.ac.uk/archive/hslarchive.html).
>>
>>   
> Dominique,
> 
> I read on the nlpy page that MA27 is for sparse matrices.
> My matrices are not sparse, but dense.
> Should I still use it ?

If they are not too large, you certainly can use it. Otherwise, you can try 
Lapack routines that perform a Bunch-Parlett (or Bunch-Kaufmann) factorization. 
That factorization is often written L B L^T, by contrast with L D L^T which is 
the Cholesky factorization. The differences between the two are that

1) there is no sense trying to compute the L D L^T factorization of a symmetric 
indefinite matrix. This Cholesky factorization is only defined for symmetric 
positive definite matrices.

2) The B differs from the D in that it is block diagonal with blocks of order at 
most 2 along the diagonal.

Dominique


From fredmfp at gmail.com  Tue Jul  3 12:20:29 2007
From: fredmfp at gmail.com (fred)
Date: Tue, 03 Jul 2007 18:20:29 +0200
Subject: [SciPy-user] calling scipy methods from fortran code...
In-Reply-To: <468A67BF.7090804@gmail.com>
References: <4689324B.2040008@gmail.com>	<4689386E.1070308@obs.univ-lyon1.fr>	<46893CED.5050302@gmail.com>	<468975C8.3080306@gmail.com>	<468A0A61.6020303@gmail.com>
	<468A67BF.7090804@gmail.com>
Message-ID: <468A774D.9080403@gmail.com>

Dominique Orban a ?crit :
> fred wrote:
>   
>> Dominique Orban a ??crit :
>>     
>>> Fred,
>>>
>>> The de-facto standard for solving symmetric indefinite linear systems is a 
>>> package called MA27 from the Harwell Subroutine Library. This one is available 
>>> for free for non-commercial and research projects 
>>> (http://hsl.rl.ac.uk/archive/hslarchive.html).
>>>
>>>   
>>>       
>> Dominique,
>>
>> I read on the nlpy page that MA27 is for sparse matrices.
>> My matrices are not sparse, but dense.
>> Should I still use it ?
>>     
>
> If they are not too large, you certainly can use it. Otherwise, you can try 
> Lapack routines that perform a Bunch-Parlett (or Bunch-Kaufmann) factorization. 
> That factorization is often written L B L^T, by contrast with L D L^T which is 
> the Cholesky factorization. The differences between the two are that
>
> 1) there is no sense trying to compute the L D L^T factorization of a symmetric 
> indefinite matrix. This Cholesky factorization is only defined for symmetric 
> positive definite matrices.
>   
Yes, I know that.

I have initiated a thread about it a few months ago.
The Cholesky factorization is faster (~<2 times) than the LU, but does 
not work for
my symmetric indefinite matrices. So I kept  my code  with the LU 
version, until now.
The cvxopt is faster than LU (2  times) and work fine on my matrices. So 
I'm quite happy.
Thanks to Nils.

Anyway, I got an issue.
cvxopt.lapack.sytr? only work on float64 matrices.
Mine are generally float32.

So I got a try with

foo = asarray(self.KM, dtype='d')
KM = matrix(foo)

but it complains about my matrices are not contiguous.


What's wrong ?

-- 
http://scipy.org/FredericPetit



From barrywark at gmail.com  Tue Jul  3 15:54:28 2007
From: barrywark at gmail.com (Barry Wark)
Date: Tue, 3 Jul 2007 12:54:28 -0700
Subject: [SciPy-user] Compile fails for r3123 on OS X 10.4 (Intel)
In-Reply-To: <cd7634ce0707011750q208bb52dg710b1e8e34f35c73@mail.gmail.com>
References: <cd7634ce0706291033ga50ed5bj1e0d7537b09bb346@mail.gmail.com>
	<cd7634ce0707011750q208bb52dg710b1e8e34f35c73@mail.gmail.com>
Message-ID: <cd7634ce0707031254o5c1d9465u2ab551d9def916ca@mail.gmail.com>

I know this isn't anyone's top priority on the list, but I'd really
appreciate a pointer (or a dope slap):

I've successfully compiled previous revisions (<r3000) of scipy on
this system (OSX Intel) before. However, after checking out an older
version, the build fails with

File "Lib/odr/setup.py", line 9, in <module>
    from numpy.distutils.misc_util import get_path, Configuration, dot_join
ImportError: cannot import name get_path

I assume this is because numpy has been updated and removed get_path.

Besides ticket #449, I haven't been able to find mention of the
problem described earlier in the thread in the list archives or
online. Obviously, not many other people are having this problem, so I
assume it's a stupid mistake on my part. It's holding up deployment of
software to our lab, however, so I would be eternally grateful for any
help.

Thanks,
Barry

On 7/1/07, Barry Wark <barrywark at gmail.com> wrote:
> I've attached the output to ticket #449 on the scipy trac, which
> appears to have reported the same problem.
>
> On 6/29/07, Barry Wark <barrywark at gmail.com> wrote:
> > I've been unable to compile scipy on OS X 10.4 (intel) from the recent
> > trunk. Scipy built on this machine as of r2708. The output from
> >
> > python setup.py build
> >
> > is attached.
> >
> > I have fftw3 installed via macports (in /opt/local), and it appears
> > that the build finds it properly, but the build fails with an error:
> >
> > building extension "scipy.fftpack._fftpack" sources
> >   target build/src.macosx-10.3-fat-2.5/_fftpackmodule.c does not exist:
> >    Assuming _fftpackmodule.c was generated with "build_src --inplace" command.
> > error: '_fftpackmodule.c' missing
> >
> > I would appreciate any advice or suggestions!
> >
> > Thanks,
> > Barry
> >
> >
>


From robert.kern at gmail.com  Tue Jul  3 16:03:16 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 03 Jul 2007 15:03:16 -0500
Subject: [SciPy-user] Compile fails for r3123 on OS X 10.4 (Intel)
In-Reply-To: <cd7634ce0707031254o5c1d9465u2ab551d9def916ca@mail.gmail.com>
References: <cd7634ce0706291033ga50ed5bj1e0d7537b09bb346@mail.gmail.com>	<cd7634ce0707011750q208bb52dg710b1e8e34f35c73@mail.gmail.com>
	<cd7634ce0707031254o5c1d9465u2ab551d9def916ca@mail.gmail.com>
Message-ID: <468AAB84.4080004@gmail.com>

Barry Wark wrote:
> I know this isn't anyone's top priority on the list, but I'd really
> appreciate a pointer (or a dope slap):
> 
> I've successfully compiled previous revisions (<r3000) of scipy on
> this system (OSX Intel) before. However, after checking out an older
> version, the build fails with
> 
> File "Lib/odr/setup.py", line 9, in <module>
>     from numpy.distutils.misc_util import get_path, Configuration, dot_join
> ImportError: cannot import name get_path
> 
> I assume this is because numpy has been updated and removed get_path.
> 
> Besides ticket #449, I haven't been able to find mention of the
> problem described earlier in the thread in the list archives or
> online. Obviously, not many other people are having this problem, so I
> assume it's a stupid mistake on my part. It's holding up deployment of
> software to our lab, however, so I would be eternally grateful for any
> help.

No, it was a stupid mistake on my part. I needed to refactor that bit of code in
numpy.distutils, and I thought the function was only used internally. The
current SVN of scipy/odr/setup.py uses the correct idiom for building with
numpy.distutils. While I shouldn't have removed get_path(), it is a holdover
from the days of scipy_distutils.

-- 
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 cookedm at physics.mcmaster.ca  Tue Jul  3 16:04:23 2007
From: cookedm at physics.mcmaster.ca (David M. Cooke)
Date: Tue, 3 Jul 2007 16:04:23 -0400
Subject: [SciPy-user] Compile fails for r3123 on OS X 10.4 (Intel)
In-Reply-To: <cd7634ce0706291033ga50ed5bj1e0d7537b09bb346@mail.gmail.com>
References: <cd7634ce0706291033ga50ed5bj1e0d7537b09bb346@mail.gmail.com>
Message-ID: <20070703200423.GA5477@arbutus.physics.mcmaster.ca>

On Fri, Jun 29, 2007 at 10:33:24AM -0700, Barry Wark wrote:
> I've been unable to compile scipy on OS X 10.4 (intel) from the recent
> trunk. Scipy built on this machine as of r2708. The output from
> 
> python setup.py build
> 
> is attached.
> 
> I have fftw3 installed via macports (in /opt/local), and it appears
> that the build finds it properly, but the build fails with an error:
> 
> building extension "scipy.fftpack._fftpack" sources
>  target build/src.macosx-10.3-fat-2.5/_fftpackmodule.c does not exist:
>   Assuming _fftpackmodule.c was generated with "build_src --inplace" 
>   command.
> error: '_fftpackmodule.c' missing
> 
> I would appreciate any advice or suggestions!

I've got pretty much the same setup, and no problems. Did you delete the
build directory first (rm -rf build)?

-- 
|>|\/|<
/--------------------------------------------------------------------------\
|David M. Cooke                      http://arbutus.physics.mcmaster.ca/dmc/
|cookedm at physics.mcmaster.ca


From barrywark at gmail.com  Tue Jul  3 16:14:12 2007
From: barrywark at gmail.com (Barry Wark)
Date: Tue, 3 Jul 2007 13:14:12 -0700
Subject: [SciPy-user] Compile fails for r3123 on OS X 10.4 (Intel)
In-Reply-To: <20070703200423.GA5477@arbutus.physics.mcmaster.ca>
References: <cd7634ce0706291033ga50ed5bj1e0d7537b09bb346@mail.gmail.com>
	<20070703200423.GA5477@arbutus.physics.mcmaster.ca>
Message-ID: <cd7634ce0707031314u1f2827c3jabf4bde824eae543@mail.gmail.com>

Yes, I'm using a clean checkout from SVN. Just to double check against
your setup, I'm using fftw-3 installed via macports as my fftw library

Barry

On 7/3/07, David M. Cooke <cookedm at physics.mcmaster.ca> wrote:
> On Fri, Jun 29, 2007 at 10:33:24AM -0700, Barry Wark wrote:
> > I've been unable to compile scipy on OS X 10.4 (intel) from the recent
> > trunk. Scipy built on this machine as of r2708. The output from
> >
> > python setup.py build
> >
> > is attached.
> >
> > I have fftw3 installed via macports (in /opt/local), and it appears
> > that the build finds it properly, but the build fails with an error:
> >
> > building extension "scipy.fftpack._fftpack" sources
> >  target build/src.macosx-10.3-fat-2.5/_fftpackmodule.c does not exist:
> >   Assuming _fftpackmodule.c was generated with "build_src --inplace"
> >   command.
> > error: '_fftpackmodule.c' missing
> >
> > I would appreciate any advice or suggestions!
>
> I've got pretty much the same setup, and no problems. Did you delete the
> build directory first (rm -rf build)?
>
> --
> |>|\/|<
> /--------------------------------------------------------------------------\
> |David M. Cooke                      http://arbutus.physics.mcmaster.ca/dmc/
> |cookedm at physics.mcmaster.ca
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From lorrmann at physik.uni-wuerzburg.de  Tue Jul  3 18:18:23 2007
From: lorrmann at physik.uni-wuerzburg.de (Volker Lorrmann)
Date: Wed, 04 Jul 2007 00:18:23 +0200
Subject: [SciPy-user] optimization question
Message-ID: <468ACB2F.8030309@physik.uni-wuerzburg.de>

Hello list,

i wanna fit a function to some measured datapoints T_m(x_i). The 
function i wanna fit  is something like that,T_fit(a,b,c,d,x_i), where 
a,b,c are the fitting-parameters. Fitting this with 
scipy.optimize.leastsq would be easy (btw. is there anoterh way to fit 
this, like fmin, fmin_powell, ...?).

The problem is, that a is a and b are _fixed_ scalar parameters. But c 
and d are variables, that depend on x_i, c=c(x_i) and d=d(x_i). And in 
fact, c(x_i) and d(x_i) are the variables i?m mainly interested in. (a 
and b are nearly exactly known, so i can reduce the fitting_function to 
T(c(x_i),d(x_i),x_i)).


I hope you can see what my problem is, its late here and i?m tired, so 
maybe i haven?t explained very well.


Maybe the following will help

           ---
           \                                          2
minimize    |  {T_meas(x_i) - T(a,b,c(x_i),d(x_i),x_i)}
           /
           ---
           i

is what i?m lookig for. Is this possible with scipy.optimize.leastsq, or 
should i use some other routine therefor? And if so, which on?

Thanks so far
    Volker



From shao at msg.ucsf.edu  Tue Jul  3 18:34:33 2007
From: shao at msg.ucsf.edu (Lin Shao)
Date: Tue, 3 Jul 2007 15:34:33 -0700
Subject: [SciPy-user] optimization question
In-Reply-To: <468ACB2F.8030309@physik.uni-wuerzburg.de>
References: <468ACB2F.8030309@physik.uni-wuerzburg.de>
Message-ID: <baa383440707031534i42c816ve745589994cf97a3@mail.gmail.com>

Can't you express c(x) and d(x) as functions in terms of x? If so,
then I assume you'll get a bunch of scalar parameters c1,c2... for
c(x) and d1,d2,... for d(x). Then what you need is to fit
(a,b,c1,c2...,d1,d2...) using your measured data and known x_i.

scipy.optimize.leastsq is a good choice. But for some reason, the
Jacobian matrix has to be "column major" (each column represents one
observation point) and set col_deriv to 1. Otherwise it won't work. I
reported this bug weeks ago, but nobody seems to have answered with
authority that if it's a bug or not.


From robert.kern at gmail.com  Tue Jul  3 18:35:18 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 03 Jul 2007 17:35:18 -0500
Subject: [SciPy-user] optimization question
In-Reply-To: <468ACB2F.8030309@physik.uni-wuerzburg.de>
References: <468ACB2F.8030309@physik.uni-wuerzburg.de>
Message-ID: <468ACF26.8090306@gmail.com>

Volker Lorrmann wrote:
> Hello list,
> 
> i wanna fit a function to some measured datapoints T_m(x_i). The 
> function i wanna fit  is something like that,T_fit(a,b,c,d,x_i), where 
> a,b,c are the fitting-parameters. Fitting this with 
> scipy.optimize.leastsq would be easy (btw. is there anoterh way to fit 
> this, like fmin, fmin_powell, ...?).

All leastsq() really adds to fmin*() is that you return the residuals instead of
summing up the squares of the residuals. You can do that manually and use
whichever minimizer you need.

> The problem is, that a is a and b are _fixed_ scalar parameters. But c 
> and d are variables, that depend on x_i, c=c(x_i) and d=d(x_i). And in 
> fact, c(x_i) and d(x_i) are the variables i?m mainly interested in. (a 
> and b are nearly exactly known, so i can reduce the fitting_function to 
> T(c(x_i),d(x_i),x_i)).

Well, that's a big problem. You have twice as many variables to fit as you have
datapoints. There are possibly an infinite number of solutions that exactly(!)
fit your data.

Is it possible that you can reparameterize your problem? Perhaps c(.) and d(.)
could be formulated as functions that depend on some small number of parameters
besides x_i in order to smooth things out. You would then do least-squares to
optimize those parameters.

You might also be interested in some non-parametric inverse theory techniques,
but that requires a reasonably substantial investment in reading up on the topic
and examining your model.

-- 
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 cookedm at physics.mcmaster.ca  Tue Jul  3 19:36:37 2007
From: cookedm at physics.mcmaster.ca (David M. Cooke)
Date: Tue, 3 Jul 2007 19:36:37 -0400
Subject: [SciPy-user] Compile fails for r3123 on OS X 10.4 (Intel)
In-Reply-To: <cd7634ce0707031314u1f2827c3jabf4bde824eae543@mail.gmail.com>
References: <cd7634ce0706291033ga50ed5bj1e0d7537b09bb346@mail.gmail.com>
	<20070703200423.GA5477@arbutus.physics.mcmaster.ca>
	<cd7634ce0707031314u1f2827c3jabf4bde824eae543@mail.gmail.com>
Message-ID: <20070703233637.GA9394@arbutus.physics.mcmaster.ca>

On Tue, Jul 03, 2007 at 01:14:12PM -0700, Barry Wark wrote:
> Yes, I'm using a clean checkout from SVN. Just to double check against
> your setup, I'm using fftw-3 installed via macports as my fftw library
> 
> Barry

Still don't have your problem. Scipy is r3143, numpy is r3880. OS X
10.4.10 (Intel MacBook), Python 2.5 and 2.4. I'm using gfortran from
macports, but that shouldn't affect this.

One thing I can think of is checking your numpy install, as the
_fortranmodule.c is generated by f2py.

(in general, the latest svn scipy usually needs a recent svn numpy)

> On 7/3/07, David M. Cooke <cookedm at physics.mcmaster.ca> wrote:
> > On Fri, Jun 29, 2007 at 10:33:24AM -0700, Barry Wark wrote:
> > > I've been unable to compile scipy on OS X 10.4 (intel) from the recent
> > > trunk. Scipy built on this machine as of r2708. The output from
> > >
> > > python setup.py build
> > >
> > > is attached.
> > >
> > > I have fftw3 installed via macports (in /opt/local), and it appears
> > > that the build finds it properly, but the build fails with an error:
> > >
> > > building extension "scipy.fftpack._fftpack" sources
> > >  target build/src.macosx-10.3-fat-2.5/_fftpackmodule.c does not exist:
> > >   Assuming _fftpackmodule.c was generated with "build_src --inplace"
> > >   command.
> > > error: '_fftpackmodule.c' missing
> > >
> > > I would appreciate any advice or suggestions!
> >
> > I've got pretty much the same setup, and no problems. Did you delete the
> > build directory first (rm -rf build)?
> >
> > --
> > |>|\/|<
> > /--------------------------------------------------------------------------\
> > |David M. Cooke                      http://arbutus.physics.mcmaster.ca/dmc/
> > |cookedm at physics.mcmaster.ca
> > _______________________________________________
> > 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
> 

-- 
|>|\/|<
/--------------------------------------------------------------------------\
|David M. Cooke                      http://arbutus.physics.mcmaster.ca/dmc/
|cookedm at physics.mcmaster.ca


From matthieu.brucher at gmail.com  Wed Jul  4 01:51:21 2007
From: matthieu.brucher at gmail.com (Matthieu Brucher)
Date: Wed, 4 Jul 2007 07:51:21 +0200
Subject: [SciPy-user] optimization question
In-Reply-To: <468ACB2F.8030309@physik.uni-wuerzburg.de>
References: <468ACB2F.8030309@physik.uni-wuerzburg.de>
Message-ID: <e76aa17f0707032251j2c623b20mdf4c4a3f663ba51a@mail.gmail.com>

>
> i wanna fit a function to some measured datapoints T_m(x_i). The
> function i wanna fit  is something like that,T_fit(a,b,c,d,x_i), where
> a,b,c are the fitting-parameters. Fitting this with
> scipy.optimize.leastsq would be easy (btw. is there anoterh way to fit
> this, like fmin, fmin_powell, ...?).



Yes, you can define a cost function and use it fmin, ... or the generic
optimizers (although such helper functions will be made in the near future)


The problem is, that a is a and b are _fixed_ scalar parameters. But c
> and d are variables, that depend on x_i, c=c(x_i) and d=d(x_i). And in
> fact, c(x_i) and d(x_i) are the variables i?m mainly interested in. (a
> and b are nearly exactly known, so i can reduce the fitting_function to
> T(c(x_i),d(x_i),x_i)).


Then it's a vector of parameters you have to fit ? Create your own cost
function, it's probably the easiest.

Matthieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070704/f97ab0c5/attachment.html>

From lorrmann at physik.uni-wuerzburg.de  Wed Jul  4 02:51:55 2007
From: lorrmann at physik.uni-wuerzburg.de (Volker Lorrmann)
Date: Wed, 04 Jul 2007 08:51:55 +0200
Subject: [SciPy-user] optimization question
Message-ID: <468B438B.4080303@physik.uni-wuerzburg.de>


> Yes, you can define a cost function and use it fmin, ... or the generic
> optimizers (although such helper functions will be made in the near future)
> 
> Then it's a vector of parameters you have to fit ? Create your own cost
> function, it's probably the easiest.

> Matthieu

Hi Matthieu,

thanks for your answer, but can u be little bit more exactly please. What do you mean by a costfunction?
Maybe you can give a short example, if its not to much work.

 volker



From robert.kern at gmail.com  Wed Jul  4 02:58:22 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 04 Jul 2007 01:58:22 -0500
Subject: [SciPy-user] optimization question
In-Reply-To: <468B438B.4080303@physik.uni-wuerzburg.de>
References: <468B438B.4080303@physik.uni-wuerzburg.de>
Message-ID: <468B450E.6010604@gmail.com>

Volker Lorrmann wrote:
>> Yes, you can define a cost function and use it fmin, ... or the generic
>> optimizers (although such helper functions will be made in the near future)
>>
>> Then it's a vector of parameters you have to fit ? Create your own cost
>> function, it's probably the easiest.
> 
>> Matthieu
> 
> Hi Matthieu,
> 
> thanks for your answer, but can u be little bit more exactly please. What do you mean by a costfunction?
> Maybe you can give a short example, if its not to much work.


  ---
  \                                          2
   |  {T_meas(x_i) - T(a,b,c(x_i),d(x_i),x_i)}
  /
  ---
   i

-- 
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 gael.varoquaux at normalesup.org  Wed Jul  4 03:01:10 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Wed, 4 Jul 2007 09:01:10 +0200
Subject: [SciPy-user] optimization question
In-Reply-To: <468B438B.4080303@physik.uni-wuerzburg.de>
References: <468B438B.4080303@physik.uni-wuerzburg.de>
Message-ID: <20070704070110.GA28716@clipper.ens.fr>

On Wed, Jul 04, 2007 at 08:51:55AM +0200, Volker Lorrmann wrote:
> thanks for your answer, but can u be little bit more exactly please. What do you mean by a costfunction?
> Maybe you can give a short example, if its not to much work.

http://scipy.org/Cookbook/FittingData

HTH,

Ga?l


From zunzun at zunzun.com  Wed Jul  4 03:19:52 2007
From: zunzun at zunzun.com (zunzun at zunzun.com)
Date: Wed, 4 Jul 2007 03:19:52 -0400
Subject: [SciPy-user] optimization question
In-Reply-To: <468B438B.4080303@physik.uni-wuerzburg.de>
References: <468B438B.4080303@physik.uni-wuerzburg.de>
Message-ID: <20070704071952.GB1572@zunzun.com>

On Wed, Jul 04, 2007 at 08:51:55AM +0200, Volker Lorrmann wrote:
> 
> What do you mean by a costfunction?

If I want to minimize the sum of squared errors for some
data as inputs to an equation, my "cost function" will
calculate the sum of square derrors.

If I want to minimize the inverse sum of absolute errors
for the same data and equation, my "cost function" would
calculate that instead.

The cost function calculates whatever it is your minimizer
will be minimizing.


> Maybe you can give a short example, if its not to much work.

Some badly written pseudocode to illustrate;

First define a cost function, this returns a value
to be minimized - here the sum of squared error.

function SSQ_cost_func(data, equation_function)
{
     calculate and return the sum of squared error
}

now pass this to the minimizer:

minimize(SSQ_cost_func, data, equation_function)


BTW, don't people on this mailing list ever sleep?
It's 2:00 AM here in Birmingham, Alabama.

     James Phillips
     http://zunzun.com



From lorrmann at physik.uni-wuerzburg.de  Wed Jul  4 03:20:16 2007
From: lorrmann at physik.uni-wuerzburg.de (Volker Lorrmann)
Date: Wed, 04 Jul 2007 09:20:16 +0200
Subject: [SciPy-user] optimization question
Message-ID: <468B4A30.2060408@physik.uni-wuerzburg.de>

Hi Robert,

 > All leastsq() really adds to fmin*() is that you return the residuals 
 > instead of summing up the squares of the residuals. You can do that
 > manually and use whichever minimizer you need.

Can you give me a short example, how to do so? I?ve asked this also 
Matthieu, so if he will answer, you probably don?t need to. But two 
examples would be better than one ;)

 > Well, that's a big problem. You have twice as many variables to fit 
as > you have datapoints. There are possibly an infinite number of
 > solutions that exactly(!) fit your data.

I see, indeed i need as many datapoints (equations) as variables, to 
solve the problem.

 > Is it possible that you can reparameterize your problem? Perhaps c(.) 
 > and d(.)
 > could be formulated as functions that depend on some small number of 
  > parameters
 > besides x_i in order to smooth things out. You would then do
 > least-squares to optimize those parameters.

I think i can do that. There are some physical constraints for c(x_i) 
and d(x_i), which should make it possible to reparameterize c an d.

volker




From openopt at ukr.net  Wed Jul  4 03:28:44 2007
From: openopt at ukr.net (dmitrey)
Date: Wed, 04 Jul 2007 10:28:44 +0300
Subject: [SciPy-user] optimization question
In-Reply-To: <468B4A30.2060408@physik.uni-wuerzburg.de>
References: <468B4A30.2060408@physik.uni-wuerzburg.de>
Message-ID: <468B4C2C.8050402@ukr.net>

Volker Lorrmann wrote:
> Hi Robert,
>
>  > All leastsq() really adds to fmin*() is that you return the residuals 
>  > instead of summing up the squares of the residuals. You can do that
>  > manually and use whichever minimizer you need.
>
> Can you give me a short example, how to do so? I?ve asked this also 
> Matthieu, so if he will answer, you probably don?t need to. But two 
> examples would be better than one ;)
>
>  > Well, that's a big problem. You have twice as many variables to fit 
> as > you have datapoints. There are possibly an infinite number of
>  > solutions that exactly(!) fit your data.
>
> I see, indeed i need as many datapoints (equations) as variables, to 
> solve the problem.
>
>  > Is it possible that you can reparameterize your problem? Perhaps c(.) 
>  > and d(.)
>  > could be formulated as functions that depend on some small number of 
>   > parameters
>  > besides x_i in order to smooth things out. You would then do
>  > least-squares to optimize those parameters.
>
> I think i can do that. There are some physical constraints for c(x_i) 
> and d(x_i), which should make it possible to reparameterize c an d.
>
> volker
>
>   
as one more solution, for constrained NLP problem you can try "lincher" 
solver from scikits.openopt toolkit (free, BSD license). There is also a 
ralg solver that is good for both NLP and non-smooth problems, but it's 
currently suitable for unconstrained problems only.
See example of lincher usage here:
http://openopt.blogspot.com/2007/06/constrained-nlp-example-with-gradients.html

>BTW, don't people on this mailing list ever sleep?
>It's 2:00 AM here in Birmingham, Alabama.

>     James Phillips
>     http://zunzun.com

Here in Ukraine it's 10.27AM, time to work :)

Regards, Dmitrey.







From david at ar.media.kyoto-u.ac.jp  Wed Jul  4 03:26:18 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Wed, 04 Jul 2007 16:26:18 +0900
Subject: [SciPy-user] optimization question
In-Reply-To: <20070704071952.GB1572@zunzun.com>
References: <468B438B.4080303@physik.uni-wuerzburg.de>
	<20070704071952.GB1572@zunzun.com>
Message-ID: <468B4B9A.1010705@ar.media.kyoto-u.ac.jp>

zunzun at zunzun.com wrote:
> BTW, don't people on this mailing list ever sleep?
> It's 2:00 AM here in Birmingham, Alabama.
Well, it is 4 p.m here in Japan :)

David


From matthieu.brucher at gmail.com  Wed Jul  4 03:40:39 2007
From: matthieu.brucher at gmail.com (Matthieu Brucher)
Date: Wed, 4 Jul 2007 09:40:39 +0200
Subject: [SciPy-user] optimization question
In-Reply-To: <20070704071952.GB1572@zunzun.com>
References: <468B438B.4080303@physik.uni-wuerzburg.de>
	<20070704071952.GB1572@zunzun.com>
Message-ID: <e76aa17f0707040040hc74593s5128c979d3f261f6@mail.gmail.com>

>
> BTW, don't people on this mailing list ever sleep?
> It's 2:00 AM here in Birmingham, Alabama.
>

It's 09:40 AM in France ;)

Matthieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070704/94872483/attachment.html>

From matthieu.brucher at gmail.com  Wed Jul  4 03:48:03 2007
From: matthieu.brucher at gmail.com (Matthieu Brucher)
Date: Wed, 4 Jul 2007 09:48:03 +0200
Subject: [SciPy-user] optimization question
In-Reply-To: <468B4A30.2060408@physik.uni-wuerzburg.de>
References: <468B4A30.2060408@physik.uni-wuerzburg.de>
Message-ID: <e76aa17f0707040048r382002acp7f04dae639194c71@mail.gmail.com>

2007/7/4, Volker Lorrmann <lorrmann at physik.uni-wuerzburg.de>:
>
> Hi Robert,
>
> > All leastsq() really adds to fmin*() is that you return the residuals
> > instead of summing up the squares of the residuals. You can do that
> > manually and use whichever minimizer you need.
>
> Can you give me a short example, how to do so? I?ve asked this also
> Matthieu, so if he will answer, you probably don?t need to. But two
> examples would be better than one ;)


My view is heavily influenced by my own optimizers, but here is the thing :

you have points y (possibly multidimensional) and points x (multidimensional
too) and a fitting function f(x, parameters)
You construct a new class (for instance) that take x and y as parameters :
class cost_fun(object:
  def __init__(self, x, y):
    self.x = x
    self.y = y


and you define a call method that will be used by fmin, for instance :
  def __call__(self, parameters):
    cost = 0
    for (xi, yi) in (x, y):
      cost += numpy.sum((yi - f(xi, parameters))**2) # you can use some
parameters instead of all parameters, just iterate over the part you need
    return cost

And voil?, your basic cost function is created.
Here a quadratic cost is used (this is what leastsq may do), but you can use
a robust cost instead.

I hope I was clear.

Matthieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070704/507e3338/attachment.html>

From subscriptions at smart-knowhow.de  Wed Jul  4 05:23:04 2007
From: subscriptions at smart-knowhow.de (Andrew Smart)
Date: Wed, 4 Jul 2007 11:23:04 +0200
Subject: [SciPy-user] How to "move" data within an array
Message-ID: <0ML21M-1I614g2fqj-0000xv@mrelayeu.kundenserver.de>

Hi folks,

I'm using numpy arrays for storing data which is generated within an engine.
I'm using the topmost dimension as time axis: every row represents a full
set of data created by the engine while one round.

Say: i have an array for storing prices (e.g. 10 different prices are
generated within one engine round). I'm storing/using the last 5 rounds, so
I get an array with the dimensions (5,10).

If the engine runs longer than 5 rounds I have to "remove" the oldest record
and move the younger records one position back. 

Since I've a lot of such arrays I would like to use the most efficient
method avaiable in numpy. On a pure memory-orientated view this would be
just to copy ("move") the memory blocks from the younger 4 rows one row
further, thus having the first row for the new data.

In the C API I see some functions like copyswap() and memmove() which
indicate that such operations are possible at the C API level. But I'm not
sure the correct approach on the Python level.

Taking slices may be one options - but the new slice will then occupy new
memory, causing memory fragmentation...
Looping over all data items, all rows is time consuming and surely wasting
resources...

Any pointers/ideas?

Thanks,
Andrew 




From lbolla at gmail.com  Wed Jul  4 05:33:29 2007
From: lbolla at gmail.com (lorenzo bolla)
Date: Wed, 4 Jul 2007 11:33:29 +0200
Subject: [SciPy-user] How to "move" data within an array
In-Reply-To: <0ML21M-1I614g2fqj-0000xv@mrelayeu.kundenserver.de>
References: <0ML21M-1I614g2fqj-0000xv@mrelayeu.kundenserver.de>
Message-ID: <80c99e790707040233s3593f4d9g769d9814e7366dcb@mail.gmail.com>

why not using a list of 1D arrays?
but why do you want to physically move your rows? you can simply use an
integer as a pointer to the row of the "current time": then you update this
integer every timestep (+1), taking its "modulo 5" to cycle through
the rows.

my two cents.
lorenzo.

On 7/4/07, Andrew Smart <subscriptions at smart-knowhow.de> wrote:
>
> Hi folks,
>
> I'm using numpy arrays for storing data which is generated within an
> engine.
> I'm using the topmost dimension as time axis: every row represents a full
> set of data created by the engine while one round.
>
> Say: i have an array for storing prices (e.g. 10 different prices are
> generated within one engine round). I'm storing/using the last 5 rounds,
> so
> I get an array with the dimensions (5,10).
>
> If the engine runs longer than 5 rounds I have to "remove" the oldest
> record
> and move the younger records one position back.
>
> Since I've a lot of such arrays I would like to use the most efficient
> method avaiable in numpy. On a pure memory-orientated view this would be
> just to copy ("move") the memory blocks from the younger 4 rows one row
> further, thus having the first row for the new data.
>
> In the C API I see some functions like copyswap() and memmove() which
> indicate that such operations are possible at the C API level. But I'm not
> sure the correct approach on the Python level.
>
> Taking slices may be one options - but the new slice will then occupy new
> memory, causing memory fragmentation...
> Looping over all data items, all rows is time consuming and surely wasting
> resources...
>
> Any pointers/ideas?
>
> Thanks,
> Andrew
>
>
> _______________________________________________
> 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/20070704/dcc84016/attachment.html>

From subscriptions at smart-knowhow.de  Wed Jul  4 05:44:59 2007
From: subscriptions at smart-knowhow.de (Andrew Smart)
Date: Wed, 4 Jul 2007 11:44:59 +0200
Subject: [SciPy-user] How to "move" data within an array
In-Reply-To: <80c99e790707040233s3593f4d9g769d9814e7366dcb@mail.gmail.com>
Message-ID: <0ML2xA-1I61Pt1N7i-0003mi@mrelayeu.kundenserver.de>

Hi Lorenzo,
 
thanks for your 2cents. 

The pointer method isn't practical for my purposes: I want to have the
ability to access the "historical" data within the engine on various ways,
e.g. "price average of the last 3 periods", where the array itself stores
still 5 periods. The pointer method would require to re-calculate the time
axis and especially to manage the "wrap", like: current time row is 3, so "3
periods back" would be rows 2, 1 and 5.

I would like to use the numpy functions as sum(), avg() etc. on the arrays,
so having single 1d arrays (one row = one array) does not really make sense.

But thanks for the idea,
Andrew
________________________________

	Von: scipy-user-bounces at scipy.org
[mailto:scipy-user-bounces at scipy.org] Im Auftrag von lorenzo bolla
	Gesendet: Mittwoch, 4. Juli 2007 11:33
	An: SciPy Users List
	Betreff: Re: [SciPy-user] How to "move" data within an array
	
	
	why not using a list of 1D arrays?
	but why do you want to physically move your rows? you can simply use
an integer as a pointer to the row of the "current time": then you update
this integer every timestep (+1), taking its "modulo 5" to cycle through the
rows. 
	 
	my two cents.
	lorenzo.
	 
	On 7/4/07, Andrew Smart <subscriptions at smart-knowhow.de> wrote: 

		Hi folks,
		
		I'm using numpy arrays for storing data which is generated
within an engine.
		I'm using the topmost dimension as time axis: every row
represents a full 
		set of data created by the engine while one round.
		
		Say: i have an array for storing prices (e.g. 10 different
prices are
		generated within one engine round). I'm storing/using the
last 5 rounds, so
		I get an array with the dimensions (5,10). 
		
		If the engine runs longer than 5 rounds I have to "remove"
the oldest record
		and move the younger records one position back.
		
		Since I've a lot of such arrays I would like to use the most
efficient 
		method avaiable in numpy. On a pure memory-orientated view
this would be
		just to copy ("move") the memory blocks from the younger 4
rows one row
		further, thus having the first row for the new data.
		
		In the C API I see some functions like copyswap() and
memmove() which
		indicate that such operations are possible at the C API
level. But I'm not
		sure the correct approach on the Python level.
		
		Taking slices may be one options - but the new slice will
then occupy new 
		memory, causing memory fragmentation...
		Looping over all data items, all rows is time consuming and
surely wasting
		resources...
		
		Any pointers/ideas?
		
		Thanks,
		Andrew
		
		
		_______________________________________________ 
		SciPy-user mailing list
		SciPy-user at scipy.org
		http://projects.scipy.org/mailman/listinfo/scipy-user
		





From lbolla at gmail.com  Wed Jul  4 07:21:38 2007
From: lbolla at gmail.com (lorenzo bolla)
Date: Wed, 4 Jul 2007 13:21:38 +0200
Subject: [SciPy-user] How to "move" data within an array
In-Reply-To: <0ML2xA-1I61Pt1N7i-0003mi@mrelayeu.kundenserver.de>
References: <80c99e790707040233s3593f4d9g769d9814e7366dcb@mail.gmail.com>
	<0ML2xA-1I61Pt1N7i-0003mi@mrelayeu.kundenserver.de>
Message-ID: <80c99e790707040421l62aee169p2fff8b4611415e90@mail.gmail.com>

gotcha. you could use numpy.roll, then.


In [10]: x = numpy.array([[1,2,3],[4,5,6]])

In [11]: x
Out[11]:
array([[1, 2, 3],
       [4, 5, 6]])

In [12]: numpy.roll(x,1)
Out[12]:
array([[6, 1, 2],
       [3, 4, 5]])

In [13]: numpy.roll(x,2)
Out[13]:
array([[5, 6, 1],
       [2, 3, 4]])

In [14]: numpy.roll(x,3)
Out[14]:
array([[4, 5, 6],
       [1, 2, 3]])


lorenzo.


On 7/4/07, Andrew Smart <subscriptions at smart-knowhow.de> wrote:
>
> Hi Lorenzo,
>
> thanks for your 2cents.
>
> The pointer method isn't practical for my purposes: I want to have the
> ability to access the "historical" data within the engine on various ways,
> e.g. "price average of the last 3 periods", where the array itself stores
> still 5 periods. The pointer method would require to re-calculate the time
> axis and especially to manage the "wrap", like: current time row is 3, so
> "3
> periods back" would be rows 2, 1 and 5.
>
> I would like to use the numpy functions as sum(), avg() etc. on the
> arrays,
> so having single 1d arrays (one row = one array) does not really make
> sense.
>
> But thanks for the idea,
> Andrew
> ________________________________
>
>        Von: scipy-user-bounces at scipy.org
> [mailto:scipy-user-bounces at scipy.org] Im Auftrag von lorenzo bolla
>        Gesendet: Mittwoch, 4. Juli 2007 11:33
>        An: SciPy Users List
>        Betreff: Re: [SciPy-user] How to "move" data within an array
>
>
>        why not using a list of 1D arrays?
>        but why do you want to physically move your rows? you can simply
> use
> an integer as a pointer to the row of the "current time": then you update
> this integer every timestep (+1), taking its "modulo 5" to cycle through
> the
> rows.
>
>        my two cents.
>        lorenzo.
>
>        On 7/4/07, Andrew Smart <subscriptions at smart-knowhow.de> wrote:
>
>                Hi folks,
>
>                I'm using numpy arrays for storing data which is generated
> within an engine.
>                I'm using the topmost dimension as time axis: every row
> represents a full
>                set of data created by the engine while one round.
>
>                Say: i have an array for storing prices (e.g. 10 different
> prices are
>                generated within one engine round). I'm storing/using the
> last 5 rounds, so
>                I get an array with the dimensions (5,10).
>
>                If the engine runs longer than 5 rounds I have to "remove"
> the oldest record
>                and move the younger records one position back.
>
>                Since I've a lot of such arrays I would like to use the
> most
> efficient
>                method avaiable in numpy. On a pure memory-orientated view
> this would be
>                just to copy ("move") the memory blocks from the younger 4
> rows one row
>                further, thus having the first row for the new data.
>
>                In the C API I see some functions like copyswap() and
> memmove() which
>                indicate that such operations are possible at the C API
> level. But I'm not
>                sure the correct approach on the Python level.
>
>                Taking slices may be one options - but the new slice will
> then occupy new
>                memory, causing memory fragmentation...
>                Looping over all data items, all rows is time consuming and
> surely wasting
>                resources...
>
>                Any pointers/ideas?
>
>                Thanks,
>                Andrew
>
>
>                _______________________________________________
>                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/20070704/f3df8eee/attachment.html>

From lbolla at gmail.com  Wed Jul  4 07:23:35 2007
From: lbolla at gmail.com (lorenzo bolla)
Date: Wed, 4 Jul 2007 13:23:35 +0200
Subject: [SciPy-user] How to "move" data within an array
In-Reply-To: <80c99e790707040421l62aee169p2fff8b4611415e90@mail.gmail.com>
References: <80c99e790707040233s3593f4d9g769d9814e7366dcb@mail.gmail.com>
	<0ML2xA-1I61Pt1N7i-0003mi@mrelayeu.kundenserver.de>
	<80c99e790707040421l62aee169p2fff8b4611415e90@mail.gmail.com>
Message-ID: <80c99e790707040423s532a42afwf1882bc8a78b3572@mail.gmail.com>

it's better to specify the axis, though:

In [20]: numpy.roll(x,1,axis = 1)
Out[20]:
array([[3, 1, 2],
       [6, 4, 5]])

In [21]: numpy.roll(x,1)
Out[21]:
array([[6, 1, 2],
       [3, 4, 5]])


lorenzo.

On 7/4/07, lorenzo bolla <lbolla at gmail.com> wrote:
>
> gotcha. you could use numpy.roll, then.
>
>
> In [10]: x = numpy.array([[1,2,3],[4,5,6]])
>
> In [11]: x
> Out[11]:
> array([[1, 2, 3],
>        [4, 5, 6]])
>
> In [12]: numpy.roll(x,1)
> Out[12]:
> array([[6, 1, 2],
>        [3, 4, 5]])
>
> In [13]: numpy.roll(x,2)
> Out[13]:
> array([[5, 6, 1],
>        [2, 3, 4]])
>
> In [14]: numpy.roll(x,3)
> Out[14]:
> array([[4, 5, 6],
>        [1, 2, 3]])
>
>
> lorenzo.
>
>
> On 7/4/07, Andrew Smart <subscriptions at smart-knowhow.de> wrote:
> >
> > Hi Lorenzo,
> >
> > thanks for your 2cents.
> >
> > The pointer method isn't practical for my purposes: I want to have the
> > ability to access the "historical" data within the engine on various
> > ways,
> > e.g. "price average of the last 3 periods", where the array itself
> > stores
> > still 5 periods. The pointer method would require to re-calculate the
> > time
> > axis and especially to manage the "wrap", like: current time row is 3,
> > so "3
> > periods back" would be rows 2, 1 and 5.
> >
> > I would like to use the numpy functions as sum(), avg() etc. on the
> > arrays,
> > so having single 1d arrays (one row = one array) does not really make
> > sense.
> >
> > But thanks for the idea,
> > Andrew
> > ________________________________
> >
> >        Von: scipy-user-bounces at scipy.org
> > [mailto:scipy-user-bounces at scipy.org] Im Auftrag von lorenzo bolla
> >        Gesendet: Mittwoch, 4. Juli 2007 11:33
> >        An: SciPy Users List
> >        Betreff: Re: [SciPy-user] How to "move" data within an array
> >
> >
> >        why not using a list of 1D arrays?
> >        but why do you want to physically move your rows? you can simply
> > use
> > an integer as a pointer to the row of the "current time": then you
> > update
> > this integer every timestep (+1), taking its "modulo 5" to cycle through
> > the
> > rows.
> >
> >        my two cents.
> >        lorenzo.
> >
> >        On 7/4/07, Andrew Smart < subscriptions at smart-knowhow.de> wrote:
> >
> >                Hi folks,
> >
> >                I'm using numpy arrays for storing data which is
> > generated
> > within an engine.
> >                I'm using the topmost dimension as time axis: every row
> > represents a full
> >                set of data created by the engine while one round.
> >
> >                Say: i have an array for storing prices (e.g. 10
> > different
> > prices are
> >                generated within one engine round). I'm storing/using the
> >
> > last 5 rounds, so
> >                I get an array with the dimensions (5,10).
> >
> >                If the engine runs longer than 5 rounds I have to
> > "remove"
> > the oldest record
> >                and move the younger records one position back.
> >
> >                Since I've a lot of such arrays I would like to use the
> > most
> > efficient
> >                method avaiable in numpy. On a pure memory-orientated
> > view
> > this would be
> >                just to copy ("move") the memory blocks from the younger
> > 4
> > rows one row
> >                further, thus having the first row for the new data.
> >
> >                In the C API I see some functions like copyswap() and
> > memmove() which
> >                indicate that such operations are possible at the C API
> > level. But I'm not
> >                sure the correct approach on the Python level.
> >
> >                Taking slices may be one options - but the new slice will
> > then occupy new
> >                memory, causing memory fragmentation...
> >                Looping over all data items, all rows is time consuming
> > and
> > surely wasting
> >                resources...
> >
> >                Any pointers/ideas?
> >
> >                Thanks,
> >                Andrew
> >
> >
> >                _______________________________________________
> >                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/20070704/376b14f4/attachment.html>

From subscriptions at smart-knowhow.de  Wed Jul  4 10:50:35 2007
From: subscriptions at smart-knowhow.de (Andrew Smart)
Date: Wed, 4 Jul 2007 16:50:35 +0200
Subject: [SciPy-user] How to "move" data within an array
In-Reply-To: <80c99e790707040423s532a42afwf1882bc8a78b3572@mail.gmail.com>
Message-ID: <0MKwpI-1I66Be08GV-00044o@mrelayeu.kundenserver.de>

I found roll() in the meantime also as one option. It's a feasible approach,
but still may cause memory fragmentation. I'll take roll() for the time
being - and some time later I'll see if someone makes me a C based function
which does the same without copying the array...
 
But thanks for your answers!
 
Andrew
 
 

smart:elligence Unternehmensberatung
Goethestr. 55
40237 D?sseldorf
Tel.: 0211 6015 699
Mob: 0170 223 4440

VERTRAULICHKEITSHINWEIS:
Der Inhalt dieser E-Mail ist vertraulich und f?r den Nutzer der E-Mail
Adresse bestimmt, an den die Nachricht geschickt wurde; sie kann dar?ber
hinaus durch besondere Bestimmungen gesch?tzt sein. Wenn Sie nicht der
Adressat dieser E-Mail sind, d?rfen Sie diese nicht kopieren, weiterleiten,
weitergeben oder sie ganz oder teilweise in irgendeiner Weise nutzen. Wenn
Sie diese E-Mail f?lschlicherweise erhalten haben, benachrichtigen Sie bitte
den Absender, indem Sie auf diese Nachricht antworten.

  

 


  _____  

Von: scipy-user-bounces at scipy.org [mailto:scipy-user-bounces at scipy.org] Im
Auftrag von lorenzo bolla
Gesendet: Mittwoch, 4. Juli 2007 13:24
An: SciPy Users List
Betreff: Re: [SciPy-user] How to "move" data within an array


it's better to specify the axis, though:

In [20]: numpy.roll(x,1,axis = 1)
Out[20]: 
array([[3, 1, 2],
       [6, 4, 5]])

In [21]: numpy.roll(x,1)
Out[21]: 
array([[6, 1, 2],
       [3, 4, 5]])



 
lorenzo.
 
On 7/4/07, lorenzo bolla <lbolla at gmail.com> wrote: 

gotcha. you could use numpy.roll, then.
 
 
In [10]: x = numpy.array([[1,2,3],[4,5,6]])

In [11]: x
Out[11]: 
array([[1, 2, 3],
       [4, 5, 6]])

In [12]: numpy.roll(x,1)
Out[12]: 
array([[6, 1, 2],
       [3, 4, 5]])

In [13]: numpy.roll(x,2)
Out[13]: 
array([[5, 6, 1],
       [2, 3, 4]])

In [14]: numpy.roll(x,3)
Out[14]: 
array([[4, 5, 6],
       [1, 2, 3]])




 
lorenzo.

 
On 7/4/07, Andrew Smart <subscriptions at smart-knowhow.de
<mailto:subscriptions at smart-knowhow.de> > wrote: 

Hi Lorenzo,

thanks for your 2cents.

The pointer method isn't practical for my purposes: I want to have the 
ability to access the "historical" data within the engine on various ways,
e.g. "price average of the last 3 periods", where the array itself stores
still 5 periods. The pointer method would require to re-calculate the time 
axis and especially to manage the "wrap", like: current time row is 3, so "3
periods back" would be rows 2, 1 and 5.

I would like to use the numpy functions as sum(), avg() etc. on the arrays, 
so having single 1d arrays (one row = one array) does not really make sense.

But thanks for the idea,
Andrew
________________________________

       Von: scipy-user-bounces at scipy.org
[mailto:scipy-user-bounces at scipy.org] Im Auftrag von lorenzo bolla 
       Gesendet: Mittwoch, 4. Juli 2007 11:33
       An: SciPy Users List 
       Betreff: Re: [SciPy-user] How to "move" data within an array


       why not using a list of 1D arrays?
       but why do you want to physically move your rows? you can simply use 
an integer as a pointer to the row of the "current time": then you update 
this integer every timestep (+1), taking its "modulo 5" to cycle through the
rows.

       my two cents.
       lorenzo. 

       On 7/4/07, Andrew Smart <  <mailto:subscriptions at smart-knowhow.de>
subscriptions at smart-knowhow.de> wrote:

               Hi folks, 

               I'm using numpy arrays for storing data which is generated
within an engine.
               I'm using the topmost dimension as time axis: every row 
represents a full
               set of data created by the engine while one round. 

               Say: i have an array for storing prices (e.g. 10 different
prices are
               generated within one engine round). I'm storing/using the 
last 5 rounds, so
               I get an array with the dimensions (5,10). 

               If the engine runs longer than 5 rounds I have to "remove"
the oldest record
               and move the younger records one position back. 

               Since I've a lot of such arrays I would like to use the most 
efficient
               method avaiable in numpy. On a pure memory-orientated view
this would be
               just to copy ("move") the memory blocks from the younger 4 
rows one row
               further, thus having the first row for the new data. 

               In the C API I see some functions like copyswap() and
memmove() which
               indicate that such operations are possible at the C API 
level. But I'm not
               sure the correct approach on the Python level. 

               Taking slices may be one options - but the new slice will
then occupy new
               memory, causing memory fragmentation... 
               Looping over all data items, all rows is time consuming and 
surely wasting
               resources...

               Any pointers/ideas?

               Thanks,
               Andrew 


               _______________________________________________ 
               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/20070704/0c5720af/attachment.html>

From peridot.faceted at gmail.com  Wed Jul  4 11:49:52 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Wed, 4 Jul 2007 11:49:52 -0400
Subject: [SciPy-user] How to "move" data within an array
In-Reply-To: <0MKwpI-1I66Be08GV-00044o@mrelayeu.kundenserver.de>
References: <80c99e790707040423s532a42afwf1882bc8a78b3572@mail.gmail.com>
	<0MKwpI-1I66Be08GV-00044o@mrelayeu.kundenserver.de>
Message-ID: <ce557a360707040849y4f66bff3m5f9c474e4b703772@mail.gmail.com>

On 04/07/07, Andrew Smart <subscriptions at smart-knowhow.de> wrote:

> I found roll() in the meantime also as one option. It's a feasible approach,
> but still may cause memory fragmentation. I'll take roll() for the time
> being - and some time later I'll see if someone makes me a C based function
> which does the same without copying the array...

I don't think memory fragmentation should be a concern. For one thing,
as you've described the problem, the array size isn't changing, so it
will almost certainly be copied back and forth between two memory
blocks. For another, for an array big enough for you to care about,
malloc() will request a new hunk of memory from the OS, and then free
it back to the OS when you're done with it. (Well, on Linux anyway.)

If you do want to move the array in place, you could try
A[:-1]=A[1:]
I don't know if that is smart enough to use memmove() if the copy is
of a contiguous block, but at the least it will be done in place by C
code. (The implementation of that optimization is complicated by the
semantics of that operation - since source and destination overlap,
the order of the indexing matters.)

A true in-place roll() is tricky and might have very poor cache
behaviour (because roll() tries to preserve all the array elements, so
rolling a 7-element array only has to loop around once, skipping by 3,
while rolling a 9-element array has to loop around nine times,
skipping by three).

Memory allocation is cheap and efficient, so eliminating temporaries
is less helpful than it might seem. In particular, it's worth knowing
that numpy array memory is a leaf with respect to the python garbage
collector - it doesn't need to be traversed, it just disappears when
the array objects pointing to it go. Also, as I said above, at least
on Linux, big monolithic allocations like numpy's go on fresh pages
from the OS, and are given back to it when done.

Is this actually the slowdown in your application?

Anne


From subscriptions at smart-knowhow.de  Wed Jul  4 13:14:27 2007
From: subscriptions at smart-knowhow.de (Andrew Smart)
Date: Wed, 4 Jul 2007 19:14:27 +0200
Subject: [SciPy-user] How to "move" data within an array
In-Reply-To: <ce557a360707040849y4f66bff3m5f9c474e4b703772@mail.gmail.com>
Message-ID: <0ML25U-1I68Qr40eI-0006EF@mrelayeu.kundenserver.de>

Hi Anne,

thanks for your very helpful input.

Currently I don't have any slowdowns since I'm working on the first version
of the software. I just tried to prevent mistakes by design. I already know
that I will have very large arrays - and since I'm new to numpy I thoughed
"ask the pro's". Better ask than to have to refactor the software later
because I took the wrong approach.

Numpy is definitivly a very sophisticated package with lots of power - but
this implies also that there may be facts very important for good
performance "hidden" below.

I'll take the roll() approach and test it with lots of data. The target
plattform is primarily Linux/Unix, but I'll have to make showcases on
Windows too... We'll see.

Thanks a lot,
Andrew


> -----Urspr?ngliche Nachricht-----
> Von: scipy-user-bounces at scipy.org 
> [mailto:scipy-user-bounces at scipy.org] Im Auftrag von Anne Archibald
> Gesendet: Mittwoch, 4. Juli 2007 17:50
> An: SciPy Users List
> Betreff: Re: [SciPy-user] How to "move" data within an array
> 
> On 04/07/07, Andrew Smart <subscriptions at smart-knowhow.de> wrote:
> 
> > I found roll() in the meantime also as one option. It's a feasible 
> > approach, but still may cause memory fragmentation. I'll 
> take roll() 
> > for the time being - and some time later I'll see if 
> someone makes me 
> > a C based function which does the same without copying the array...
> 
> I don't think memory fragmentation should be a concern. For 
> one thing, as you've described the problem, the array size 
> isn't changing, so it will almost certainly be copied back 
> and forth between two memory blocks. For another, for an 
> array big enough for you to care about,
> malloc() will request a new hunk of memory from the OS, and 
> then free it back to the OS when you're done with it. (Well, 
> on Linux anyway.)
> 
> If you do want to move the array in place, you could try 
> A[:-1]=A[1:] I don't know if that is smart enough to use 
> memmove() if the copy is of a contiguous block, but at the 
> least it will be done in place by C code. (The implementation 
> of that optimization is complicated by the semantics of that 
> operation - since source and destination overlap, the order 
> of the indexing matters.)
> 
> A true in-place roll() is tricky and might have very poor 
> cache behaviour (because roll() tries to preserve all the 
> array elements, so rolling a 7-element array only has to loop 
> around once, skipping by 3, while rolling a 9-element array 
> has to loop around nine times, skipping by three).
> 
> Memory allocation is cheap and efficient, so eliminating 
> temporaries is less helpful than it might seem. In 
> particular, it's worth knowing that numpy array memory is a 
> leaf with respect to the python garbage collector - it 
> doesn't need to be traversed, it just disappears when the 
> array objects pointing to it go. Also, as I said above, at 
> least on Linux, big monolithic allocations like numpy's go on 
> fresh pages from the OS, and are given back to it when done.
> 
> Is this actually the slowdown in your application?
> 
> Anne
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
> 



From lorrmann at physik.uni-wuerzburg.de  Wed Jul  4 13:26:23 2007
From: lorrmann at physik.uni-wuerzburg.de (Volker Lorrmann)
Date: Wed, 04 Jul 2007 19:26:23 +0200
Subject: [SciPy-user] optimization question
In-Reply-To: <468B438B.4080303@physik.uni-wuerzburg.de>
References: <468B438B.4080303@physik.uni-wuerzburg.de>
Message-ID: <468BD83F.9050502@physik.uni-wuerzburg.de>

So many answers. :)
Thanks to all authors

  volker


From strawman at astraw.com  Wed Jul  4 15:05:28 2007
From: strawman at astraw.com (Andrew Straw)
Date: Wed, 04 Jul 2007 12:05:28 -0700
Subject: [SciPy-user] How to "move" data within an array
In-Reply-To: <0ML2xA-1I61Pt1N7i-0003mi@mrelayeu.kundenserver.de>
References: <0ML2xA-1I61Pt1N7i-0003mi@mrelayeu.kundenserver.de>
Message-ID: <468BEF78.6090101@astraw.com>

Another way to use a list of arrays is to it as a FIFO (First In First
Out): (However, I realize this won't work with your other requirements
such as sum()).

LoA = []

for new_array in new_array_source():
    LoA.append(new_array)
    if len(LoA) > 5:
        old_array = LoA.pop(0)

Andrew Smart wrote:
> Hi Lorenzo,
>  
> thanks for your 2cents. 
> 
> The pointer method isn't practical for my purposes: I want to have the
> ability to access the "historical" data within the engine on various ways,
> e.g. "price average of the last 3 periods", where the array itself stores
> still 5 periods. The pointer method would require to re-calculate the time
> axis and especially to manage the "wrap", like: current time row is 3, so "3
> periods back" would be rows 2, 1 and 5.
> 
> I would like to use the numpy functions as sum(), avg() etc. on the arrays,
> so having single 1d arrays (one row = one array) does not really make sense.
> 
> But thanks for the idea,
> Andrew
> ________________________________
> 
> 	Von: scipy-user-bounces at scipy.org
> [mailto:scipy-user-bounces at scipy.org] Im Auftrag von lorenzo bolla
> 	Gesendet: Mittwoch, 4. Juli 2007 11:33
> 	An: SciPy Users List
> 	Betreff: Re: [SciPy-user] How to "move" data within an array
> 	
> 	
> 	why not using a list of 1D arrays?
> 	but why do you want to physically move your rows? you can simply use
> an integer as a pointer to the row of the "current time": then you update
> this integer every timestep (+1), taking its "modulo 5" to cycle through the
> rows. 
> 	 
> 	my two cents.
> 	lorenzo.
> 	 
> 	On 7/4/07, Andrew Smart <subscriptions at smart-knowhow.de> wrote: 
> 
> 		Hi folks,
> 		
> 		I'm using numpy arrays for storing data which is generated
> within an engine.
> 		I'm using the topmost dimension as time axis: every row
> represents a full 
> 		set of data created by the engine while one round.
> 		
> 		Say: i have an array for storing prices (e.g. 10 different
> prices are
> 		generated within one engine round). I'm storing/using the
> last 5 rounds, so
> 		I get an array with the dimensions (5,10). 
> 		
> 		If the engine runs longer than 5 rounds I have to "remove"
> the oldest record
> 		and move the younger records one position back.
> 		
> 		Since I've a lot of such arrays I would like to use the most
> efficient 
> 		method avaiable in numpy. On a pure memory-orientated view
> this would be
> 		just to copy ("move") the memory blocks from the younger 4
> rows one row
> 		further, thus having the first row for the new data.
> 		
> 		In the C API I see some functions like copyswap() and
> memmove() which
> 		indicate that such operations are possible at the C API
> level. But I'm not
> 		sure the correct approach on the Python level.
> 		
> 		Taking slices may be one options - but the new slice will
> then occupy new 
> 		memory, causing memory fragmentation...
> 		Looping over all data items, all rows is time consuming and
> surely wasting
> 		resources...
> 		
> 		Any pointers/ideas?
> 		
> 		Thanks,
> 		Andrew
> 		
> 		
> 		_______________________________________________ 
> 		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 lev at columbia.edu  Wed Jul  4 16:56:25 2007
From: lev at columbia.edu (Lev Givon)
Date: Wed, 4 Jul 2007 16:56:25 -0400
Subject: [SciPy-user] scipy code contribution question
Message-ID: <20070704205625.GC28755@localhost.ee.columbia.edu>

I recently had occasion to translate one of the functions in Matlab's
signal processing toolkit not currently available in scipy into Python
for my own nefarious purposes (i.e., reworking some extant Matlab code
in Python). Might there be any copyright obstacles to my contributing
it to scipy? I personally suspect not because

- Mathworks' copyright specifically protects its Matlab code and
  cannot constrain one from reimplementing non-patented and/or publicly  
  available algorithms in other languages;
- my reimplementation is not an exact translation of said Matlab code;
  and
- the code of several scipy functions seems to have been directly
  inspired by that of corresponding functions included in Matlab.

Please feel free to disabuse me of any misconceptions I may have
expressed above.

							L.G.


From edschofield at gmail.com  Wed Jul  4 18:33:19 2007
From: edschofield at gmail.com (Ed Schofield)
Date: Wed, 4 Jul 2007 23:33:19 +0100
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <20070704205625.GC28755@localhost.ee.columbia.edu>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>
Message-ID: <1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>

On 7/4/07, Lev Givon <lev at columbia.edu> wrote:
> I recently had occasion to translate one of the functions in Matlab's
> signal processing toolkit not currently available in scipy into Python
> for my own nefarious purposes (i.e., reworking some extant Matlab code
> in Python). Might there be any copyright obstacles to my contributing
> it to scipy? I personally suspect not because
>
> - Mathworks' copyright specifically protects its Matlab code and
>   cannot constrain one from reimplementing non-patented and/or publicly
>   available algorithms in other languages;
> - my reimplementation is not an exact translation of said Matlab code;
>   and
> - the code of several scipy functions seems to have been directly
>   inspired by that of corresponding functions included in Matlab.


I think a translation of software into a different language would be
considered a "derivative work":

http://www.copyright.gov/circs/circ14.html
http://ec.europa.eu/internal_market/copyright/documents/documents_en.htm

This definition applies to a motion picture based on a play, so it
would also cover a less-than-exact translation. If several scipy
functions were "directly inspired by" functions from Matlab (or
Numerical Recipes, or other sources), this would pose us no problem --
but if there are any that look like they might be "derivative works",
and therefore infringing, let us know so we can root them out. I think
that we should err on the side of caution, with a zero-tolerance
policy for contaminated code.

Back to your original question, then: I'd suggest that you post some
documentation here about what steps the code takes, using your
judgment about how detailed to make it, and ask for a volunteer to
write a clean-room implementation. Sorry that this requires more work,
but I think it's important for the legitimacy of the project.

-- Ed


From lev at columbia.edu  Wed Jul  4 20:14:21 2007
From: lev at columbia.edu (Lev Givon)
Date: Wed, 4 Jul 2007 20:14:21 -0400
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>
	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>
Message-ID: <20070705001420.GB30029@localhost.ee.columbia.edu>

Received from Ed Schofield on Wed, Jul 04, 2007 at 06:33:19PM EDT:
> On 7/4/07, Lev Givon <lev at columbia.edu> wrote:
> > I recently had occasion to translate one of the functions in Matlab's
> > signal processing toolkit not currently available in scipy into Python
> > for my own nefarious purposes (i.e., reworking some extant Matlab code
> > in Python). Might there be any copyright obstacles to my contributing
> > it to scipy? I personally suspect not because
> >
> > - Mathworks' copyright specifically protects its Matlab code and
> >   cannot constrain one from reimplementing non-patented and/or publicly
> >   available algorithms in other languages;
> > - my reimplementation is not an exact translation of said Matlab code;
> >   and
> > - the code of several scipy functions seems to have been directly
> >   inspired by that of corresponding functions included in Matlab.
> 
> 
> I think a translation of software into a different language would be
> considered a "derivative work":
> 
> http://www.copyright.gov/circs/circ14.html
> http://ec.europa.eu/internal_market/copyright/documents/documents_en.htm
> 
> This definition applies to a motion picture based on a play, so it
> would also cover a less-than-exact translation. If several scipy
> functions were "directly inspired by" functions from Matlab (or
> Numerical Recipes, or other sources), this would pose us no problem --
> but if there are any that look like they might be "derivative works",
> and therefore infringing, let us know so we can root them out. I think
> that we should err on the side of caution, with a zero-tolerance
> policy for contaminated code.
> 
> Back to your original question, then: I'd suggest that you post some
> documentation here about what steps the code takes, using your
> judgment about how detailed to make it, and ask for a volunteer to
> write a clean-room implementation. Sorry that this requires more work,
> but I think it's important for the legitimacy of the project.
> 
> -- Ed

Perhaps I am reading too deeply into the code, but it seems that a few
(but not all) of the scipy filter design functions (e.g., buttord,
cheb1ord) bear more than a passing similarity to their Matlab
counterparts insofar as their internal structure or logical flow are
concerned. (On the other hand, perhaps the published descriptions of
the filter design techniques in question are sufficiently explicit as
to lead to structural similarity in most independent implementations.)

Insofar as the item that prompted my first message, I desire a filter
design function similar to others currently available in scipy to
facilitate the construction of FIR filters with the Remez algorithm
(i.e., something similar to the remezord or firpmord functions in
Matlab) given some typical design parameters (e.g., band cutoff
frequencies, ripple constraints, etc.). As the relevant algorithms
needed by such a filter design function are not overly complicated, I
could look them up an appropriate DSP text or paper and try
implementing them completely from scratch (unless some generous soul
reading this list has already done so :-)

   	       	   	    		     	  L.G.


From david at ar.media.kyoto-u.ac.jp  Thu Jul  5 01:33:33 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Thu, 05 Jul 2007 14:33:33 +0900
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <20070705001420.GB30029@localhost.ee.columbia.edu>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>
	<20070705001420.GB30029@localhost.ee.columbia.edu>
Message-ID: <468C82AD.9090104@ar.media.kyoto-u.ac.jp>

Lev Givon wrote:
>
> Perhaps I am reading too deeply into the code, but it seems that a few
> (but not all) of the scipy filter design functions (e.g., buttord,
> cheb1ord) bear more than a passing similarity to their Matlab
> counterparts insofar as their internal structure or logical flow are
> concerned. (On the other hand, perhaps the published descriptions of
> the filter design techniques in question are sufficiently explicit as
> to lead to structural similarity in most independent implementations.)
As everybody here I guess, IANAL, but the safest thing to do is to 
explicitely ask the mathworks. What is derivative work is a gray area, 
and there is no way to be sure before going to court; implementing 
something by reading source code is certainly not a good idea, and is 
explictely forbidden in most open source projects which "copy" known 
softwares (ReactOS and Wine for windows, samba, etc...); reverse 
engineering is more or less acceptable, but not always, and not in the 
same way in every country.
>
> Insofar as the item that prompted my first message, I desire a filter
> design function similar to others currently available in scipy to
> facilitate the construction of FIR filters with the Remez algorithm
> (i.e., something similar to the remezord or firpmord functions in
> Matlab) given some typical design parameters (e.g., band cutoff
> frequencies, ripple constraints, etc.). As the relevant algorithms
> needed by such a filter design function are not overly complicated, I
> could look them up an appropriate DSP text or paper and try
> implementing them completely from scratch (unless some generous soul
> reading this list has already done so :-)
>
That would be the best thing to do. When you reimplement something for 
scipy, but coming from open source software, you can also ask with the 
copyright holders if this is Ok to copy some parts (eg rereleasing them 
under BSD; that way, it can be incorporated in scipy, and the original 
authors can still use GPL). Even if there is a strong focus on the law 
in the licenses, open source is really about what's considered fair and 
not fair, and at least in my experience, I had no difficulty in taking 
some part of matlab scripts under the GPL, adapt them for scipy and put 
them under the BSD (eg getting the explicit consent of the GPL code to 
release the parts I used under the BSD).

David


From wbaxter at gmail.com  Thu Jul  5 02:18:32 2007
From: wbaxter at gmail.com (Bill Baxter)
Date: Thu, 5 Jul 2007 15:18:32 +0900
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <468C82AD.9090104@ar.media.kyoto-u.ac.jp>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>
	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>
	<20070705001420.GB30029@localhost.ee.columbia.edu>
	<468C82AD.9090104@ar.media.kyoto-u.ac.jp>
Message-ID: <e86a5fd00707042318l440d4f21qc2ed7358f1ee077c@mail.gmail.com>

On 7/5/07, David Cournapeau <david at ar.media.kyoto-u.ac.jp> wrote:
>
> Lev Givon wrote:
> >
> As everybody here I guess, IANAL, but the safest thing to do is to
> explicitely ask the mathworks.


The license I got with my copy of Matlab is pretty clear:

"""
Except as expressly provided by this Agreement, including the
attached Addendum, Licensee may not adapt, translate, or convert
"M-files", "MDL-files" or "P-code" contained in the Programs in
order to create software, a principal purpose of which is to
perform the same or similar functions as Programs licensed by
MathWorks or which is intended to replace any component of the
Programs.  The Licensee may not incorporate or use "M-files",
"P-code", source code, or any other part of the Programs in or
as part of another computer program without the consent of
MathWorks.
"""

You can contact them and try, but I don't expect they'd be willing to give
you explicit consent to translate their M-files in this case, since is
precisely for the purpose of creating software that does the same thing as
Matlab.  And even worse it's for creating *free* software that does the same
thing as Matlab.

On the other hand, if Octave has an implementation of this particular
function you're after, then you are much more likely to get the developers
permission to translate it to Python and put it under BSD.

--bb
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070705/bb24a085/attachment.html>

From wbaxter at gmail.com  Thu Jul  5 02:24:12 2007
From: wbaxter at gmail.com (Bill Baxter)
Date: Thu, 5 Jul 2007 15:24:12 +0900
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <e86a5fd00707042318l440d4f21qc2ed7358f1ee077c@mail.gmail.com>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>
	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>
	<20070705001420.GB30029@localhost.ee.columbia.edu>
	<468C82AD.9090104@ar.media.kyoto-u.ac.jp>
	<e86a5fd00707042318l440d4f21qc2ed7358f1ee077c@mail.gmail.com>
Message-ID: <e86a5fd00707042324pf0163ffq7f35f53bdf02949c@mail.gmail.com>

On 7/5/07, Bill Baxter <wbaxter at gmail.com> wrote:
>
> On 7/5/07, David Cournapeau <david at ar.media.kyoto-u.ac.jp> wrote:
> >
> > Lev Givon wrote:
> > >
> > As everybody here I guess, IANAL, but the safest thing to do is to
> > explicitely ask the mathworks.
>
>
> The license I got with my copy of Matlab is pretty clear:
>
> """
> Except as expressly provided by this Agreement, including the
> attached Addendum, Licensee may not adapt, translate, or convert
> "M-files", "MDL-files" or "P-code" contained in the Programs in
> order to create software, a principal purpose of which is to
> perform the same or similar functions as Programs licensed by
> MathWorks or which is intended to replace any component of the
> Programs.  The Licensee may not incorporate or use "M-files",
> "P-code", source code, or any other part of the Programs in or
> as part of another computer program without the consent of
> MathWorks.
> """
>
> You can contact them and try, but I don't expect they'd be willing to give
> you explicit consent to translate their M-files in this case, since is
> precisely for the purpose of creating software that does the same thing as
> Matlab.  And even worse it's for creating *free* software that does the same
> thing as Matlab.
>
> On the other hand, if Octave has an implementation of this particular
> function you're after, then you are much more likely to get the developers
> permission to translate it to Python and put it under BSD.
>

And here appears to be just such a beast:

http://velveeta.che.wisc.edu/octave/lists/archive//octave-sources.2001/msg00011.html


---bb
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070705/45b38793/attachment.html>

From icy.flame.gm at gmail.com  Thu Jul  5 03:36:24 2007
From: icy.flame.gm at gmail.com (iCy-fLaME)
Date: Thu, 5 Jul 2007 08:36:24 +0100
Subject: [SciPy-user] How does concatenate work?
Message-ID: <bfb379e10707050036s2ef49330h7301c6fcb2affea5@mail.gmail.com>

If I have a (m,  n) array, is concatenate is right way to go about
making it a (m+1, n) or (m, n+1) array?

I tried the following, couldn't figure out how to do it correctly. Any
helps are greatly appreciated.

>>> from numpy import linspace, zeros
>>> from numpy import concatenate
>>>
>>> a = linspace(1, 20, 20).reshape((5,4))
>>> a
array([[  1.,   2.,   3.,   4.],
       [  5.,   6.,   7.,   8.],
       [  9.,  10.,  11.,  12.],
       [ 13.,  14.,  15.,  16.],
       [ 17.,  18.,  19.,  20.]])
>>>
>>> b = zeros(4)
>>> b
array([ 0.,  0.,  0.,  0.])
>>>
>>> concatenate((a,b), axis=0)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
ValueError: arrays must have same number of dimensions
>>>
>>> concatenate((a,b), axis=1)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
ValueError: arrays must have same number of dimensions





icy


From openopt at ukr.net  Thu Jul  5 03:43:09 2007
From: openopt at ukr.net (dmitrey)
Date: Thu, 05 Jul 2007 10:43:09 +0300
Subject: [SciPy-user] How does concatenate work?
In-Reply-To: <bfb379e10707050036s2ef49330h7301c6fcb2affea5@mail.gmail.com>
References: <bfb379e10707050036s2ef49330h7301c6fcb2affea5@mail.gmail.com>
Message-ID: <468CA10D.3010608@ukr.net>

try
b = zeros((1,4))

concatenate((a,b), 0)
array([[  1.,   2.,   3.,   4.],
       [  5.,   6.,   7.,   8.],
       [  9.,  10.,  11.,  12.],
       [ 13.,  14.,  15.,  16.],
       [ 17.,  18.,  19.,  20.],
       [  0.,   0.,   0.,   0.]])
you may use also
concatenate((a,b.reshape(1,4)), 0)
or concatenate((a,b.reshape(1,4)))
or numpy.hstack((a,b.reshape(1,4)))
HTH,D

iCy-fLaME wrote:
> If I have a (m,  n) array, is concatenate is right way to go about
> making it a (m+1, n) or (m, n+1) array?
>
> I tried the following, couldn't figure out how to do it correctly. Any
> helps are greatly appreciated.
>
>   
>>>> from numpy import linspace, zeros
>>>> from numpy import concatenate
>>>>
>>>> a = linspace(1, 20, 20).reshape((5,4))
>>>> a
>>>>         
> array([[  1.,   2.,   3.,   4.],
>        [  5.,   6.,   7.,   8.],
>        [  9.,  10.,  11.,  12.],
>        [ 13.,  14.,  15.,  16.],
>        [ 17.,  18.,  19.,  20.]])
>   
>>>> b = zeros(4)
>>>> b
>>>>         
> array([ 0.,  0.,  0.,  0.])
>   
>>>> concatenate((a,b), axis=0)
>>>>         
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> ValueError: arrays must have same number of dimensions
>   
>>>> concatenate((a,b), axis=1)
>>>>         
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> ValueError: arrays must have same number of dimensions
>
>
>
>
>
> icy
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>
>   



From icy.flame.gm at gmail.com  Thu Jul  5 03:55:30 2007
From: icy.flame.gm at gmail.com (iCy-fLaME)
Date: Thu, 5 Jul 2007 08:55:30 +0100
Subject: [SciPy-user] How does concatenate work?
In-Reply-To: <468CA10D.3010608@ukr.net>
References: <bfb379e10707050036s2ef49330h7301c6fcb2affea5@mail.gmail.com>
	<468CA10D.3010608@ukr.net>
Message-ID: <bfb379e10707050055x20040549o1c35b641aeea2641@mail.gmail.com>

Aha.. that works. I see they both have to have the same number of
dimensions for it to work, thanks a lot dmitrey.

On 7/5/07, dmitrey <openopt at ukr.net> wrote:
> try
> b = zeros((1,4))
>
> concatenate((a,b), 0)
> array([[  1.,   2.,   3.,   4.],
>        [  5.,   6.,   7.,   8.],
>        [  9.,  10.,  11.,  12.],
>        [ 13.,  14.,  15.,  16.],
>        [ 17.,  18.,  19.,  20.],
>        [  0.,   0.,   0.,   0.]])
> you may use also
> concatenate((a,b.reshape(1,4)), 0)
> or concatenate((a,b.reshape(1,4)))
> or numpy.hstack((a,b.reshape(1,4)))
> HTH,D
>
> iCy-fLaME wrote:
> > If I have a (m,  n) array, is concatenate is right way to go about
> > making it a (m+1, n) or (m, n+1) array?
> >
> > I tried the following, couldn't figure out how to do it correctly. Any
> > helps are greatly appreciated.
> >
> >
> >>>> from numpy import linspace, zeros
> >>>> from numpy import concatenate
> >>>>
> >>>> a = linspace(1, 20, 20).reshape((5,4))
> >>>> a
> >>>>
> > array([[  1.,   2.,   3.,   4.],
> >        [  5.,   6.,   7.,   8.],
> >        [  9.,  10.,  11.,  12.],
> >        [ 13.,  14.,  15.,  16.],
> >        [ 17.,  18.,  19.,  20.]])
> >
> >>>> b = zeros(4)
> >>>> b
> >>>>
> > array([ 0.,  0.,  0.,  0.])
> >
> >>>> concatenate((a,b), axis=0)
> >>>>
> > Traceback (most recent call last):
> >   File "<stdin>", line 1, in ?
> > ValueError: arrays must have same number of dimensions
> >
> >>>> concatenate((a,b), axis=1)
> >>>>
> > Traceback (most recent call last):
> >   File "<stdin>", line 1, in ?
> > ValueError: arrays must have same number of dimensions
> >
> >
> >
> >
> >
> > icy
> > _______________________________________________
> > 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
>


-- 
iCy-fLaME
------------------------------------------------
The body maybe wounded, but it is the mind that hurts.


From ryanlists at gmail.com  Thu Jul  5 08:36:28 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Thu, 5 Jul 2007 07:36:28 -0500
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <e86a5fd00707042324pf0163ffq7f35f53bdf02949c@mail.gmail.com>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>
	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>
	<20070705001420.GB30029@localhost.ee.columbia.edu>
	<468C82AD.9090104@ar.media.kyoto-u.ac.jp>
	<e86a5fd00707042318l440d4f21qc2ed7358f1ee077c@mail.gmail.com>
	<e86a5fd00707042324pf0163ffq7f35f53bdf02949c@mail.gmail.com>
Message-ID: <c5b438120707050536v7f863ee4jcc1b914cf4e7ad73@mail.gmail.com>

I certainly don't claim to be an expert on derivative work, but it
seems like you aren't creating a derivative work if Mathworks never
showed you the source code - which I assume they didn't.

It seems like an overly literal interpretation of the license snippet
that Bill posted would say that if I ever used Matlab, then I can
never write code that does the same things as Matlab or any of its
toolboxes.  If that is legally defended, then I can't contribute
anything to scipy.

I appreciate that we want to be careful, but it doesn't make sense
that we could get in trouble by creating functions with the same names
as Matlab functions if we've never seen their source code.  I mean,
can they copyright plot(x,y)?  If so, than matplotlib is in trouble
(unless John had never used Matlab when he came up with that syntax).
What about fft(x)?

The link that Bill posted with the GPL code poses as big a problem in
my mind.  If any of us looks at that code, then any similar algorithm
would need to be GPL'ed as a derivative work.  So, don't open that
link :)

FWIW,

Ryan

On 7/5/07, Bill Baxter <wbaxter at gmail.com> wrote:
>
>
>
> On 7/5/07, Bill Baxter <wbaxter at gmail.com> wrote:
> > On 7/5/07, David Cournapeau <david at ar.media.kyoto-u.ac.jp> wrote:
> >
> > > Lev Givon wrote:
> > > >
> > > As everybody here I guess, IANAL, but the safest thing to do is to
> > > explicitely ask the mathworks.
> >
> >
> > The license I got with my copy of Matlab is pretty clear:
> >
> > """
> > Except as expressly provided by this Agreement, including the
> > attached Addendum, Licensee may not adapt, translate, or convert
> > "M-files", "MDL-files" or "P-code" contained in the Programs in
> > order to create software, a principal purpose of which is to
> > perform the same or similar functions as Programs licensed by
> > MathWorks or which is intended to replace any component of the
> > Programs.  The Licensee may not incorporate or use "M-files",
> > "P-code", source code, or any other part of the Programs in or
> > as part of another computer program without the consent of
> > MathWorks.
> > """
> >
> > You can contact them and try, but I don't expect they'd be willing to give
> you explicit consent to translate their M-files in this case, since is
> precisely for the purpose of creating software that does the same thing as
> Matlab.  And even worse it's for creating *free* software that does the same
> thing as Matlab.
> >
> > On the other hand, if Octave has an implementation of this particular
> function you're after, then you are much more likely to get the developers
> permission to translate it to Python and put it under BSD.
>
> And here appears to be just such a beast:
>
> http://velveeta.che.wisc.edu/octave/lists/archive//octave-sources.2001/msg00011.html
>
>
> ---bb
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>


From giovanni.samaey at cs.kuleuven.be  Thu Jul  5 09:25:01 2007
From: giovanni.samaey at cs.kuleuven.be (Giovanni Samaey)
Date: Thu, 5 Jul 2007 15:25:01 +0200
Subject: [SciPy-user] installation problems on Mac OS X
Message-ID: <13A99D27-BD5E-4D53-86C1-770507B6C1BF@cs.kuleuven.be>

Hi all,

I am trying to install scipy on my shiny new macbook:
I followed instructions as on www.scipy.org/Installing_SciPy/Mac_OS_X/.
Installation of vecLib was done when the computer shipped (I found a  
vecLib).  The link to the vecLib on the instruction page is dead and  
I couldn't find a download on apple.com, so I don't know if that  
installation is complete, as the instructions ask.

Installation goes smoothly, byt scipy.test() gives 2 errors.  I tried  
both  python 2.4.4 and python 2.5.1; no difference. I have the most  
recent gfortran and I use gcc 4.0.   Scipy is taken from svn.

======================================================================
FAIL: check_dot (scipy.lib.tests.test_blas.test_fblas1_simple)
----------------------------------------------------------------------
Traceback (most recent call last):
   File "/Library/Frameworks/Python.framework/Versions/2.4/lib/ 
python2.4/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 "/Library/Frameworks/Python.framework/Versions/2.4/lib/ 
python2.4/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: 3.4625784615075479e-37j
  DESIRED: (-9+2j)

======================================================================
FAIL: check_dot (scipy.linalg.tests.test_blas.test_fblas1_simple)
----------------------------------------------------------------------
Traceback (most recent call last):
   File "/Library/Frameworks/Python.framework/Versions/2.4/lib/ 
python2.4/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 "/Library/Frameworks/Python.framework/Versions/2.4/lib/ 
python2.4/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: 3.4782945282523978e-37j
  DESIRED: (-9+2j)

----------------------------------------------------------------------
Ran 1632 tests in 3.592s

FAILED (failures=2)




From lev at columbia.edu  Thu Jul  5 09:55:18 2007
From: lev at columbia.edu (Lev Givon)
Date: Thu, 5 Jul 2007 09:55:18 -0400
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <c5b438120707050536v7f863ee4jcc1b914cf4e7ad73@mail.gmail.com>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>
	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>
	<20070705001420.GB30029@localhost.ee.columbia.edu>
	<468C82AD.9090104@ar.media.kyoto-u.ac.jp>
	<e86a5fd00707042318l440d4f21qc2ed7358f1ee077c@mail.gmail.com>
	<e86a5fd00707042324pf0163ffq7f35f53bdf02949c@mail.gmail.com>
	<c5b438120707050536v7f863ee4jcc1b914cf4e7ad73@mail.gmail.com>
Message-ID: <20070705135518.GD30029@localhost.ee.columbia.edu>

Received from Ryan Krauss on Thu, Jul 05, 2007 at 08:36:28AM EDT:
> I certainly don't claim to be an expert on derivative work, but it
> seems like you aren't creating a derivative work if Mathworks never
> showed you the source code - which I assume they didn't.
> 
> It seems like an overly literal interpretation of the license snippet
> that Bill posted would say that if I ever used Matlab, then I can
> never write code that does the same things as Matlab or any of its
> toolboxes.  If that is legally defended, then I can't contribute
> anything to scipy.
> 
> I appreciate that we want to be careful, but it doesn't make sense
> that we could get in trouble by creating functions with the same names
> as Matlab functions if we've never seen their source code.  I mean,
> can they copyright plot(x,y)?  If so, than matplotlib is in trouble
> (unless John had never used Matlab when he came up with that syntax).
> What about fft(x)?
> 

I agree with your view regarding the implementation of Matlab-like
interfaces; in my previous messages, I was specifically alluding to
functions included in Matlab whose source is visible (i.e., most of
the toolkit functions).

> The link that Bill posted with the GPL code poses as big a problem in
> my mind.  If any of us looks at that code, then any similar algorithm
> would need to be GPL'ed as a derivative work.  So, don't open that
> link :)
> 

No need to worry; one side effect of prolonged scipy use is that
scientific C code begins to take on a marked resemblance to Greek (at
least in the eyes of barbaroi such as myself :-)

						L.G.


From ryanlists at gmail.com  Thu Jul  5 10:15:39 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Thu, 5 Jul 2007 09:15:39 -0500
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <20070705135518.GD30029@localhost.ee.columbia.edu>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>
	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>
	<20070705001420.GB30029@localhost.ee.columbia.edu>
	<468C82AD.9090104@ar.media.kyoto-u.ac.jp>
	<e86a5fd00707042318l440d4f21qc2ed7358f1ee077c@mail.gmail.com>
	<e86a5fd00707042324pf0163ffq7f35f53bdf02949c@mail.gmail.com>
	<c5b438120707050536v7f863ee4jcc1b914cf4e7ad73@mail.gmail.com>
	<20070705135518.GD30029@localhost.ee.columbia.edu>
Message-ID: <c5b438120707050715oe6dbef6i88212fa479cfda5f@mail.gmail.com>

Thanks for the clarification.  If you had access to their source, then
I don't see how you could contribute this to scipy.  I can't imagine
them giving you permision.

On 7/5/07, Lev Givon <lev at columbia.edu> wrote:
> Received from Ryan Krauss on Thu, Jul 05, 2007 at 08:36:28AM EDT:
> > I certainly don't claim to be an expert on derivative work, but it
> > seems like you aren't creating a derivative work if Mathworks never
> > showed you the source code - which I assume they didn't.
> >
> > It seems like an overly literal interpretation of the license snippet
> > that Bill posted would say that if I ever used Matlab, then I can
> > never write code that does the same things as Matlab or any of its
> > toolboxes.  If that is legally defended, then I can't contribute
> > anything to scipy.
> >
> > I appreciate that we want to be careful, but it doesn't make sense
> > that we could get in trouble by creating functions with the same names
> > as Matlab functions if we've never seen their source code.  I mean,
> > can they copyright plot(x,y)?  If so, than matplotlib is in trouble
> > (unless John had never used Matlab when he came up with that syntax).
> > What about fft(x)?
> >
>
> I agree with your view regarding the implementation of Matlab-like
> interfaces; in my previous messages, I was specifically alluding to
> functions included in Matlab whose source is visible (i.e., most of
> the toolkit functions).
>
> > The link that Bill posted with the GPL code poses as big a problem in
> > my mind.  If any of us looks at that code, then any similar algorithm
> > would need to be GPL'ed as a derivative work.  So, don't open that
> > link :)
> >
>
> No need to worry; one side effect of prolonged scipy use is that
> scientific C code begins to take on a marked resemblance to Greek (at
> least in the eyes of barbaroi such as myself :-)
>
>                                                 L.G.
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From barrywark at gmail.com  Thu Jul  5 14:57:57 2007
From: barrywark at gmail.com (Barry Wark)
Date: Thu, 5 Jul 2007 11:57:57 -0700
Subject: [SciPy-user] Compile fails for r3123 on OS X 10.4 (Intel)
In-Reply-To: <20070703233637.GA9394@arbutus.physics.mcmaster.ca>
References: <cd7634ce0706291033ga50ed5bj1e0d7537b09bb346@mail.gmail.com>
	<20070703200423.GA5477@arbutus.physics.mcmaster.ca>
	<cd7634ce0707031314u1f2827c3jabf4bde824eae543@mail.gmail.com>
	<20070703233637.GA9394@arbutus.physics.mcmaster.ca>
Message-ID: <cd7634ce0707051157g2992e2b4i2426d2ec5b401f71@mail.gmail.com>

OK. This is a weird one for me...

numpy r3881 builds fine. Scipy r3145 (a fresh checkout) does not build
still. It looks like the scipy setup/config can't find many of the
sources in sub-packages. Most relevant may be these lines from the
build log:

non-existing path in 'scipy/fftpack': 'tests'
could not resolve pattern in 'scipy/fftpack': 'dfftpack/*.f'
non-existing path in 'scipy/fftpack': 'fftpack.pyf'
non-existing path in 'scipy/fftpack': 'src/zfft.c'
non-existing path in 'scipy/fftpack': 'src/drfft.c'
non-existing path in 'scipy/fftpack': 'src/zrfft.c'
non-existing path in 'scipy/fftpack': 'src/zfftnd.c'
non-existing path in 'scipy/fftpack': 'src/zfft_djbfft.c'
non-existing path in 'scipy/fftpack': 'src/zfft_fftpack.c'
non-existing path in 'scipy/fftpack': 'src/zfft_fftw.c'
non-existing path in 'scipy/fftpack': 'src/zfft_fftw3.c'
non-existing path in 'scipy/fftpack': 'src/zfft_mkl.c'
non-existing path in 'scipy/fftpack': 'convolve.pyf'
non-existing path in 'scipy/fftpack': 'src/convolve.c'

Similar lines for many other packages are printed as well. Obviously
none of these lines appear on a machine that does build scipy
correctly.

Can any of the numpy.distutils gurus point me in a helpful direction?

Thanks!
Barry

On 7/3/07, David M. Cooke <cookedm at physics.mcmaster.ca> wrote:
> On Tue, Jul 03, 2007 at 01:14:12PM -0700, Barry Wark wrote:
> > Yes, I'm using a clean checkout from SVN. Just to double check against
> > your setup, I'm using fftw-3 installed via macports as my fftw library
> >
> > Barry
>
> Still don't have your problem. Scipy is r3143, numpy is r3880. OS X
> 10.4.10 (Intel MacBook), Python 2.5 and 2.4. I'm using gfortran from
> macports, but that shouldn't affect this.
>
> One thing I can think of is checking your numpy install, as the
> _fortranmodule.c is generated by f2py.
>
> (in general, the latest svn scipy usually needs a recent svn numpy)
>
> > On 7/3/07, David M. Cooke <cookedm at physics.mcmaster.ca> wrote:
> > > On Fri, Jun 29, 2007 at 10:33:24AM -0700, Barry Wark wrote:
> > > > I've been unable to compile scipy on OS X 10.4 (intel) from the recent
> > > > trunk. Scipy built on this machine as of r2708. The output from
> > > >
> > > > python setup.py build
> > > >
> > > > is attached.
> > > >
> > > > I have fftw3 installed via macports (in /opt/local), and it appears
> > > > that the build finds it properly, but the build fails with an error:
> > > >
> > > > building extension "scipy.fftpack._fftpack" sources
> > > >  target build/src.macosx-10.3-fat-2.5/_fftpackmodule.c does not exist:
> > > >   Assuming _fftpackmodule.c was generated with "build_src --inplace"
> > > >   command.
> > > > error: '_fftpackmodule.c' missing
> > > >
> > > > I would appreciate any advice or suggestions!
> > >
> > > I've got pretty much the same setup, and no problems. Did you delete the
> > > build directory first (rm -rf build)?
> > >
> > > --
> > > |>|\/|<
> > > /--------------------------------------------------------------------------\
> > > |David M. Cooke                      http://arbutus.physics.mcmaster.ca/dmc/
> > > |cookedm at physics.mcmaster.ca
> > > _______________________________________________
> > > 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
> >
>
> --
> |>|\/|<
> /--------------------------------------------------------------------------\
> |David M. Cooke                      http://arbutus.physics.mcmaster.ca/dmc/
> |cookedm at physics.mcmaster.ca
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From fredmfp at gmail.com  Thu Jul  5 16:15:20 2007
From: fredmfp at gmail.com (fred)
Date: Thu, 05 Jul 2007 22:15:20 +0200
Subject: [SciPy-user] [f2py] f90 format file...
Message-ID: <468D5158.6070304@gmail.com>

Hi,

I wrote my fortran code in F90 format (I guess called "free" ?),
and it seems that f2py can understand it.

But how can I enable it ?

If I wrote subroutine foo() at column 0,
f2py complains.

Any clue ?

TIA.

Cheers,

-- 
http://scipy.org/FredericPetit



From fredmfp at gmail.com  Thu Jul  5 16:29:13 2007
From: fredmfp at gmail.com (fred)
Date: Thu, 05 Jul 2007 22:29:13 +0200
Subject: [SciPy-user] [f2py] f90 format file...
In-Reply-To: <468D5158.6070304@gmail.com>
References: <468D5158.6070304@gmail.com>
Message-ID: <468D5499.3030200@gmail.com>

Hmm, renaming file from *.f to *.f90 seems to fix it.

Sorry.

Cheers,

-- 
http://scipy.org/FredericPetit



From steveire at gmail.com  Thu Jul  5 17:52:53 2007
From: steveire at gmail.com (Stephen Kelly)
Date: Thu, 05 Jul 2007 22:52:53 +0100
Subject: [SciPy-user] Fraction representation
Message-ID: <468d6837.1401420a.34d7.0973@mx.google.com>

Hi. I wonder if it is possible to add a fraction representation.

My casio calculator has a feature to display a real number as a fraction
(eg, 0.1 as 1,10. The delimiter is not a comma though). Would it be
possible to add such a feature to scipy?

Thanks

Steve.



From fredmfp at gmail.com  Thu Jul  5 18:35:41 2007
From: fredmfp at gmail.com (fred)
Date: Fri, 06 Jul 2007 00:35:41 +0200
Subject: [SciPy-user] [f2py] f90 format file...
In-Reply-To: <468D5499.3030200@gmail.com>
References: <468D5158.6070304@gmail.com> <468D5499.3030200@gmail.com>
Message-ID: <468D723D.1010900@gmail.com>

fred a ?crit :
> Hmm, renaming file from *.f to *.f90 seems to fix it.
>
I still have issue with this, in fact.

I use intel fortran compiler.

If fortran file are named *.f, my code works really fine.
But I don't like the "fixed format", I prefer the "free" one.
So I renamed my fortran files to *.f90, and indented the lines as I like.
But in this case, my code does not work any more.

Typically, I use in my python code something like:

KM = calc_KM(KM,...)

With *.f files, KM is the right matrix.
With *.f90 files, KM is None.


Any clue ?

TIA.


Cheers,

-- 
http://scipy.org/FredericPetit



From labbuhl at hotmail.com  Thu Jul  5 19:32:06 2007
From: labbuhl at hotmail.com (Lee Abbuhl)
Date: Thu, 05 Jul 2007 19:32:06 -0400
Subject: [SciPy-user] [f2py] f90 format file...
In-Reply-To: <468D723D.1010900@gmail.com>
Message-ID: <BAY105-F29B5C39DD37D83FF4FF518B3020@phx.gbl>




>From: fred <fredmfp 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] [f2py] f90 format file...
>Date: Fri, 06 Jul 2007 00:35:41 +0200
>
>fred a ?crit :
> > Hmm, renaming file from *.f to *.f90 seems to fix it.
> >
>I still have issue with this, in fact.
>
>I use intel fortran compiler.
>
>If fortran file are named *.f, my code works really fine.
>But I don't like the "fixed format", I prefer the "free" one.
>So I renamed my fortran files to *.f90, and indented the lines as I like.
>But in this case, my code does not work any more.
>
>Typically, I use in my python code something like:
>
>KM = calc_KM(KM,...)
>
>With *.f files, KM is the right matrix.
>With *.f90 files, KM is None.
>
>
>Any clue ?
>
>TIA.
>
>
>Cheers,
>
>--
>http://scipy.org/FredericPetit
>
>_______________________________________________
>SciPy-user mailing list
>SciPy-user at scipy.org
>http://projects.scipy.org/mailman/listinfo/scipy-user

_________________________________________________________________
Don't get caught with egg on your face. Play Chicktionary!? 
http://club.live.com/chicktionary.aspx?icid=chick_hotmailtextlink2



From lbolla at gmail.com  Fri Jul  6 02:53:49 2007
From: lbolla at gmail.com (lorenzo bolla)
Date: Fri, 6 Jul 2007 08:53:49 +0200
Subject: [SciPy-user] [f2py] f90 format file...
In-Reply-To: <468D723D.1010900@gmail.com>
References: <468D5158.6070304@gmail.com> <468D5499.3030200@gmail.com>
	<468D723D.1010900@gmail.com>
Message-ID: <80c99e790707052353s6d68e966h5ba3a181bcc07a38@mail.gmail.com>

I use f2py with .f90 files and the intel compiler, and it works fine for
me... can you submit a piece of .f90 code?
by the way, I specify the intel compiler by this command line:
f2py -c <file.f90> -m <module> --fcompiler=intele --compiler=intel

lorenzo.




On 7/6/07, fred <fredmfp at gmail.com> wrote:
>
> fred a ?crit :
> > Hmm, renaming file from *.f to *.f90 seems to fix it.
> >
> I still have issue with this, in fact.
>
> I use intel fortran compiler.
>
> If fortran file are named *.f, my code works really fine.
> But I don't like the "fixed format", I prefer the "free" one.
> So I renamed my fortran files to *.f90, and indented the lines as I like.
> But in this case, my code does not work any more.
>
> Typically, I use in my python code something like:
>
> KM = calc_KM(KM,...)
>
> With *.f files, KM is the right matrix.
> With *.f90 files, KM is None.
>
>
> Any clue ?
>
> TIA.
>
>
> Cheers,
>
> --
> http://scipy.org/FredericPetit
>
> _______________________________________________
> 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/20070706/c27c75ad/attachment.html>

From fredmfp at gmail.com  Fri Jul  6 05:25:46 2007
From: fredmfp at gmail.com (fred)
Date: Fri, 06 Jul 2007 11:25:46 +0200
Subject: [SciPy-user] [f2py] f90 format file...
In-Reply-To: <80c99e790707052353s6d68e966h5ba3a181bcc07a38@mail.gmail.com>
References: <468D5158.6070304@gmail.com>
	<468D5499.3030200@gmail.com>	<468D723D.1010900@gmail.com>
	<80c99e790707052353s6d68e966h5ba3a181bcc07a38@mail.gmail.com>
Message-ID: <468E0A9A.90804@gmail.com>

lorenzo bolla a ?crit :
> I use f2py with .f90 files and the intel compiler, and it works fine 
> for me... can you submit a piece of .f90 code?
> by the way, I specify the intel compiler by this command line:
> f2py -c <file.f90> -m <module> --fcompiler=intele --compiler=intel
No problem.
Please see the attached files.

TIA.

-- 
http://scipy.org/FredericPetit

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: Makefile
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070706/96d542d8/attachment.ksh>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test.py
Type: text/x-python
Size: 318 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070706/96d542d8/attachment.py>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: test.f
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070706/96d542d8/attachment-0001.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: test.f90
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070706/96d542d8/attachment-0002.ksh>

From fredmfp at gmail.com  Fri Jul  6 05:33:52 2007
From: fredmfp at gmail.com (fred)
Date: Fri, 06 Jul 2007 11:33:52 +0200
Subject: [SciPy-user] [f2py] f90 format file...
In-Reply-To: <80c99e790707052353s6d68e966h5ba3a181bcc07a38@mail.gmail.com>
References: <468D5158.6070304@gmail.com>
	<468D5499.3030200@gmail.com>	<468D723D.1010900@gmail.com>
	<80c99e790707052353s6d68e966h5ba3a181bcc07a38@mail.gmail.com>
Message-ID: <468E0C80.4000005@gmail.com>

lorenzo bolla a ?crit :
> I use f2py with .f90 files and the intel compiler, and it works fine 
> for me... can you submit a piece of .f90 code?
> by the way, I specify the intel compiler by this command line:
> f2py -c <file.f90> -m <module> --fcompiler=intele --compiler=intel
PS: I use two debian etch boxes, i686 & x86_64.
I get the same issue on both.

PS2: f2py 2_3649, scipy 0.5.2, numpy 1.0.2
 
Cheers,

-- 
http://scipy.org/FredericPetit



From nwagner at iam.uni-stuttgart.de  Fri Jul  6 05:45:32 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Fri, 06 Jul 2007 11:45:32 +0200
Subject: [SciPy-user] B-orthogonal complement of the null space of A
Message-ID: <468E0F3C.4070502@iam.uni-stuttgart.de>

Hi all,

Consider  two symmetric matrices A and B, where B is positive definite
and A is positive semidefinite.
How can I compute the B-orthogonal complement of the null space of A
with scipy.linalg ?

Any hint would be appreciated.
 
Nils



From lbolla at gmail.com  Fri Jul  6 05:52:47 2007
From: lbolla at gmail.com (lorenzo bolla)
Date: Fri, 6 Jul 2007 11:52:47 +0200
Subject: [SciPy-user] [f2py] f90 format file...
In-Reply-To: <468E0C80.4000005@gmail.com>
References: <468D5158.6070304@gmail.com> <468D5499.3030200@gmail.com>
	<468D723D.1010900@gmail.com>
	<80c99e790707052353s6d68e966h5ba3a181bcc07a38@mail.gmail.com>
	<468E0C80.4000005@gmail.com>
Message-ID: <80c99e790707060252t1094f83cn253d9b9f36965e63@mail.gmail.com>

quite strange. it doesn't happen to me. here is what I get

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

> f2py --fcompiler=intele -c test.f90 -m test
running build
running config_fc
running build_src
building extension "test" sources
f2py options: []
f2py:> /tmp/tmpkQ6dXv/src.linux-ia64-2.5/testmodule.c
creating /tmp/tmpkQ6dXv
creating /tmp/tmpkQ6dXv/src.linux-ia64-2.5
Reading fortran codes...
        Reading file 'test.f90' (format:free)
Post-processing...
        Block: test
                        Block: calc_km
Post-processing (stage 2)...
Building modules...
        Building module "test"...
                Constructing wrapper function "calc_km"...
                  km = calc_km(km)
        Wrote C/API module "test" to file "/tmp/tmpkQ6dXv/src.linux-ia64-2.5
/testmodule.c"
  adding '/tmp/tmpkQ6dXv/src.linux-ia64-2.5/fortranobject.c' to sources.
  adding '/tmp/tmpkQ6dXv/src.linux-ia64-2.5' to include_dirs.
copying
/xlv1/labsoi_devices/bollalo001/lib/python2.5/site-packages/numpy/f2py/src/fortranobject.c
-> /tmp/tmpkQ6dXv/src.linux-ia64-2.5
copying
/xlv1/labsoi_devices/bollalo001/lib/python2.5/site-packages/numpy/f2py/src/fortranobject.h
-> /tmp/tmpkQ6dXv/src.linux-ia64-2.5
running build_ext
customize UnixCCompiler
customize UnixCCompiler using build_ext
Could not locate executable ifc
Could not locate executable efort
Could not locate executable efort
customize IntelItaniumFCompiler
customize IntelItaniumFCompiler using build_ext
building 'test' extension
compiling C sources
C compiler: gcc -pthread -fno-strict-aliasing -DNDEBUG -g -O3 -Wall
-Wstrict-prototypes -fPIC

creating /tmp/tmpkQ6dXv/tmp
creating /tmp/tmpkQ6dXv/tmp/tmpkQ6dXv
creating /tmp/tmpkQ6dXv/tmp/tmpkQ6dXv/src.linux-ia64-2.5
compile options:
'-I/tmp/tmpkQ6dXv/src.linux-ia64-2.5-I/xlv1/labsoi_devices/bollalo001/lib/python2.5/site-packages/numpy/core/include
-I/xlv1/labsoi_devices/bollalo001/include/python2.5 -c'
gcc: /tmp/tmpkQ6dXv/src.linux-ia64-2.5/fortranobject.c
gcc: /tmp/tmpkQ6dXv/src.linux-ia64-2.5/testmodule.c
compiling Fortran sources
Fortran f77 compiler: /opt/intel/fc/bin/ifort -FI -w90 -w95 -KPIC -cm -O3
-unroll
Fortran f90 compiler: /opt/intel/fc/bin/ifort -FR -KPIC -cm -O3 -unroll
Fortran fix compiler: /opt/intel/fc/bin/ifort -FI -KPIC -cm -O3 -unroll
compile options:
'-I/tmp/tmpkQ6dXv/src.linux-ia64-2.5-I/xlv1/labsoi_devices/bollalo001/lib/python2.5/site-packages/numpy/core/include
-I/xlv1/labsoi_devices/bollalo001/include/python2.5 -c'
ifort:f90: test.f90
ifort: Command line warning: ignoring option '-u'; no argument required
/opt/intel/fc/bin/ifort -shared -nofor_main
/tmp/tmpkQ6dXv/tmp/tmpkQ6dXv/src.linux-ia64-2.5/testmodule.o
/tmp/tmpkQ6dXv/tmp/tmpkQ6dXv/src.linux-ia64-2.5/fortranobject.o
/tmp/tmpkQ6dXv/test.o -o ./test.so
Removing build directory /tmp/tmpkQ6dXv


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

what is the exact error message you get?
lorenzo






On 7/6/07, fred <fredmfp at gmail.com> wrote:
>
> lorenzo bolla a ?crit :
> > I use f2py with .f90 files and the intel compiler, and it works fine
> > for me... can you submit a piece of .f90 code?
> > by the way, I specify the intel compiler by this command line:
> > f2py -c <file.f90> -m <module> --fcompiler=intele --compiler=intel
> PS: I use two debian etch boxes, i686 & x86_64.
> I get the same issue on both.
>
> PS2: f2py 2_3649, scipy 0.5.2, numpy 1.0.2
>
> Cheers,
>
> --
> http://scipy.org/FredericPetit
>
> _______________________________________________
> 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/20070706/d4dfc2b8/attachment.html>

From fredmfp at gmail.com  Fri Jul  6 06:02:30 2007
From: fredmfp at gmail.com (fred)
Date: Fri, 06 Jul 2007 12:02:30 +0200
Subject: [SciPy-user] [f2py] f90 format file...
In-Reply-To: <80c99e790707060252t1094f83cn253d9b9f36965e63@mail.gmail.com>
References: <468D5158.6070304@gmail.com>
	<468D5499.3030200@gmail.com>	<468D723D.1010900@gmail.com>	<80c99e790707052353s6d68e966h5ba3a181bcc07a38@mail.gmail.com>	<468E0C80.4000005@gmail.com>
	<80c99e790707060252t1094f83cn253d9b9f36965e63@mail.gmail.com>
Message-ID: <468E1336.5090507@gmail.com>

lorenzo bolla a ?crit :
> quite strange. it doesn't happen to me. here is what I get
>  
I notice you use python 2.5.
I use python 2.4. 
> what is the exact error message you get?
I don't get error message.
With test_f90, I get KM = None.
With test_f, I get the right result.

-- 
http://scipy.org/FredericPetit



From lbolla at gmail.com  Fri Jul  6 06:13:58 2007
From: lbolla at gmail.com (lorenzo bolla)
Date: Fri, 6 Jul 2007 12:13:58 +0200
Subject: [SciPy-user] [f2py] f90 format file...
In-Reply-To: <468E1336.5090507@gmail.com>
References: <468D5158.6070304@gmail.com> <468D5499.3030200@gmail.com>
	<468D723D.1010900@gmail.com>
	<80c99e790707052353s6d68e966h5ba3a181bcc07a38@mail.gmail.com>
	<468E0C80.4000005@gmail.com>
	<80c99e790707060252t1094f83cn253d9b9f36965e63@mail.gmail.com>
	<468E1336.5090507@gmail.com>
Message-ID: <80c99e790707060313m3b9d59ecl144686735d8eab5c@mail.gmail.com>

gotcha. I suspect you are using the wrong "comment" in the .f90 file, then.
you use:
!!!f2py
and you should use:
!f2py

with !!!f2py the directive is not read and KM is not "intent(out)", i.e.,
the subroutine returns None. see the signature of the subroutine below, as
parsed by f2py:
------
Building modules...
        Building module "test"...
                Constructing wrapper function "calc_km"...
                  calc_km(km,[nx,ny,nz])
        Wrote C/API module "test" to file "/tmp/tmp0yiO43/src.linux-ia64-2.5
/testmodule.c"
------

with !f2py the directive intent(out) is read and the function is correctly
parsed, and returns KM:
 ------
Building modules...
        Building module "test"...
                Constructing wrapper function "calc_km"...
                  km = calc_km(km)
        Wrote C/API module "test" to file "/tmp/tmpDYUhc6/src.linux-ia64-2.5
/testmodule.c"

------


lorenzo









On 7/6/07, fred <fredmfp at gmail.com> wrote:
>
> lorenzo bolla a ?crit :
> > quite strange. it doesn't happen to me. here is what I get
> >
> I notice you use python 2.5.
> I use python 2.4.
> > what is the exact error message you get?
> I don't get error message.
> With test_f90, I get KM = None.
> With test_f, I get the right result.
>
> --
> http://scipy.org/FredericPetit
>
> _______________________________________________
> 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/20070706/2eab6789/attachment.html>

From fredmfp at gmail.com  Fri Jul  6 06:16:52 2007
From: fredmfp at gmail.com (fred)
Date: Fri, 06 Jul 2007 12:16:52 +0200
Subject: [SciPy-user] [f2py] f90 format file...
In-Reply-To: <80c99e790707060252t1094f83cn253d9b9f36965e63@mail.gmail.com>
References: <468D5158.6070304@gmail.com>
	<468D5499.3030200@gmail.com>	<468D723D.1010900@gmail.com>	<80c99e790707052353s6d68e966h5ba3a181bcc07a38@mail.gmail.com>	<468E0C80.4000005@gmail.com>
	<80c99e790707060252t1094f83cn253d9b9f36965e63@mail.gmail.com>
Message-ID: <468E1694.3020102@gmail.com>

lorenzo bolla a ?crit :
> quite strange. it doesn't happen to me. here is what I get
I have attached the compilation log file.

I can see (at least) one difference.

test_f:
Building module "test_f"
...-> km = calc_km(km)

test_f90:
Building module "test_f90"
...-> calc_km(km)

You have the same case as in the first case...

Any clue, f2py dev ?


TIA.

-- 
http://scipy.org/FredericPetit

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: log
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070706/54929f78/attachment.ksh>

From fredmfp at gmail.com  Fri Jul  6 06:21:14 2007
From: fredmfp at gmail.com (fred)
Date: Fri, 06 Jul 2007 12:21:14 +0200
Subject: [SciPy-user] [f2py] f90 format file...
In-Reply-To: <80c99e790707060313m3b9d59ecl144686735d8eab5c@mail.gmail.com>
References: <468D5158.6070304@gmail.com>
	<468D5499.3030200@gmail.com>	<468D723D.1010900@gmail.com>	<80c99e790707052353s6d68e966h5ba3a181bcc07a38@mail.gmail.com>	<468E0C80.4000005@gmail.com>	<80c99e790707060252t1094f83cn253d9b9f36965e63@mail.gmail.com>	<468E1336.5090507@gmail.com>
	<80c99e790707060313m3b9d59ecl144686735d8eab5c@mail.gmail.com>
Message-ID: <468E179A.1050909@gmail.com>

lorenzo bolla a ?crit :
> gotcha. I suspect you are using the wrong "comment" in the .f90 file, 
> then.
> you use:
> !!!f2py
> and you should use:
> !f2py
Damn !

Thanks a lot !

Cheers,

-- 
http://scipy.org/FredericPetit



From travis at enthought.com  Fri Jul  6 08:09:41 2007
From: travis at enthought.com (Travis Vaught)
Date: Fri, 6 Jul 2007 07:09:41 -0500
Subject: [SciPy-user] ANN: SciPy Conference Early Registration Reminder
Message-ID: <5A9C5264-C85B-4DD8-836D-E63582058720@enthought.com>

Greetings,

The *SciPy 2007 Conference on Scientific Computing with Python* early  
registration deadline is July 15, 2007.  After this date, the price  
for registration will increase from $150 to $200.

More information on the Conference is here: http://www.scipy.org/ 
SciPy2007
The registration page is here: https://www.enthought.com/scipy07/

The Conference is to be held on August 16-17.  Tutorial Sessions are  
being offered on August 14-15 (http://www.scipy.org/SciPy2007/ 
Tutorials).  The price to attend Tutorials is $75.  The Saturday  
following the Conference will hold a Sprint session for those  
interested in pitching in on particular development efforts.   
(suggestions welcome: http://www.scipy.org/SciPy2007/Sprints)

Today is the deadline for abstract submissions for those wanting to  
present at the conference.  Please email to abstracts at scipy.org by  
midnight US Central Time.  From the conference web page:

"If you are using Python in Scientific Computing, we'd love to hear  
from you. If you are interested in presenting at the conference, you  
may submit an abstract in Plain Text, PDF or MS Word formats to  
abstracts at scipy.org -- the deadline for abstract submission is July  
6, 2007. Papers and/or presentation slides are acceptable and are due  
by August 3, 2007. Presentations will be allowed 30-35 minutes,  
depending on the final schedule."

We're looking forward to another great gathering.

Best,

Travis


From cookedm at physics.mcmaster.ca  Fri Jul  6 08:58:32 2007
From: cookedm at physics.mcmaster.ca (David M. Cooke)
Date: Fri, 6 Jul 2007 08:58:32 -0400
Subject: [SciPy-user] installation problems on Mac OS X
In-Reply-To: <13A99D27-BD5E-4D53-86C1-770507B6C1BF@cs.kuleuven.be>
References: <13A99D27-BD5E-4D53-86C1-770507B6C1BF@cs.kuleuven.be>
Message-ID: <F5320826-43C3-47CF-A868-67BE214A9295@physics.mcmaster.ca>

On Jul 5, 2007, at 09:25 , Giovanni Samaey wrote:

> Hi all,
>
> I am trying to install scipy on my shiny new macbook:
> I followed instructions as on www.scipy.org/Installing_SciPy/ 
> Mac_OS_X/.
> Installation of vecLib was done when the computer shipped (I found a
> vecLib).  The link to the vecLib on the instruction page is dead and
> I couldn't find a download on apple.com, so I don't know if that
> installation is complete, as the instructions ask.

vecLib ships with OS X, so you don't need to install it (vecLib  
itself may be deprecated, I think it's suggested to use the  
Accelerate framework, which includes vecLib as a subframework). I  
think the page is wrong; it's part of the OS, not of the Developer  
Tools.

> Installation goes smoothly, byt scipy.test() gives 2 errors.  I tried
> both  python 2.4.4 and python 2.5.1; no difference. I have the most
> recent gfortran and I use gcc 4.0.   Scipy is taken from svn.
>
> ======================================================================
> FAIL: check_dot (scipy.lib.tests.test_blas.test_fblas1_simple)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>    File "/Library/Frameworks/Python.framework/Versions/2.4/lib/
> python2.4/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 "/Library/Frameworks/Python.framework/Versions/2.4/lib/
> python2.4/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: 3.4625784615075479e-37j
>   DESIRED: (-9+2j)
>
> ======================================================================
> FAIL: check_dot (scipy.linalg.tests.test_blas.test_fblas1_simple)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>    File "/Library/Frameworks/Python.framework/Versions/2.4/lib/
> python2.4/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 "/Library/Frameworks/Python.framework/Versions/2.4/lib/
> python2.4/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: 3.4782945282523978e-37j
>   DESIRED: (-9+2j)
>
> ----------------------------------------------------------------------
> Ran 1632 tests in 3.592s
>
> FAILED (failures=2)

Don't worry, I get these too. There's a bug there, but no one's been  
bothered enough to find it. I doubt it'll impact you unless you call  
those wrapped BLAS routines directly.

-- 
|>|\/|<
/------------------------------------------------------------------\
|David M. Cooke              http://arbutus.physics.mcmaster.ca/dmc/
|cookedm at physics.mcmaster.ca



From lorenzo.isella at gmail.com  Fri Jul  6 10:29:00 2007
From: lorenzo.isella at gmail.com (Lorenzo Isella)
Date: Fri, 6 Jul 2007 16:29:00 +0200
Subject: [SciPy-user] Moving Grid in SciPy
Message-ID: <a2b3004b0707060729yebc0b23v905adaae63cf6c69@mail.gmail.com>

Dear All,
A question which I hope will not look too general for this mailing list.
I do quite a lot of work of population equations, which generally are
a set of 1st order ODE's:

dn_k/dt=f_k(t,....)    for k=1....N

where n_k is the population of the k-th level.
Now, in my specific problem the levels at low kappa get progressively
empty as the population moves toward high k.
This means that, after a certain time, formulating the problem as a
"fixed" (hope you understand what I mean here) set of ODE's, I am
waisting cpu time on empty levels.  It would be better to "recycle"
them e.g. by sending levels with k=1,2 into levels with
k=N+1, N+2 and so on.
This method is called "moving grid" and has been used for quite a
while in my field, but before diving into the (maybe too) abundant
literature, I would like to find out if there is such a tool already
implemented in Python or SciPy directly or whose implementation is
rather trivial.
At the moment I am using integrate.odeint for my calculations (of
course still on a fixed grid).
Any suggestions are really welcome.
Many thanks

Lorenzo


From peridot.faceted at gmail.com  Fri Jul  6 10:50:45 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Fri, 6 Jul 2007 10:50:45 -0400
Subject: [SciPy-user] B-orthogonal complement of the null space of A
In-Reply-To: <468E0F3C.4070502@iam.uni-stuttgart.de>
References: <468E0F3C.4070502@iam.uni-stuttgart.de>
Message-ID: <ce557a360707060750r5910bfaib3ea0086013da106@mail.gmail.com>

On 06/07/07, Nils Wagner <nwagner at iam.uni-stuttgart.de> wrote:
> Hi all,
>
> Consider  two symmetric matrices A and B, where B is positive definite
> and A is positive semidefinite.
> How can I compute the B-orthogonal complement of the null space of A
> with scipy.linalg ?

Find the nullspace of A - which is numerically tricky, of course -
using eig or svd; the B-orthogonal complement of a set of vectors is
the nullspace of a rectangular matrix VB, where V has a row for each
vector in the nullspace. svd should find that nullspace for you.

This is not a particularly numerically stable operation. I suspect
that the question itself is ill-conditioned, so that no algorithm can
reliably solve it, but there may be a better approach. The svd is
fairly reliable, though, and lets you make your own decisions about
when a linear combination is zero enough.

Anne


From giovanni.samaey at cs.kuleuven.be  Fri Jul  6 11:09:44 2007
From: giovanni.samaey at cs.kuleuven.be (Giovanni Samaey)
Date: Fri, 6 Jul 2007 17:09:44 +0200
Subject: [SciPy-user] installation problems on Mac OS X
In-Reply-To: <F5320826-43C3-47CF-A868-67BE214A9295@physics.mcmaster.ca>
References: <13A99D27-BD5E-4D53-86C1-770507B6C1BF@cs.kuleuven.be>
	<F5320826-43C3-47CF-A868-67BE214A9295@physics.mcmaster.ca>
Message-ID: <3D69D48E-A4D6-4C39-8AD7-5D326C330FC9@cs.kuleuven.be>

>  I
> think the page is wrong; it's part of the OS, not of the Developer
> Tools.

Thanks !  I will try to update the wiki next week.
>>
>
> Don't worry, I get these too. There's a bug there, but no one's been
> bothered enough to find it. I doubt it'll impact you unless you call
> those wrapped BLAS routines directly.

thanks!  Do you mean that calling scipy.dot is OK? It looks to be, in  
any case.

Best,

Giovanni

>
> -- 
> |>|\/|<
> /------------------------------------------------------------------\
> |David M. Cooke              http://arbutus.physics.mcmaster.ca/dmc/
> |cookedm at physics.mcmaster.ca
>
> _______________________________________________
> 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/20070706/3e188a71/attachment.html>

From gary.pajer at gmail.com  Mon Jul  9 13:00:00 2007
From: gary.pajer at gmail.com (Gary Pajer)
Date: Mon, 9 Jul 2007 13:00:00 -0400
Subject: [SciPy-user] Fraction representation
In-Reply-To: <468d6837.1401420a.34d7.0973@mx.google.com>
References: <468d6837.1401420a.34d7.0973@mx.google.com>
Message-ID: <88fe22a0707091000p5e7054d1uace1ea9bb4fd51d8@mail.gmail.com>

On 7/5/07, Stephen Kelly <steveire at gmail.com> wrote:
> Hi. I wonder if it is possible to add a fraction representation.
>
> My casio calculator has a feature to display a real number as a fraction
> (eg, 0.1 as 1,10. The delimiter is not a comma though). Would it be
> possible to add such a feature to scipy?

I believe that SAGE has a rational data type and it has scipy (I
think!), but I don't know anything about how things play together in
SAGE.  In fact, I don't know much about SAGE at all, so you should
check for yourself.

http://modular.math.washington.edu/sage/

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


From fredmfp at gmail.com  Mon Jul  9 14:32:12 2007
From: fredmfp at gmail.com (fred)
Date: Mon, 9 Jul 2007 20:32:12 +0200
Subject: [SciPy-user] LU: linalg.solve vs. dgetrf/dgetrs fortran subroutines
Message-ID: <18066.32556.432189.679611@gargle.gargle.HOWL>

Hi,

I want to compare execution times between linalg.solve from scipy
and dgetrf/dgetrs fortran subroutines (before using cvxopt...)

Here my functions:

    def calc_KW_LU(self, KV):
        from scipy.linalg import solve
        from sys import stdout
        t0 = time.time()
        print 'Computing kriging weights (scipy LU version)...',
        stdout.flush()
        KW = solve(self.KM, KV)
        self.KW = KW[:-1].squeeze()
        print '('+str(time.time()-t0)+' s)',
        print 'done !'
        stdout.flush()


subroutine calc_KW(KW, KM, KV, nbx, nby, nbz)
  
  implicit none
  
!f2py real intent(in, out) :: KW
!f2py real*8 intent(in) :: KM, KV
!f2py integer intent(in) :: nbx, nby, nbz
  
  integer :: nbx, nby, nbz
  real :: KW((2*nbx+1)*(2*nby+1)*(2*nbz+1))
  real*8 :: KM((2*nbx+1)*(2*nby+1)*(2*nbz+1)+1, (2*nbx+1)*(2*nby+1)*(2*nbz+1)+1)
  real*8 :: KV((2*nbx+1)*(2*nby+1)*(2*nbz+1)+1)
  
  integer :: ndim
  integer :: info
  integer, dimension(:), allocatable :: ipvt
  
  ndim = (2*nbx+1)*(2*nby+1)*(2*nbz+1)+1
  
  allocate(ipvt(ndim))
  
  call dgetrf(ndim, ndim, KM, ndim, ipvt, info)
  call dgetrs('N', ndim, 1, KM, ndim, ipvt, KV, ndim, info)
  
  KW = KV(1:size(KV)-1)
  
end subroutine calc_KW

This last is called by my python method:

    def calc_KW_LU_Fmodule(self, KV):
        from scipy import float32, zeros
        from sys import stdout
        from calc_KW_LU import calc_kw
        t0 = time.time()
        nbx, nby, nbz = self.nbx, self.nby, self.nbz
        ndx, ndy, ndz = (2*nbx+1), (2*nby+1), (2*nbz+1)
        ndt = ndx*ndy*ndz
        KM, KV = self.KM, self.KV
        print 'Computing kriging weights (LU Fortran module version)...',
        stdout.flush()
        KW = zeros((ndt,1), dtype=float32)
        KW = calc_kw(KW, KM, KV, nbx, nby, nbz)
        self.KW = KW.squeeze()
        print '('+str(time.time()-t0)+' s)',
        print 'done !'
        stdout.flush()

The execution times are very different:

1) On my Core 2 Duo T7200 (i686, Debian etch):

Computing kriging weights (LU version)... (142.409274817 s) done !

Computing kriging weights (LU Fortran module version)... (157.362993002 s) done !

2) On my Pentium 930 D (x86_64, Debian etch), differences are bigger:

Computing kriging weights (LU version)... (283.214334011 s) done !

Computing kriging weights (LU Fortran module version)... (583.182439089 s) done !

Could someone give me any clues ?

TIA.


Cheers,

-- 
http://scipy.org/FredericPetit


From rshepard at appl-ecosys.com  Mon Jul  9 14:34:57 2007
From: rshepard at appl-ecosys.com (Rich Shepard)
Date: Mon, 9 Jul 2007 11:34:57 -0700 (PDT)
Subject: [SciPy-user] NumPy Informational Message
Message-ID: <Pine.LNX.4.64.0707091132290.7552@salmo.appl-ecosys.com>

   When I run the application I'm developing, python responds with:

Overwriting info=<function info at 0xb5ef45a4> from scipy.misc (was
<function info at 0xb605b95c> from numpy.lib.utils)

   Don't recall having seen this before, and I've not changed numpy or scipy
so I do not know why I'm getting this message. Clarification would be
appreciated.

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 rmay at ou.edu  Mon Jul  9 17:05:36 2007
From: rmay at ou.edu (Ryan May)
Date: Mon, 09 Jul 2007 16:05:36 -0500
Subject: [SciPy-user] SciPy 2007
Message-ID: <4692A320.5010302@ou.edu>

Hi,

I was curious if anyone had any recommendations for a hotel to stay at 
for the SciPy 2007 conference.  I saw the link to local lodging on 
CalTech's site, but I was curious if anyone had some place they'd 
recommend based on experience (or even _not_ recommend).

Also, would anyone have any idea how long things would go on Friday? 
I'm trying to figure out when I should fly back.

Thanks,

Ryan

-- 
Ryan May
Graduate Research Assistant
School of Meteorology
University of Oklahoma


From robert.kern at gmail.com  Mon Jul  9 17:51:36 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Mon, 09 Jul 2007 16:51:36 -0500
Subject: [SciPy-user] SciPy 2007
In-Reply-To: <4692A320.5010302@ou.edu>
References: <4692A320.5010302@ou.edu>
Message-ID: <4692ADE8.4090704@gmail.com>

Ryan May wrote:
> Hi,
> 
> I was curious if anyone had any recommendations for a hotel to stay at 
> for the SciPy 2007 conference.  I saw the link to local lodging on 
> CalTech's site, but I was curious if anyone had some place they'd 
> recommend based on experience (or even _not_ recommend).

I like the Saga Motor Hotel. No frills, but cheap and close. It's website is
down currently, but here are their address and phone number.

www.thesagamotorhotel.com
1633 E Colorado Blvd
Pasadena, CA 91106
(626) 795-0431

I'm not a fan of the Vagabond Inn although it is also cheap and close.

> Also, would anyone have any idea how long things would go on Friday? 
> I'm trying to figure out when I should fly back.

The talks will end around 5. It's okay to shuffle out early to catch a flight.
However, people do get together for dinner afterwards. There is also a sprint
day on Saturday, so I imagine there will be some hacking on Friday night, too.

-- 
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 rmay at ou.edu  Mon Jul  9 14:57:19 2007
From: rmay at ou.edu (Ryan May)
Date: Mon, 09 Jul 2007 13:57:19 -0500
Subject: [SciPy-user] SciPy 2007 Lodging
Message-ID: <4692850F.6010204@ou.edu>

Hi,

I was curious if anyone had any recommendations for a hotel to stay at 
for the SciPy 2007 conference.  I saw the link to local lodging on 
CalTech's site, but I was curious if anyone had some place they'd 
recommend based on experience (or even _not_ recommend).

Thanks,

Ryan

-- 
Ryan May
Graduate Research Assistant
School of Meteorology
University of Oklahoma


From kmaheswaran at hotmail.com  Mon Jul  9 18:13:56 2007
From: kmaheswaran at hotmail.com (Krishnan Maheswaran)
Date: Mon, 09 Jul 2007 22:13:56 +0000
Subject: [SciPy-user] TimeSeriesPackage
Message-ID: <BAY137-F325FE31958AF50D4275B5BBD060@phx.gbl>

Hi All,
I would appreciate any help on installing the scipy timeseries module ( I 
had no problem installing the maskedarray package). Please see errors below.

Many Thanks,
KM





Tasks:
  i       - Show python/platform/machine information
  ie      - Show environment information
  c       - Show C compilers information
  c<name> - Set C compiler (current:None)
  f       - Show Fortran compilers information
  f<name> - Set Fortran compiler (current:None)
  e       - Edit proposed sys.argv[1:].

Task aliases:
  0         - Configure
  1         - Build
  2         - Install
  2<prefix> - Install with prefix.
  3         - Inplace build
  4         - Source distribution
  5         - Binary distribution

Proposed sys.argv = 
['C:\\Python25\\Lib\\site-packages\\timeseries\\setup.py']



Tasks:
  i       - Show python/platform/machine information
  ie      - Show environment information
  c       - Show C compilers information
  c<name> - Set C compiler (current:None)
  f       - Show Fortran compilers information
  f<name> - Set Fortran compiler (current:None)
  e       - Edit proposed sys.argv[1:].

Task aliases:
  0         - Configure
  1         - Build
  2         - Install
  2<prefix> - Install with prefix.
  3         - Inplace build
  4         - Source distribution
  5         - Binary distribution

Proposed sys.argv = 
['C:\\Python25\\Lib\\site-packages\\timeseries\\setup.py', 'install']

------------------------------------------------------------------------
running install
running build
running config_cc
unifing config_cc, config, build_clib, build_ext, build commands --compiler 
options
running config_fc
unifing config_fc, config, build_clib, build_ext, build commands --fcompiler 
options
running build_src
building extension "timeseries.cseries" sources
running build_py
creating build
creating build\lib.win32-2.5
creating build\lib.win32-2.5\timeseries
copying .\const.py -> build\lib.win32-2.5\timeseries
copying .\parser.py -> build\lib.win32-2.5\timeseries
copying .\reportlib.py -> build\lib.win32-2.5\timeseries
copying .\tcore.py -> build\lib.win32-2.5\timeseries
copying .\tdates.py -> build\lib.win32-2.5\timeseries
copying .\tmulti.py -> build\lib.win32-2.5\timeseries
copying .\tseries.py -> build\lib.win32-2.5\timeseries
copying .\__init__.py -> build\lib.win32-2.5\timeseries
creating build\lib.win32-2.5\timeseries\lib
copying .\lib\filters.py -> build\lib.win32-2.5\timeseries\lib
copying .\lib\interpolate.py -> build\lib.win32-2.5\timeseries\lib
copying .\lib\moving_funcs.py -> build\lib.win32-2.5\timeseries\lib
copying .\lib\__init__.py -> build\lib.win32-2.5\timeseries\lib
creating build\lib.win32-2.5\timeseries\io
copying .\io\__init__.py -> build\lib.win32-2.5\timeseries\io
creating build\lib.win32-2.5\timeseries\plotlib
copying .\plotlib\mpl_timeseries.py -> 
build\lib.win32-2.5\timeseries\plotlib
copying .\plotlib\__init__.py -> build\lib.win32-2.5\timeseries\plotlib
running build_ext
No module named msvccompiler in numpy.distutils, trying from distutils, 
trying from distutils. The .net framework sdk needs to be installed before 
building extension for python.

_________________________________________________________________
Advertisement: New jobsjobsjobs.com.au. Find thousands of jobs online now! 
http://a.ninemsn.com.au/b.aspx?URL=http%3A%2F%2Fad%2Eau%2Edoubleclick%2Enet%2Fclk%3B114014868%3B17770752%3Bi%3Fhttp%3A%2F%2Fwww%2Ejobsjobsjobs%2Ecom%2Eau&_t=762242361&_r=Hotmail_email_tagline_July07&_m=EXT



From pgmdevlist at gmail.com  Mon Jul  9 18:32:22 2007
From: pgmdevlist at gmail.com (Pierre GM)
Date: Mon, 9 Jul 2007 18:32:22 -0400
Subject: [SciPy-user] TimeSeriesPackage
In-Reply-To: <BAY137-F325FE31958AF50D4275B5BBD060@phx.gbl>
References: <BAY137-F325FE31958AF50D4275B5BBD060@phx.gbl>
Message-ID: <200707091832.23670.pgmdevlist@gmail.com>

On Monday 09 July 2007 18:13:56 Krishnan Maheswaran wrote:
> Hi All,
Hello !

> I would appreciate any help on installing the scipy timeseries module
Yay, another user ! More feedback !

> ( I 
> had no problem installing the maskedarray package). Please see errors
> below.

So, looks like you're trying to install the package on Windows. Here's an 
excerpt from the archives (from Matt Knox, the codeveloper of TimeSeries):

"""
cygwin works pretty seamlessly on windows with python 2.5, so I'd recommend 
that approach. It doesn't require any special witchcraft that used to be 
needed for compiling extensions on windows with earlier versions of python. 
Here are some good general instructions for compiling extensions with cygwin 
and python 2.5: ?http://boodebr.org/main/python/build-windows-extensions

Once you've followed the setup steps there, just open the cygwin shell and you 
can do "python setup.py build_ext -i" to build the extension in place in that 
same directory. Or I *believe* "python setup.py install" will compile the 
extension and install it in your site-packages directory if memory serves me 
correctly... but I haven't tried that in a while and don't have the necessary 
stuff setup on my pc here to test. You can even create a nice binary installer 
by doing "python setup.py bdist_wininst" if you want to pass it around to your 
friends and family :)
"""

I remember following this method when I had to install the package on my boss' 
computer (a Windows machine), and that did work seamlessly indeed when you 
have Python 2.5. Let us know how it goes.

Sincerely,
P.


From matthieu.brucher at gmail.com  Tue Jul 10 02:18:10 2007
From: matthieu.brucher at gmail.com (Matthieu Brucher)
Date: Tue, 10 Jul 2007 08:18:10 +0200
Subject: [SciPy-user] TimeSeriesPackage
In-Reply-To: <BAY137-F325FE31958AF50D4275B5BBD060@phx.gbl>
References: <BAY137-F325FE31958AF50D4275B5BBD060@phx.gbl>
Message-ID: <e76aa17f0707092318k55cbc8b6rb948f39e561dc43f@mail.gmail.com>

Hi,

What compiler are you using ?

The problem with Python for Windows out of the box is that it requires
Visual Studio 2003 and not Visual Studio 2005. So you must find the free
toolkit from Microsoft, but Microsoft does not support anymore (moved to
2005 which is perfectly logical, gcc does not support the old 3.x series as
well) or compile Python with VS 2005.
What is more, for the Express version of VS 2005 (the free one, contrary to
what Pierre's link says, Microsoft provides a free compiler), you have to
install the platform SDK.

Matthieu


2007/7/10, Krishnan Maheswaran <kmaheswaran at hotmail.com>:
>
> Hi All,
> I would appreciate any help on installing the scipy timeseries module ( I
> had no problem installing the maskedarray package). Please see errors
> below.
>
> Many Thanks,
> KM
>
>
>
>
>
> Tasks:
>   i       - Show python/platform/machine information
>   ie      - Show environment information
>   c       - Show C compilers information
>   c<name> - Set C compiler (current:None)
>   f       - Show Fortran compilers information
>   f<name> - Set Fortran compiler (current:None)
>   e       - Edit proposed sys.argv[1:].
>
> Task aliases:
>   0         - Configure
>   1         - Build
>   2         - Install
>   2<prefix> - Install with prefix.
>   3         - Inplace build
>   4         - Source distribution
>   5         - Binary distribution
>
> Proposed sys.argv =
> ['C:\\Python25\\Lib\\site-packages\\timeseries\\setup.py']
>
>
>
> Tasks:
>   i       - Show python/platform/machine information
>   ie      - Show environment information
>   c       - Show C compilers information
>   c<name> - Set C compiler (current:None)
>   f       - Show Fortran compilers information
>   f<name> - Set Fortran compiler (current:None)
>   e       - Edit proposed sys.argv[1:].
>
> Task aliases:
>   0         - Configure
>   1         - Build
>   2         - Install
>   2<prefix> - Install with prefix.
>   3         - Inplace build
>   4         - Source distribution
>   5         - Binary distribution
>
> Proposed sys.argv =
> ['C:\\Python25\\Lib\\site-packages\\timeseries\\setup.py', 'install']
>
> ------------------------------------------------------------------------
> running install
> running build
> running config_cc
> unifing config_cc, config, build_clib, build_ext, build commands
> --compiler
> options
> running config_fc
> unifing config_fc, config, build_clib, build_ext, build commands
> --fcompiler
> options
> running build_src
> building extension "timeseries.cseries" sources
> running build_py
> creating build
> creating build\lib.win32-2.5
> creating build\lib.win32-2.5\timeseries
> copying .\const.py -> build\lib.win32-2.5\timeseries
> copying .\parser.py -> build\lib.win32-2.5\timeseries
> copying .\reportlib.py -> build\lib.win32-2.5\timeseries
> copying .\tcore.py -> build\lib.win32-2.5\timeseries
> copying .\tdates.py -> build\lib.win32-2.5\timeseries
> copying .\tmulti.py -> build\lib.win32-2.5\timeseries
> copying .\tseries.py -> build\lib.win32-2.5\timeseries
> copying .\__init__.py -> build\lib.win32-2.5\timeseries
> creating build\lib.win32-2.5\timeseries\lib
> copying .\lib\filters.py -> build\lib.win32-2.5\timeseries\lib
> copying .\lib\interpolate.py -> build\lib.win32-2.5\timeseries\lib
> copying .\lib\moving_funcs.py -> build\lib.win32-2.5\timeseries\lib
> copying .\lib\__init__.py -> build\lib.win32-2.5\timeseries\lib
> creating build\lib.win32-2.5\timeseries\io
> copying .\io\__init__.py -> build\lib.win32-2.5\timeseries\io
> creating build\lib.win32-2.5\timeseries\plotlib
> copying .\plotlib\mpl_timeseries.py ->
> build\lib.win32-2.5\timeseries\plotlib
> copying .\plotlib\__init__.py -> build\lib.win32-2.5\timeseries\plotlib
> running build_ext
> No module named msvccompiler in numpy.distutils, trying from distutils,
> trying from distutils. The .net framework sdk needs to be installed before
> building extension for python.
>
> _________________________________________________________________
> Advertisement: New jobsjobsjobs.com.au. Find thousands of jobs online now!
>
> http://a.ninemsn.com.au/b.aspx?URL=http%3A%2F%2Fad%2Eau%2Edoubleclick%2Enet%2Fclk%3B114014868%3B17770752%3Bi%3Fhttp%3A%2F%2Fwww%2Ejobsjobsjobs%2Ecom%2Eau&_t=762242361&_r=Hotmail_email_tagline_July07&_m=EXT
>
> _______________________________________________
> 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/20070710/b3e20d93/attachment.html>

From kmaheswaran at hotmail.com  Tue Jul 10 10:35:28 2007
From: kmaheswaran at hotmail.com (Krishnan Maheswaran)
Date: Tue, 10 Jul 2007 14:35:28 +0000
Subject: [SciPy-user] TimeSeriesPackage
In-Reply-To: <200707091832.23670.pgmdevlist@gmail.com>
Message-ID: <BAY137-F317A3A0C7B2822DC4DD2EDBD050@phx.gbl>

Thanks. I built the binaries using MinGW.


>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] TimeSeriesPackage
>Date: Mon, 9 Jul 2007 18:32:22 -0400
>
>On Monday 09 July 2007 18:13:56 Krishnan Maheswaran wrote:
> > Hi All,
>Hello !
>
> > I would appreciate any help on installing the scipy timeseries module
>Yay, another user ! More feedback !
>
> > ( I
> > had no problem installing the maskedarray package). Please see errors
> > below.
>
>So, looks like you're trying to install the package on Windows. Here's an
>excerpt from the archives (from Matt Knox, the codeveloper of TimeSeries):
>
>"""
>cygwin works pretty seamlessly on windows with python 2.5, so I'd recommend
>that approach. It doesn't require any special witchcraft that used to be
>needed for compiling extensions on windows with earlier versions of python.
>Here are some good general instructions for compiling extensions with 
>cygwin
>and python 2.5:  http://boodebr.org/main/python/build-windows-extensions
>
>Once you've followed the setup steps there, just open the cygwin shell and 
>you
>can do "python setup.py build_ext -i" to build the extension in place in 
>that
>same directory. Or I *believe* "python setup.py install" will compile the
>extension and install it in your site-packages directory if memory serves 
>me
>correctly... but I haven't tried that in a while and don't have the 
>necessary
>stuff setup on my pc here to test. You can even create a nice binary 
>installer
>by doing "python setup.py bdist_wininst" if you want to pass it around to 
>your
>friends and family :)
>"""
>
>I remember following this method when I had to install the package on my 
>boss'
>computer (a Windows machine), and that did work seamlessly indeed when you
>have Python 2.5. Let us know how it goes.
>
>Sincerely,
>P.
>_______________________________________________
>SciPy-user mailing list
>SciPy-user at scipy.org
>http://projects.scipy.org/mailman/listinfo/scipy-user

_________________________________________________________________
Advertisement: Are you paid what you're worth? Find out: SEEK Salary Centre 
http://a.ninemsn.com.au/b.aspx?URL=http%3A%2F%2Fninemsn%2Eseek%2Ecom%2Eau%2Fcareer%2Dresources%2Fsalary%2Dcentre%2F%3Ftracking%3Dsk%3Ahet%3Asc%3Anine%3A0%3Ahot%3Atext&_t=764565661&_r=july07_endtext_salary&_m=EXT



From jakobsybren at gmail.com  Wed Jul  4 13:59:09 2007
From: jakobsybren at gmail.com (Jakob van Bethlehem)
Date: Wed, 4 Jul 2007 17:59:09 +0000 (UTC)
Subject: [SciPy-user] scipy import warning - how to remove
References: <20070702085805.99798.qmail@web27407.mail.ukl.yahoo.com>
Message-ID: <loom.20070704T195247-91@post.gmane.org>

Robert VERGNES <robert.vergnes <at> yahoo.fr> writes:

> 
> 
> Hello,each time I do a from scipy import * I get a warning. How to remove it
?This has a bad result on py2exe which seems to believe it is an error and stop
starting my application.Any ideas ?(how to force not to importe scipytest
?)ThanxRobertPlatform - Windows - Scipy 0.5.2from scipy import 

I'm not sure whether it is possible to prevent the module from loading. You can
however prevent the message quite easily (on Linux anyway) by going to the
packages-dir of your scipy installation (...../site-packages/scipy)
In this directory give the following command:

find . -name "*.py" -exec sed -i -e "s/ScipyTest/NumpyTest/g" "{}" \;

This will replace the deprecated definitions/calls of ScipyTest by the new
NumpyTest, including the also deprecated class ScipyTestCase. (I suppose this is
precisely what has been done in the svn-version)




From openopt at ukr.net  Wed Jul 11 03:11:13 2007
From: openopt at ukr.net (dmitrey)
Date: Wed, 11 Jul 2007 10:11:13 +0300
Subject: [SciPy-user] matlab isequal equivalent in python or numpy
Message-ID: <46948291.1040301@ukr.net>

I need check does array1 equal to array2 or no, lots of times.
In MATLAB I just used

if isequal(array1, array2) ...

I guess it uses binary code comparison, w/o translating to numbers, 
because isequal works with any types of objects, not only numeric ones

in python I use

if array1.shape==array2.shape and all(array1==array2):...

is it possible to use something more faster and/or better?
D.


From lbolla at gmail.com  Wed Jul 11 06:03:06 2007
From: lbolla at gmail.com (lorenzo bolla)
Date: Wed, 11 Jul 2007 12:03:06 +0200
Subject: [SciPy-user] matlab isequal equivalent in python or numpy
In-Reply-To: <46948291.1040301@ukr.net>
References: <46948291.1040301@ukr.net>
Message-ID: <80c99e790707110303q75b7002u44cee56b30b7804f@mail.gmail.com>

can't you use "numpy.all" instead of "all"?


In [786]: x = numpy.arange(10)

In [787]: y = numpy.arange(10)

In [788]: z = numpy.arange(11)

In [789]: all(x==y)
Out[789]: True

In [790]: all(x==z)
------------------------------------
<type 'exceptions.TypeError'>

In [791]: numpy.all(x==y)
Out[791]: True

In [792]: numpy.all(x==z)
Out[792]: False



hth,

lorenzo






On 7/11/07, dmitrey <openopt at ukr.net> wrote:
>
> I need check does array1 equal to array2 or no, lots of times.
> In MATLAB I just used
>
> if isequal(array1, array2) ...
>
> I guess it uses binary code comparison, w/o translating to numbers,
> because isequal works with any types of objects, not only numeric ones
>
> in python I use
>
> if array1.shape==array2.shape and all(array1==array2):...
>
> is it possible to use something more faster and/or better?
> D.
> _______________________________________________
> 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/20070711/75290b9c/attachment.html>

From nwagner at iam.uni-stuttgart.de  Wed Jul 11 11:00:24 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Wed, 11 Jul 2007 17:00:24 +0200
Subject: [SciPy-user] B-orthogonal complement of the null space of A
In-Reply-To: <ce557a360707060750r5910bfaib3ea0086013da106@mail.gmail.com>
References: <468E0F3C.4070502@iam.uni-stuttgart.de>
	<ce557a360707060750r5910bfaib3ea0086013da106@mail.gmail.com>
Message-ID: <4694F088.5020503@iam.uni-stuttgart.de>

Anne Archibald wrote:
> On 06/07/07, Nils Wagner <nwagner at iam.uni-stuttgart.de> wrote:
>   
>> Hi all,
>>
>> Consider  two symmetric matrices A and B, where B is positive definite
>> and A is positive semidefinite.
>> How can I compute the B-orthogonal complement of the null space of A
>> with scipy.linalg ?
>>     
>
> Find the nullspace of A - which is numerically tricky, of course -
> using eig or svd; the B-orthogonal complement of a set of vectors is
> the nullspace of a rectangular matrix VB, where V has a row for each
> vector in the nullspace. svd should find that nullspace for you.
>
> This is not a particularly numerically stable operation. I suspect
> that the question itself is ill-conditioned, so that no algorithm can
> reliably solve it, but there may be a better approach. The svd is
> fairly reliable, though, and lets you make your own decisions about
> when a linear combination is zero enough.
>
> Anne
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   
Hi Anne,

Thank you very much for your detailed information !
I have enclosed a small script (which is hopefully correct)
illustrating the computation of the B-orthogonal complement of the null
space of A for dense matrices.
How can I extend it to sparse matrices ? I mean what can be used instead
of linalg.qr and linalg.svd ?

Nils
 


From nwagner at iam.uni-stuttgart.de  Wed Jul 11 11:03:23 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Wed, 11 Jul 2007 17:03:23 +0200
Subject: [SciPy-user] B-orthogonal complement of the null space of A
In-Reply-To: <4694F088.5020503@iam.uni-stuttgart.de>
References: <468E0F3C.4070502@iam.uni-stuttgart.de>	<ce557a360707060750r5910bfaib3ea0086013da106@mail.gmail.com>
	<4694F088.5020503@iam.uni-stuttgart.de>
Message-ID: <4694F13B.9050009@iam.uni-stuttgart.de>

Nils Wagner wrote:
> Anne Archibald wrote:
>   
>> On 06/07/07, Nils Wagner <nwagner at iam.uni-stuttgart.de> wrote:
>>   
>>     
>>> Hi all,
>>>
>>> Consider  two symmetric matrices A and B, where B is positive definite
>>> and A is positive semidefinite.
>>> How can I compute the B-orthogonal complement of the null space of A
>>> with scipy.linalg ?
>>>     
>>>       
>> Find the nullspace of A - which is numerically tricky, of course -
>> using eig or svd; the B-orthogonal complement of a set of vectors is
>> the nullspace of a rectangular matrix VB, where V has a row for each
>> vector in the nullspace. svd should find that nullspace for you.
>>
>> This is not a particularly numerically stable operation. I suspect
>> that the question itself is ill-conditioned, so that no algorithm can
>> reliably solve it, but there may be a better approach. The svd is
>> fairly reliable, though, and lets you make your own decisions about
>> when a linear combination is zero enough.
>>
>> Anne
>> _______________________________________________
>> SciPy-user mailing list
>> SciPy-user at scipy.org
>> http://projects.scipy.org/mailman/listinfo/scipy-user
>>   
>>     
> Hi Anne,
>
> Thank you very much for your detailed information !
> I have enclosed a small script (which is hopefully correct)
> illustrating the computation of the B-orthogonal complement of the null
> space of A for dense matrices.
> How can I extend it to sparse matrices ? I mean what can be used instead
> of linalg.qr and linalg.svd ?
>
> Nils
>  
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>   
Oops, I missed out the attachment.

Nils

-------------- next part --------------
A non-text attachment was scrubbed...
Name: orthogonalcomplement.py
Type: text/x-python
Size: 674 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070711/4063feb3/attachment.py>

From fie.pye at atlas.cz  Wed Jul 11 15:18:00 2007
From: fie.pye at atlas.cz (fie.pye at atlas.cz)
Date: Wed, 11 Jul 2007 19:18:00 GMT
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
Message-ID: <ad2b96c6499c4b6cba0af6f87d6f9405@c0391fd0ef7044bc95651d396bc8cf1d>

>---------------------------------------------------------
>Od: David Cournapeau
>P?ijato: 3.7.2007 4:16:32
>P?edm?t: Re: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
>
>fie.pye at atlas.cz wrote:
>
>> Hello.
>
>>
>
>> I would like to install NumPy and SciPy modules. I compiled and install BLAS, LAPACK and ATLAS libraries. After installation and importing NumPy I obtained message:
>
>>
>
>>>>> import numpy
>
>> Traceback (most recent call last):
>
>> File "", line 1, in 
>
>> File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/__init__.py", line 43, in 
>
>> import linalg
>
>> File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in 
>
>> from linalg import *
>
>> File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in 
>
>> from numpy.linalg import lapack_lite
>
>> ImportError: liblapack.so: cannot open shared object file: No such file or directory
>
>> It seams to me that BLAS and LAPACK installation is OK.
>
>Well, it may not be :) If you do (once you exported LD_LIBRARY_PATH):
>
>
>
>ldd 
>
>/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so
>
>
>
>What do you get ? The most useful information, actually, is to give us 
>
>the log of python setup.py config &> log when building numpy and scipy.
>
>>
>
>> # LAPACK installation Make.inc
>
>> PLAT = _LINUX
>
>> FORTRAN = gfortran
>
>> OPTS = -funroll-all-loops -O3 -fpic
>
>> DRVOPTS = $(OPTS)
>
>> NOOPT =
>
>Here you forgot NOOPT = -fpic. Then, how did you compile atlas ? You 
>
>should try first without atlas at all, I think: compile blas, lapack, 
>
>and check that python setup.py config picks up the libraries.
>
>
>
>As a note, there are source rpms available for blas/lapack/atlas/scipy: 
>
>the binary rpms do not support CENTOS, but I think building the rpm on 
>
>CENTOS should not be too difficult. Everything is taken care of in the 
>
>rpms: options, compatible fortran ABI, full pic version of liblapack to 
>
>build complete LAPACK based on ATLAS, etc...
>
>
>
>http://software.opensuse.org/download/home:/ashigabou/ (pick up source 
>
>rpm for FC 6 or 7, this should be the more similar to centos).
>
>
>
>If this does not work, I would like to hear about it,
>
>
>
>David
>
>_______________________________________________
>
>SciPy-user mailing list
>
>SciPy-user at scipy.org
>
>http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>
Hi David.

Thank you for your response. Based on your notes I recompiled BLAS and LAPACK libraries and build and installed python2.5.1 and numpy again. Unfortunately the error appears again

>>> import numpy
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/__init__.py", line 43, in <module>
    import linalg
  File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in <module>
    from linalg import *
  File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in <module>
    from numpy.linalg import lapack_lite
ImportError: /usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so: undefined symbol: _gfortran_concat_string

Installation information:

# ldd /usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so
	libg2c.so.0 => /usr/lib64/libg2c.so.0 (0x00002aaaaae52000)
	libm.so.6 => /lib64/libm.so.6 (0x00002aaaab073000)
	libgcc_s.so.1 => /lib64/libgcc_s.so.1 (0x00002aaaab2f6000)
	libc.so.6 => /lib64/libc.so.6 (0x00002aaaab504000)
	/lib64/ld-linux-x86-64.so.2 (0x0000555555554000)

# /usr/local/lib64/blas
-rw-r--r-- 1 root root 612242 Jul 10 23:25 libblas.a
-rwxr-xr-x 1 root root 358087 Jul 10 23:25 libblas.so
lrwxrwxrwx 1 root root      9 Jul 10 23:27 libfblas.a -> libblas.a

# /usr/local/lib64/lapack
lrwxrwxrwx 1 root root       11 Jul 10 23:36 libflapack.a -> liblapack.a
-rw-r--r-- 1 root root 10938922 Jul 10 23:34 liblapack.a
-rwxr-xr-x 1 root root  6948895 Jul 10 23:38 liblapack.so

for python2.5.1 configuration see attached file: python_config.log.gz
for numpy building and installation see attached files: numpy_build.log.gz, numpy_install.log.gz

fie pye


-------------- next part --------------
A non-text attachment was scrubbed...
Name: python_config.log.gz
Type: application/x-gzip
Size: 22512 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070711/12bd03e0/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: numpy_build.log.gz
Type: application/x-gzip
Size: 3747 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070711/12bd03e0/attachment-0001.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: numpy_install.log.gz
Type: application/x-gzip
Size: 6180 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070711/12bd03e0/attachment-0002.bin>

From david at ar.media.kyoto-u.ac.jp  Thu Jul 12 00:05:53 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Thu, 12 Jul 2007 13:05:53 +0900
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
In-Reply-To: <ad2b96c6499c4b6cba0af6f87d6f9405@c0391fd0ef7044bc95651d396bc8cf1d>
References: <ad2b96c6499c4b6cba0af6f87d6f9405@c0391fd0ef7044bc95651d396bc8cf1d>
Message-ID: <4695A8A1.9030204@ar.media.kyoto-u.ac.jp>


> Hi David.
>
> Thank you for your response. Based on your notes I recompiled BLAS and LAPACK libraries and build and installed python2.5.1 and numpy again. Unfortunately the error appears again
>
>>>> import numpy
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/__init__.py", line 43, in <module>
>     import linalg
>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in <module>
>     from linalg import *
>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in <module>
>     from numpy.linalg import lapack_lite
> ImportError: /usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so: undefined symbol: _gfortran_concat_string
>
This is typically because you did not use the same fortran compiler 
everywhere (g77 vs gfortran): the default fortran compiler on fedora 
core (at least since 5) is gfortran, but this is not the default for 
numpy or scipy. Again, this is exactly for this kind of things that I 
packaged blas/lapack/atlas/numpy/scipy. Did you try building the source 
rpms instead of building everything from sources ?

cheers,

David


From rhc28 at cornell.edu  Thu Jul 12 01:14:38 2007
From: rhc28 at cornell.edu (Rob Clewley)
Date: Thu, 12 Jul 2007 01:14:38 -0400
Subject: [SciPy-user] ANN: PyDSTool 0.85 released
Message-ID: <a749952d0707112214k6040a2a6qd71b11773b0417d7@mail.gmail.com>

We have released an update to the PyDSTool dynamical systems and
modeling package at http://sourceforge.net/projects/pydstool.  There
are lots of minor improvements and fixes in this version, but a
powerful new feature is the support for user-defined functions for
continuation using PyCont.

We have also added some tools for basic phase-plane analysis,
calculation of phase response curves for oscillators, and support for
some special math functions.  Please see the release notes for more
details, and the documentation at http://pydstool.sourceforge.net.

We plan to continue to add features and improve our code over the
coming months, and we look forward to hearing feedback on our
progress.  In particular, contributions are very welcome.  Thanks
again to the Scipy, Numpy, Matplotlib, and SWIG developers for
facilitating our work!

-- The PyDSTool developers


From v.vauchey at gmail.com  Thu Jul 12 03:08:03 2007
From: v.vauchey at gmail.com (vincent vauchey)
Date: Thu, 12 Jul 2007 09:08:03 +0200
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
In-Reply-To: <4695A8A1.9030204@ar.media.kyoto-u.ac.jp>
References: <ad2b96c6499c4b6cba0af6f87d6f9405@c0391fd0ef7044bc95651d396bc8cf1d>
	<4695A8A1.9030204@ar.media.kyoto-u.ac.jp>
Message-ID: <36fdacbc0707120008v1142067ap2e126aeac2326d6a@mail.gmail.com>

Hello,

there are many time I obtain the same problem because fortran compiler used
was not the good, like you i tried to install lapack and other libraries,
but that failed. Now i used following line to install numpy ans scipy and on
many different computer with rh3, and i haven't problem.

"

python setup.py build build_ext -lg2c install

"

It is maybe useless, but try that.

Regards.

Vincent

2007/7/12, David Cournapeau <david at ar.media.kyoto-u.ac.jp>:
>
>
> > Hi David.
> >
> > Thank you for your response. Based on your notes I recompiled BLAS and
> LAPACK libraries and build and installed python2.5.1 and numpy again.
> Unfortunately the error appears again
> >
> >>>> import numpy
> > Traceback (most recent call last):
> >   File "<stdin>", line 1, in <module>
> >   File
> "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/__init__.py",
> line 43, in <module>
> >     import linalg
> >   File
> "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/__init__.py",
> line 4, in <module>
> >     from linalg import *
> >   File
> "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/linalg.py",
> line 25, in <module>
> >     from numpy.linalg import lapack_lite
> > ImportError:
> /usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so:
> undefined symbol: _gfortran_concat_string
> >
> This is typically because you did not use the same fortran compiler
> everywhere (g77 vs gfortran): the default fortran compiler on fedora
> core (at least since 5) is gfortran, but this is not the default for
> numpy or scipy. Again, this is exactly for this kind of things that I
> packaged blas/lapack/atlas/numpy/scipy. Did you try building the source
> rpms instead of building everything from sources ?
>
> cheers,
>
> David
> _______________________________________________
> 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/20070712/35790107/attachment.html>

From asefu at fooie.net  Thu Jul 12 21:12:21 2007
From: asefu at fooie.net (Fahd Sultan)
Date: Thu, 12 Jul 2007 21:12:21 -0400
Subject: [SciPy-user] scipy fblas.so functions not found
In-Reply-To: <467CF54A.1080406@ar.media.kyoto-u.ac.jp>
References: <loom.20070614T135305-673@post.gmane.org>	<20070614134122.GB21936@localhost.ee.columbia.edu>	<20070614140423.GC21936@localhost.ee.columbia.edu>	<loom.20070614T182651-110@post.gmane.org>	<20070614193211.GG14029@avicenna.cc.columbia.edu>	<loom.20070615T185440-402@post.gmane.org>	<467B68B4.1060903@fooie.net>	<467B6908.70901@ar.media.kyoto-u.ac.jp>	<467C4F2A.2010206@fooie.net>
	<467CF54A.1080406@ar.media.kyoto-u.ac.jp>
Message-ID: <4696D175.4060900@fooie.net>

Hi,

I figured out the issues. I had to alter the spec files for redhat 4 AS. 

for instance I had to add to each spec file:

%define is_redhat %(test -e /etc/redhat-release && echo 1 || echo 0)

I had to add a condition BuildRequires and requires to use gcc4-gfortan 
and lapack  as opposed to gcc-fortran and lapack3

--from python-numpy.spec

%if %is_redhat
BuildRequires:  gcc4-gfortran python python-devel lapack-devel > 3.0 
refblas3-devel
requires:       gcc4-gfortran lapack > 3.0 refblas3
%endif

"--record=INSTALLED_FILES" gave me an error when I tested manually so I 
also added:

%if %is_redhat
python setup.py install --root=$RPM_BUILD_ROOT \
        --prefix=/usr
%endif

and

%if %is_redhat
    %define NUMPY_EGG 0
%endif



David Cournapeau wrote:
> Fahd Sultan wrote:
>   
>> I had to recompile the rpms from your src rpms since the rpms dont 
>> install on Redhat AS 4.4.
>>     
> Mmmh, this distribution is not supported by the build system, so if 
> package convention differ, it may cause problems I see unnoticed, since 
> every rpm distribution decides it is a good idea to change conventions. 
> I assumed wrongly that you used fedora core, not the official RH thing.
>
> The distribution I am supporting for now are FC (5, 6, and 7) on x86 and 
> x86_64 and opensuse (both x86 and x86_64). That does not mean I am not 
> interested in improving the rpm for AS support, of course; it will just 
> be trickier because I do not have access to such a distribution, and 
> packaging is highly distribution dependent.
>   
>> during the build of python-numpy I get this error:
>> (any ideas?) 
>> I couldn't find an aborted install anywhere, not sure whats going on 
>> with this RPM.
>>
>>     
> Do you use 32 or 64 bits arch ? The problem seems to be related to the 
> library location. What would be useful for me is a complete log of the 
> rpm building (rpmbuild -ba python-numpy.spec &> build.log). Below are a 
> more detailed explanation on the problem:
>
> If you do not know anything about rpm packaging, here is the problem: 
> build rpm from sources involve a series of steps, the last one being 
> install, where you list the files to be installed. For example, all 
> python2.4 files are installed in /usr/lib/python2.4/site-packages/ 
> (/usr/lib64/python2.4/site-packages for 64 bits arch). You can see in 
> the python-numpy.spec file:
>
> %files
> %defattr(-,root,root,-)
> %{_bindir}/*
> # the following does not work on 64 bits arch.
> #%{py_sitedir}/*
> # I shamelessly copied the install from official fedora core numeric.
> %{_libdir}/python*/site-packages/numpy/
> # the egg is not produced on fedora 7 (something to do with different
> # configuration wrt setuptools on the build farm ?)
> %if 0%{?fedora_version}
>     %define NUMPY_EGG 0
> %else
>     %define NUMPY_EGG 1
> %endif
>
> %if %{NUMPY_EGG}
>     %{_libdir}/python*/site-packages/numpy*.egg-info
> %endif
> # Why the following are installed ???
> %{_libdir}/python*/site-packages/COMPATIBILITY
> %{_libdir}/python*/site-packages/scipy_compatibility
> %{_libdir}/python*/site-packages/site.cfg.example
>
> Now, in your case, it seems like there is nothing in
>
> /var/tmp/python-numpy-1.0.3-build/usr/lib64/python*/site-packages/numpy
>
> which is a bit strange... As I don't have access to redhat AS, it would be useful for me to know which files are where, to see if the fs layout is different.
>
> David
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>   


From david.warde.farley at utoronto.ca  Fri Jul 13 00:40:27 2007
From: david.warde.farley at utoronto.ca (David Warde-Farley)
Date: Fri, 13 Jul 2007 00:40:27 -0400
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
Message-ID: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>

Hi folks,

I've recently acquired myself a shiny new Mac and have vowed not to  
install Matlab on it, so I've been investigating going to SciPy  
entirely for my scientific computing needs. I installed MacPython 2.5  
and the SciPy Superpack listed on the SciPy website.

My problem is this: if I start IPython with ipython -pylab, and say,

plot(arange(0,9),arange(0,9)**2)

then close the plot window, and repeat the same command, IPython  
crashes completely. Is this normal behaviour? The obvious solution is  
"don't close plot windows" but that seems rather unsatisfactory.

Any ideas, hints, fixes, workarounds?

Thanks,

David


From david at ar.media.kyoto-u.ac.jp  Fri Jul 13 00:30:09 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Fri, 13 Jul 2007 13:30:09 +0900
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>
References: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>
Message-ID: <4696FFD1.9020504@ar.media.kyoto-u.ac.jp>

David Warde-Farley wrote:
> Hi folks,
>
> I've recently acquired myself a shiny new Mac and have vowed not to  
> install Matlab on it, so I've been investigating going to SciPy  
> entirely for my scientific computing needs. I installed MacPython 2.5  
> and the SciPy Superpack listed on the SciPy website.
>
> My problem is this: if I start IPython with ipython -pylab, and say,
>
> plot(arange(0,9),arange(0,9)**2)
>
> then close the plot window, and repeat the same command, IPython  
> crashes completely. Is this normal behaviour?
A crash is never a normal behaviour :)
>  The obvious solution is  
> "don't close plot windows" but that seems rather unsatisfactory.
>   
Indeed. The first thing would be to check whether the problem is 
matplotlib, ipython or scipy: If you repeat arange with the same 
arguments as given to the plot command, do you have any crash ? If you 
plot several times the same figure but in a script run by python instead 
of ipython, does it crash to ?

David


From david.warde.farley at utoronto.ca  Fri Jul 13 00:57:55 2007
From: david.warde.farley at utoronto.ca (David Warde-Farley)
Date: Fri, 13 Jul 2007 00:57:55 -0400
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <4696FFD1.9020504@ar.media.kyoto-u.ac.jp>
References: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>
	<4696FFD1.9020504@ar.media.kyoto-u.ac.jp>
Message-ID: <35104B3E-45F8-450F-A2AB-D9FCB80BEB16@utoronto.ca>

On 13-Jul-07, at 12:30 AM, David Cournapeau wrote:

> A crash is never a normal behaviour :)

Eh, by normal I mean "documented" I guess :)

> Indeed. The first thing would be to check whether the problem is
> matplotlib, ipython or scipy: If you repeat arange with the same
> arguments as given to the plot command, do you have any crash ? If you
> plot several times the same figure but in a script run by python  
> instead
> of ipython, does it crash to ?

Plotting several times in the same figure isn't a problem, even at  
the interpreter. The problem seems to be closing the plot window, and  
issuing a command that tries to communicate with it i.e. using clf()  
will cause a crash as well. I've been able to reproduce this at a  
normal python interpreter with

from pylab import *
ion() # interactive mode, accomplished with -pylab in ipython
plot(arange(0,9),arange(0,9)**2)
# Close the window when the plot appears
plot(arange(0,9),arange(0,9)) # or any other plotting command like clf 
(), I've also tried hist()

David


From fperez.net at gmail.com  Fri Jul 13 01:29:13 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Thu, 12 Jul 2007 23:29:13 -0600
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <35104B3E-45F8-450F-A2AB-D9FCB80BEB16@utoronto.ca>
References: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>
	<4696FFD1.9020504@ar.media.kyoto-u.ac.jp>
	<35104B3E-45F8-450F-A2AB-D9FCB80BEB16@utoronto.ca>
Message-ID: <db6b5ecc0707122229n4b5150d3xfb878ff198c94ab@mail.gmail.com>

On 7/12/07, David Warde-Farley <david.warde.farley at utoronto.ca> wrote:
> On 13-Jul-07, at 12:30 AM, David Cournapeau wrote:
>
> > A crash is never a normal behaviour :)
>
> Eh, by normal I mean "documented" I guess :)

No, not really.  We'll need a bit more info to try and figure out
what's going on:

1. Versions of IPython, Python and Matplotlib you're running.  I'm not
an OSX user, so I'm not exactly sure what the 'superpack' ships.

2. What mpl backend are you running?  And what's the version of the
underlying toolkit for python (tk, wx, pygtk, qt...)?

This should most certainly NOT be happening.

Cheers,

f


From faltet at carabos.com  Fri Jul 13 07:34:46 2007
From: faltet at carabos.com (Francesc Altet)
Date: Fri, 13 Jul 2007 13:34:46 +0200
Subject: [SciPy-user] [ANN] PyTables and PyTables Pro 2.0 released
Message-ID: <200707131334.47121.faltet@carabos.com>

========================================
 Announcing PyTables & PyTables Pro 2.0
========================================

PyTables is a library for managing hierarchical datasets and designed to
efficiently cope with extremely large amounts of data with support for
full 64-bit file addressing.  PyTables runs on top of the HDF5 library
and NumPy package for achieving maximum throughput and convenient use.

After more than one year of continuous development and about five months
of alpha, beta and release candidates, we are very happy to announce
that the PyTables and PyTables Pro 2.0 are here.  We are pretty
confident that the 2.0 versions are ready to be used in production
scenarios, bringing higher performance, better portability (specially in
64-bit environments) and more stability than the 1.x series.

You can download a source package of the PyTables 2.0 with generated PDF
and HTML docs and binaries for Windows from
http://www.pytables.org/download/stable/

For an on-line version of the manual, visit:
http://www.pytables.org/docs/manual-2.0

In case you want to know more in detail what has changed in this
version, have a look at ``RELEASE_NOTES.txt``.  Find the HTML version
for this document at:
http://www.pytables.org/moin/ReleaseNotes/Release_2.0

If you are a user of PyTables 1.x, probably it is worth for you to look
at ``MIGRATING_TO_2.x.txt`` file where you will find directions on how
to migrate your existing PyTables 1.x apps to the 2.0 version.  You can
find an HTML version of this document at
http://www.pytables.org/moin/ReleaseNotes/Migrating_To_2.x


Introducing PyTables Pro 2.0
============================

The main difference between PyTables Pro and regular PyTables is that
the Pro version includes OPSI, a new indexing technology, allowing to
perform data lookups in tables exceeding 10 gigarows (10**10 rows) in
less than 1 tenth of a second.  Wearing more than 15000 tests and having
passed the complete test suite in the most common platforms (Windows,
Mac OS X, Linux 32-bit and Linux 64-bit), we are pretty confident that
PyTables Pro 2.0 is ready to be used in production scenarios, bringing
maximum stability and top performance to those users who need it.
For more info about PyTables Pro, see:
http://www.carabos.com/products/pytables-pro
For the operational details and benchmarks see the OPSI white paper:
http://www.carabos.com/docs/OPSI-indexes.pdf

Coinciding with the publication of PyTables Pro we are introducing an
innovative liberation process that will allow to ultimate release the
PyTables Pro 2.x series as open source.  You may want to know that, by
buying a PyTables Pro license, you are contributing to this process. For
details, see: http://www.carabos.com/liberation


New features of PyTables 2.0 series
===================================

- A complete refactoring of many, many modules in PyTables.  With this,
  the different parts of the code are much better integrated and code
  redundancy is kept under a minimum.  A lot of new optimizations have
  been included as well, making working with it a smoother experience
  than ever before.

- NumPy is finally at the core!  That means that PyTables no longer
  needs numarray in order to operate, although it continues to be
  supported (as well as Numeric).  This also means that you should be
  able to run PyTables in scenarios combining Python 2.5 and 64-bit
  platforms (these are a source of problems with numarray/Numeric
  because they don't support this combination as of this writing).

- Most of the operations in PyTables have experimented noticeable
  speed-ups (sometimes up to 2x, like in regular Python table
  selections).  This is a consequence of both using NumPy internally and
  a considerable effort in terms of refactorization and optimization of
  the new code.

- Combined conditions are finally supported for in-kernel selections.
  So, now it is possible to perform complex selections like::

      result = [ row['var3'] for row in
                 table.where('(var2 < 20) | (var1 == "sas")') ]

  or::

      complex_cond = '((%s <= col5) & (col2 <= %s)) ' \
                     '| (sqrt(col1 + 3.1*col2 + col3*col4) > 3)'
      result = [ row['var3'] for row in
                 table.where(complex_cond % (inf, sup)) ]

  and run them at full C-speed (or perhaps more, due to the cache-tuned
  computing kernel of Numexpr, which has been integrated into PyTables).

- Now, it is possible to get fields of the ``Row`` iterator by
  specifying their position, or even ranges of positions (extended
  slicing is supported).  For example, you can do::

      result = [ row[4] for row in table    # fetch field #4
                 if row[1] < 20 ]
      result = [ row[:] for row in table    # fetch all fields
                 if row['var2'] < 20 ]
      result = [ row[1::2] for row in       # fetch odd fields
                 table.iterrows(2, 3000, 3) ]

  in addition to the classical::

      result = [row['var3'] for row in table.where('var2 < 20')]

- ``Row`` has received a new method called ``fetch_all_fields()`` in
  order to easily retrieve all the fields of a row in situations like::

      [row.fetch_all_fields() for row in table.where('column1 < 0.3')]

  The difference between ``row[:]`` and ``row.fetch_all_fields()`` is
  that the former will return all the fields as a tuple, while the
  latter will return the fields in a NumPy void type and should be
  faster.  Choose whatever fits better to your needs.

- Now, all data that is read from disk is converted, if necessary, to
  the native byteorder of the hosting machine (before, this only
  happened with ``Table`` objects).  This should help to accelerate
  applications that have to do computations with data generated in
  platforms with a byteorder different than the user machine.

- The modification of values in ``*Array`` objects (through __setitem__)
  now doesn't make a copy of the value in the case that the shape of the
  value passed is the same as the slice to be overwritten. This results
  in considerable memory savings when you are modifying disk objects
  with big array values.

- All leaf constructors (except for ``Array``) have received a new
  ``chunkshape`` argument that lets the user explicitly select the
  chunksizes for the underlying HDF5 datasets (only for advanced users).

- All leaf constructors have received a new parameter called
  ``byteorder`` that lets the user specify the byteorder of their data
  *on disk*.  This effectively allows to create datasets in other
  byteorders than the native platform.

- Native HDF5 datasets with ``H5T_ARRAY`` datatypes are fully supported
  for reading now.

- The test suites for the different packages are installed now, so you
  don't need a copy of the PyTables sources to run the tests.  Besides,
  you can run the test suite from the Python console by using::

  >>> tables.tests()


Resources
=========

Go to the PyTables web site for more details:
http://www.pytables.org/

Go to the PyTables Pro web page for more details:
http://www.carabos.com/products/pytables-pro

About the HDF5 library:
http://hdfgroup.org/HDF5/

About NumPy:
http://numpy.scipy.org/

To know more about the company behind the development of PyTables, see:
http://www.carabos.com/


Acknowledgments
===============

Thanks to many users who provided feature improvements, patches, bug
reports, support and suggestions.  See the ``THANKS`` file in the
distribution package for a (incomplete) list of contributors.  Many
thanks also to SourceForge who have helped to make and distribute this
package!  And last, but not least thanks a lot to the HDF5 and NumPy
(and numarray!) makers. Without them PyTables simply would not exist.


Share your experience
=====================

Let us know of any bugs, suggestions, gripes, kudos, etc. you may
have.


----

  **Enjoy data!**

  -- The PyTables Team

-- 
>0,0<   Francesc Altet ? ? http://www.carabos.com/
V   V   C?rabos Coop. V. ??Enjoy Data
 "-"


From david.warde.farley at utoronto.ca  Fri Jul 13 14:51:41 2007
From: david.warde.farley at utoronto.ca (David Warde-Farley)
Date: Fri, 13 Jul 2007 14:51:41 -0400
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <db6b5ecc0707122229n4b5150d3xfb878ff198c94ab@mail.gmail.com>
References: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>
	<4696FFD1.9020504@ar.media.kyoto-u.ac.jp>
	<35104B3E-45F8-450F-A2AB-D9FCB80BEB16@utoronto.ca>
	<db6b5ecc0707122229n4b5150d3xfb878ff198c94ab@mail.gmail.com>
Message-ID: <CC7BC2D9-60E1-4465-B274-54625681A3AA@utoronto.ca>


On 13-Jul-07, at 1:29 AM, Fernando Perez wrote:

> On 7/12/07, David Warde-Farley <david.warde.farley at utoronto.ca> wrote:
>> On 13-Jul-07, at 12:30 AM, David Cournapeau wrote:
>>
>>> A crash is never a normal behaviour :)
>>
>> Eh, by normal I mean "documented" I guess :)
>
> No, not really.  We'll need a bit more info to try and figure out
> what's going on:
>
> 1. Versions of IPython, Python and Matplotlib you're running.  I'm not
> an OSX user, so I'm not exactly sure what the 'superpack' ships.

IPython 0.8.2 svnr2445, Python 2.5.1(r251:54869),  
matplotlib.__version__ = 0.90.1

> 2. What mpl backend are you running?  And what's the version of the
> underlying toolkit for python (tk, wx, pygtk, qt...)?

This was with TkAgg. My Tkinter module says it's version is Rev. 50704.

> This should most certainly NOT be happening.

So, I installed wxPython 2.6 from http://pythonmac.org/packages/py25- 
fat/index.html and tried my luck with "backend : WXAgg" in  
~/.matplotlib/.matplotlibrc , and this behaviour does not occur  
(there is another small bug but it's far less annoying). So I'm  
guessing it's specific to the TK-based backends. The question is,  
does it happen on other platforms, or is it something to do with the  
Mac (or MacPython) bindings to Tk, or is it some unexpected condition  
that matplotlib isn't handling correctly. I think we can at least  
rule out IPython being at fault.

With the WXAgg backend, issuing a Cmd+Q (quit) command while a plot  
window is active will not only not close it but cause all further  
plot-related commands to lock up the interpreter. However, you can  
close plot windows and replot using the same figure # just fine using  
WXAgg, while it would crash you in TkAgg.

I should probably also note that I'm using the Intel version of the  
Superpack; the same issue might affect the PowerPC version, or it  
might not.

David



From robert.kern at gmail.com  Fri Jul 13 14:57:39 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 13 Jul 2007 13:57:39 -0500
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <CC7BC2D9-60E1-4465-B274-54625681A3AA@utoronto.ca>
References: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>	<4696FFD1.9020504@ar.media.kyoto-u.ac.jp>	<35104B3E-45F8-450F-A2AB-D9FCB80BEB16@utoronto.ca>	<db6b5ecc0707122229n4b5150d3xfb878ff198c94ab@mail.gmail.com>
	<CC7BC2D9-60E1-4465-B274-54625681A3AA@utoronto.ca>
Message-ID: <4697CB23.80305@gmail.com>

David Warde-Farley wrote:

> So, I installed wxPython 2.6 from http://pythonmac.org/packages/py25- 
> fat/index.html and tried my luck with "backend : WXAgg" in  
> ~/.matplotlib/.matplotlibrc , and this behaviour does not occur  
> (there is another small bug but it's far less annoying). So I'm  
> guessing it's specific to the TK-based backends. The question is,  
> does it happen on other platforms, or is it something to do with the  
> Mac (or MacPython) bindings to Tk, or is it some unexpected condition  
> that matplotlib isn't handling correctly. I think we can at least  
> rule out IPython being at fault.

This looks like a matplotlib problem rather than one with numpy or scipy. You
might have better luck asking on the matplotlib list.

  https://lists.sourceforge.net/lists/listinfo/matplotlib-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 fperez.net at gmail.com  Fri Jul 13 15:01:03 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Fri, 13 Jul 2007 13:01:03 -0600
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <4697CB23.80305@gmail.com>
References: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>
	<4696FFD1.9020504@ar.media.kyoto-u.ac.jp>
	<35104B3E-45F8-450F-A2AB-D9FCB80BEB16@utoronto.ca>
	<db6b5ecc0707122229n4b5150d3xfb878ff198c94ab@mail.gmail.com>
	<CC7BC2D9-60E1-4465-B274-54625681A3AA@utoronto.ca>
	<4697CB23.80305@gmail.com>
Message-ID: <db6b5ecc0707131201l7263d41cnac117fc67b4cc134@mail.gmail.com>

On 7/13/07, Robert Kern <robert.kern at gmail.com> wrote:

> This looks like a matplotlib problem rather than one with numpy or scipy. You
> might have better luck asking on the matplotlib list.
>
>   https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Indeed.  In fact, just this morning there was a thread form another
user seeing similar crashes, and they seem to have pinned it down to a
mistake in a recent commit that indeed was bad.

Have a look at the archives on matplotlib-devel for today with subject line:

Bus error using interactive plotting commands

It seems that by now they've committed a fix into SVN, so an update
may be all you need.

And if you have further issues, then Robert is right in suggesting
that this be continued on the mpl lists so that we don't create
unnecessary noise on the scipy ones.  It's important to keep each list
on topic for many reasons.

Cheers,

f


From david.warde.farley at utoronto.ca  Fri Jul 13 15:15:43 2007
From: david.warde.farley at utoronto.ca (David Warde-Farley)
Date: Fri, 13 Jul 2007 15:15:43 -0400
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <db6b5ecc0707131201l7263d41cnac117fc67b4cc134@mail.gmail.com>
References: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>
	<4696FFD1.9020504@ar.media.kyoto-u.ac.jp>
	<35104B3E-45F8-450F-A2AB-D9FCB80BEB16@utoronto.ca>
	<db6b5ecc0707122229n4b5150d3xfb878ff198c94ab@mail.gmail.com>
	<CC7BC2D9-60E1-4465-B274-54625681A3AA@utoronto.ca>
	<4697CB23.80305@gmail.com>
	<db6b5ecc0707131201l7263d41cnac117fc67b4cc134@mail.gmail.com>
Message-ID: <F50EEB0D-F4A9-4854-980A-8FDBDC65497C@utoronto.ca>

On 13-Jul-07, at 3:01 PM, Fernando Perez wrote:

> And if you have further issues, then Robert is right in suggesting
> that this be continued on the mpl lists so that we don't create
> unnecessary noise on the scipy ones.  It's important to keep each list
> on topic for many reasons.

Indeed, though I wasn't really sure where to start. I viewed it as a  
problem with the SciPy "environment", which sort of comprise IPython/ 
MPL/Numpy/Scipy/etc., so I posted here. For future reference, is the  
intended topic of this list only the packages that comprise the  
"scipy" tarball?

Regards,

David


From robert.kern at gmail.com  Fri Jul 13 15:22:01 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 13 Jul 2007 14:22:01 -0500
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <F50EEB0D-F4A9-4854-980A-8FDBDC65497C@utoronto.ca>
References: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>	<4696FFD1.9020504@ar.media.kyoto-u.ac.jp>	<35104B3E-45F8-450F-A2AB-D9FCB80BEB16@utoronto.ca>	<db6b5ecc0707122229n4b5150d3xfb878ff198c94ab@mail.gmail.com>	<CC7BC2D9-60E1-4465-B274-54625681A3AA@utoronto.ca>	<4697CB23.80305@gmail.com>	<db6b5ecc0707131201l7263d41cnac117fc67b4cc134@mail.gmail.com>
	<F50EEB0D-F4A9-4854-980A-8FDBDC65497C@utoronto.ca>
Message-ID: <4697D0D9.8020803@gmail.com>

David Warde-Farley wrote:
> On 13-Jul-07, at 3:01 PM, Fernando Perez wrote:
> 
>> And if you have further issues, then Robert is right in suggesting
>> that this be continued on the mpl lists so that we don't create
>> unnecessary noise on the scipy ones.  It's important to keep each list
>> on topic for many reasons.
> 
> Indeed, though I wasn't really sure where to start. I viewed it as a  
> problem with the SciPy "environment", which sort of comprise IPython/ 
> MPL/Numpy/Scipy/etc., so I posted here. For future reference, is the  
> intended topic of this list only the packages that comprise the  
> "scipy" tarball?

Not strictly, no. However, once we've narrowed down the problem to matplotlib,
there are more users and developers of matplotlib that hang out on the
matplotlib list to help you than there are here.

Don't worry about breaching etiquette; you didn't. It's perfectly acceptable to
start here. I'm just trying to get you the best help for your 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 fperez.net at gmail.com  Fri Jul 13 15:27:31 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Fri, 13 Jul 2007 13:27:31 -0600
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <4697D0D9.8020803@gmail.com>
References: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>
	<4696FFD1.9020504@ar.media.kyoto-u.ac.jp>
	<35104B3E-45F8-450F-A2AB-D9FCB80BEB16@utoronto.ca>
	<db6b5ecc0707122229n4b5150d3xfb878ff198c94ab@mail.gmail.com>
	<CC7BC2D9-60E1-4465-B274-54625681A3AA@utoronto.ca>
	<4697CB23.80305@gmail.com>
	<db6b5ecc0707131201l7263d41cnac117fc67b4cc134@mail.gmail.com>
	<F50EEB0D-F4A9-4854-980A-8FDBDC65497C@utoronto.ca>
	<4697D0D9.8020803@gmail.com>
Message-ID: <db6b5ecc0707131227xf8ebd08n3e474bf8c81b919a@mail.gmail.com>

On 7/13/07, Robert Kern <robert.kern at gmail.com> wrote:
> David Warde-Farley wrote:
> > On 13-Jul-07, at 3:01 PM, Fernando Perez wrote:
> >
> >> And if you have further issues, then Robert is right in suggesting
> >> that this be continued on the mpl lists so that we don't create
> >> unnecessary noise on the scipy ones.  It's important to keep each list
> >> on topic for many reasons.
> >
> > Indeed, though I wasn't really sure where to start. I viewed it as a
> > problem with the SciPy "environment", which sort of comprise IPython/
> > MPL/Numpy/Scipy/etc., so I posted here. For future reference, is the
> > intended topic of this list only the packages that comprise the
> > "scipy" tarball?
>
> Not strictly, no. However, once we've narrowed down the problem to matplotlib,
> there are more users and developers of matplotlib that hang out on the
> matplotlib list to help you than there are here.
>
> Don't worry about breaching etiquette; you didn't. It's perfectly acceptable to
> start here. I'm just trying to get you the best help for your problem.


And sorry if it sounded like I was chastising you (David), that was
most certainly NOT my intention!  I fully agree with Robert that it
was perfectly reasonable to start here, I just worded poorly my
suggestion to continue on the mpl list at this point in the bug hunt.

I really don't want to appear as creating a hostile atmosphere for
anyone to report bugs or ask for help when they aren't 100% sure of
where the problem is coming from (ipython/numpy/scipy/mpl/etc).

Regards,

f


From david.warde.farley at utoronto.ca  Fri Jul 13 15:36:59 2007
From: david.warde.farley at utoronto.ca (David Warde-Farley)
Date: Fri, 13 Jul 2007 15:36:59 -0400
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <db6b5ecc0707131227xf8ebd08n3e474bf8c81b919a@mail.gmail.com>
References: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>
	<4696FFD1.9020504@ar.media.kyoto-u.ac.jp>
	<35104B3E-45F8-450F-A2AB-D9FCB80BEB16@utoronto.ca>
	<db6b5ecc0707122229n4b5150d3xfb878ff198c94ab@mail.gmail.com>
	<CC7BC2D9-60E1-4465-B274-54625681A3AA@utoronto.ca>
	<4697CB23.80305@gmail.com>
	<db6b5ecc0707131201l7263d41cnac117fc67b4cc134@mail.gmail.com>
	<F50EEB0D-F4A9-4854-980A-8FDBDC65497C@utoronto.ca>
	<4697D0D9.8020803@gmail.com>
	<db6b5ecc0707131227xf8ebd08n3e474bf8c81b919a@mail.gmail.com>
Message-ID: <295CC612-77F6-4243-B80B-5910ACD4A435@utoronto.ca>

On 13-Jul-07, at 3:27 PM, Fernando Perez wrote:

> I really don't want to appear as creating a hostile atmosphere for
> anyone to report bugs or ask for help when they aren't 100% sure of
> where the problem is coming from (ipython/numpy/scipy/mpl/etc).

Not a problem; I just honestly wasn't sure. There is a bit of a fuzzy  
aura surrounding what 'scipy' is or isn't in my head, as I'm sure is  
the case with many.

It does seem as though the *exact* same bug was reported today on  
matplotlib-devel (what are the odds?), I shall checkout the svn  
version and see if that fixes it, and make them aware of the other WX- 
related bug as well.

Many thanks,

David


From jdh2358 at gmail.com  Fri Jul 13 16:10:35 2007
From: jdh2358 at gmail.com (John Hunter)
Date: Fri, 13 Jul 2007 15:10:35 -0500
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <295CC612-77F6-4243-B80B-5910ACD4A435@utoronto.ca>
References: <D16E908E-CEEB-486F-9D80-D350233494BD@utoronto.ca>
	<35104B3E-45F8-450F-A2AB-D9FCB80BEB16@utoronto.ca>
	<db6b5ecc0707122229n4b5150d3xfb878ff198c94ab@mail.gmail.com>
	<CC7BC2D9-60E1-4465-B274-54625681A3AA@utoronto.ca>
	<4697CB23.80305@gmail.com>
	<db6b5ecc0707131201l7263d41cnac117fc67b4cc134@mail.gmail.com>
	<F50EEB0D-F4A9-4854-980A-8FDBDC65497C@utoronto.ca>
	<4697D0D9.8020803@gmail.com>
	<db6b5ecc0707131227xf8ebd08n3e474bf8c81b919a@mail.gmail.com>
	<295CC612-77F6-4243-B80B-5910ACD4A435@utoronto.ca>
Message-ID: <88e473830707131310w57f2c3d3lce09a2f481c48c80@mail.gmail.com>

On 7/13/07, David Warde-Farley <david.warde.farley at utoronto.ca> wrote:

> It does seem as though the *exact* same bug was reported today on
> matplotlib-devel (what are the odds?), I shall checkout the svn
> version and see if that fixes it, and make them aware of the other WX-
> related bug as well.

I hope this does fix your problem, but the bug that was fixed was also
recently introduced (in a svn commit after the 0.90.1 release) so it
may not be your problem.  But if you can update from svn and see if
the problem is still there, that would be a great start.  See you on
the matplotlib-devel side :-)

JDH


From novak at ucolick.org  Sat Jul 14 19:31:59 2007
From: novak at ucolick.org (Greg Novak)
Date: Sat, 14 Jul 2007 16:31:59 -0700
Subject: [SciPy-user] vectorizing methods?
Message-ID: <ad0d4fcf0707141631r1d96702ej3586242912d47347@mail.gmail.com>

Is there a preferred way to vectorize methods, rather than functions?
The obvious thing doesn't work:

class foo:
    def bar(self, x): pass
    bar = vectorize(bar)

I came up with this:

class foo:
    __vectorMethods = ('bar', )

    def __init__(self, n):
        for name in self.__vectorMethods:
            setattr(self, name, vectorize(getattr(self, name)))

    def bar(self, x): pass

which works fine but it doesn't seem like it should be necessary to do
it in the initialization of every instance.

Thanks,
Greg


From gael.varoquaux at normalesup.org  Sun Jul 15 05:32:44 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Sun, 15 Jul 2007 11:32:44 +0200
Subject: [SciPy-user] vectorizing methods?
In-Reply-To: <ad0d4fcf0707141631r1d96702ej3586242912d47347@mail.gmail.com>
References: <ad0d4fcf0707141631r1d96702ej3586242912d47347@mail.gmail.com>
Message-ID: <20070715093244.GA28109@clipper.ens.fr>

On Sat, Jul 14, 2007 at 04:31:59PM -0700, Greg Novak wrote:
> Is there a preferred way to vectorize methods, rather than functions?
> The obvious thing doesn't work:

> class foo:
>     def bar(self, x): pass
>     bar = vectorize(bar)

This cannot work, as you are applying "vectorize" before the class is
built, so the function is still unbound. The binding process, which
happens during the building of the class, will not like what you have
done to the function to vectorize it, mainly the fact that it is no
longer a plain function.


> class foo:
>     __vectorMethods = ('bar', )

>     def __init__(self, n):
>         for name in self.__vectorMethods:
>             setattr(self, name, vectorize(getattr(self, name)))

>     def bar(self, x): pass

There you are vectorizing after the construction of the instance, so the
method is already bound.

> which works fine but it doesn't seem like it should be necessary to do
> it in the initialization of every instance.

Well, if your methods don't really refer to the object (they don't use
self at all in the body of the function), you could do something like:

class Foo(object):
    @staticmethod
    @vectorize
    def bar(x):
        return x


In the static method there is no reference to the object. The problem is
that when the class is binding the method (making a call in which the
first argument is already assigned to the instance) it is really
modifying it. If you could apply vectorize to only part of the argument,
and get a curry of the original function you could than transform 
bar(self, *args) in baz(self, *args), where *args can be vectors, and
this might work.

Is can see two ways out of this:

   1) Do some heavy magic with a metaclass and modify the binding process
      (I am talking about things I don't know, so it may not be
      feasible).

 2.a) Use a static method that you vectorize, and pass it the instance as
      the first argument using a convention bound method:

	class Foo(object):
	    @staticmethod
	    @vectorize
	    def bar(self, x):
		return self.a + x
		
	    def baz(self, x):
		return self.bar(self, x)

	    a = 1

 2.b) IMHO it would be even cleaner to only pass interesting vectors to
      the vectorize staticmethod, as it has been vectorized relatively to
      "self", but self cannot be a vector:

	class Foo(object):
	    @staticmethod
	    @vectorize
	    def bar(a, x):
		return a + x
		
	    def baz(self, x):
		return self.bar(self.a, x)

	    a = 1

      Basically this is similar to using a standard function instead of a
      method, but lets face it, the notions of bound method and
      vectorized function seem mutually incompatible as they both want to
      modify a function.

It might be possible to implement a decorator similar to staticmethod
that does the proces presented in 2.a) but hiddes it from the user, but
to do this, you would need to understand how staticmethod works, and I
suspect there is some cooperation of the class building mechanisme (so we
are back to 1), I think).

I hope 2.a) or 2.b) suit you, elsewhere maybe some one has a better idea ?

Cheers,

Ga?l


From travis at enthought.com  Sun Jul 15 13:40:42 2007
From: travis at enthought.com (Travis Vaught)
Date: Sun, 15 Jul 2007 12:40:42 -0500
Subject: [SciPy-user] ANN: SciPy 2007 Conference Updates
Message-ID: <95D2E93A-0102-4BBF-BEDA-2BAE0CC20654@enthought.com>

Greetings,

We're excited to have *Ivan Krsti?*, the director of security  
architecture for the One Laptop Per Child project as our Keynote  
Speaker this year.

The planning for the SciPy 2007 Conference is moving along.  Please  
see below for some important updates.

Schedule Available
------------------
The full schedule of talks has been posted here:
http://www.scipy.org/SciPy2007/ConferenceSchedule


Early Registration Extended
---------------------------
If you haven't yet registered for the conference, the early  
registration deadline has been extended to Wednesday, July 18th,  
2007.  For more information on the conference see:
http://www.scipy.org/SciPy2007


Student Sponsorship
-------------------
Enthought, Inc. (http://www.enthought.com) is sponsoring the  
registration fees for up to 5 college or graduate students to attend  
the conference.  To apply, please send a short description of what  
you are studying and why you'd like to attend to info at enthought.com.   
Please include telephone contact information.


BOFs & Sprints
--------------
If you're planning to attend and are interested in selecting BOF or  
Sprint Session topics, please weigh in at:
BOFs: http://www.scipy.org/SciPy2007/BoFs
Sprints: http://www.scipy.org/SciPy2007/Sprints

We're looking forward to a great conference this year!

Best,

Travis

From lev at columbia.edu  Sun Jul 15 15:31:47 2007
From: lev at columbia.edu (Lev Givon)
Date: Sun, 15 Jul 2007 15:31:47 -0400
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <20070705001420.GB30029@localhost.ee.columbia.edu>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>
	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>
	<20070705001420.GB30029@localhost.ee.columbia.edu>
Message-ID: <20070715193147.GA19666@avicenna.cc.columbia.edu>

Received from Lev Givon on Wed, Jul 04, 2007 at 08:14:21PM EDT:

(snip)
 
> Insofar as the item that prompted my first message is concerned, I
> desire a filter design function similar to others currently
> available in scipy to facilitate the construction of FIR filters
> with the Remez algorithm (i.e., something similar to the remezord or
> firpmord functions in Matlab) given some typical design parameters
> (e.g., band cutoff frequencies, ripple constraints, etc.). As the
> relevant algorithms needed by such a filter design function are not
> overly complicated, I could look them up an appropriate DSP text or
> paper and try implementing them completely from scratch (unless some
> generous soul reading this list has already done so :-)
> 
>    	       	   	    		     	  L.G.

I finished implementing and testing the aforementioned filter design
algorithms. May I add the code directly to Lib/signal/filter_design.py
(where similar functions are defined)? Or can I create my own corner
in the sandbox and put the functions in there for folks to examine and
test before they are moved into the aforementioned file?

						L.G.



From robert.kern at gmail.com  Sun Jul 15 17:35:09 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Sun, 15 Jul 2007 16:35:09 -0500
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <20070715193147.GA19666@avicenna.cc.columbia.edu>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>	<20070705001420.GB30029@localhost.ee.columbia.edu>
	<20070715193147.GA19666@avicenna.cc.columbia.edu>
Message-ID: <469A930D.4000509@gmail.com>

Lev Givon wrote:
> Received from Lev Givon on Wed, Jul 04, 2007 at 08:14:21PM EDT:
> 
> (snip)
>  
>> Insofar as the item that prompted my first message is concerned, I
>> desire a filter design function similar to others currently
>> available in scipy to facilitate the construction of FIR filters
>> with the Remez algorithm (i.e., something similar to the remezord or
>> firpmord functions in Matlab) given some typical design parameters
>> (e.g., band cutoff frequencies, ripple constraints, etc.). As the
>> relevant algorithms needed by such a filter design function are not
>> overly complicated, I could look them up an appropriate DSP text or
>> paper and try implementing them completely from scratch (unless some
>> generous soul reading this list has already done so :-)
>>
>>    	       	   	    		     	  L.G.
> 
> I finished implementing and testing the aforementioned filter design
> algorithms. May I add the code directly to Lib/signal/filter_design.py
> (where similar functions are defined)? Or can I create my own corner
> in the sandbox and put the functions in there for folks to examine and
> test before they are moved into the aforementioned file?

To clarify, did you implement them from scratch, or did you use the Matlab
source? The final location should be filter_design.py, but please submit it in
the form of a patch to the Trac. Thanks.

-- 
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 s.mientki at mailbox.kun.nl  Sun Jul 15 18:03:00 2007
From: s.mientki at mailbox.kun.nl (stef mientki)
Date: Mon, 16 Jul 2007 00:03:00 +0200
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <469A930D.4000509@gmail.com>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>	<20070705001420.GB30029@localhost.ee.columbia.edu>	<20070715193147.GA19666@avicenna.cc.columbia.edu>
	<469A930D.4000509@gmail.com>
Message-ID: <469A9994.5040503@gmail.com>

Robert Kern wrote:
> Lev Givon wrote:
>   
>> Received from Lev Givon on Wed, Jul 04, 2007 at 08:14:21PM EDT:
>>
>> (snip)
>>  
>>     
>>> Insofar as the item that prompted my first message is concerned, I
>>> desire a filter design function similar to others currently
>>> available in scipy to facilitate the construction of FIR filters
>>> with the Remez algorithm (i.e., something similar to the remezord or
>>> firpmord functions in Matlab) given some typical design parameters
>>> (e.g., band cutoff frequencies, ripple constraints, etc.). As the
>>> relevant algorithms needed by such a filter design function are not
>>> overly complicated, I could look them up an appropriate DSP text or
>>> paper and try implementing them completely from scratch (unless some
>>> generous soul reading this list has already done so :-)
>>>
>>>    	       	   	    		     	  L.G.
>>>       
very good, although not very efficient, FIR filters are my favorits.
>> I finished implementing and testing the aforementioned filter design
>> algorithms. May I add the code directly to Lib/signal/filter_design.py
>> (where similar functions are defined)? Or can I create my own corner
>> in the sandbox and put the functions in there for folks to examine and
>> test before they are moved into the aforementioned file?
>>     
>
> To clarify, did you implement them from scratch, or did you use the Matlab
> source?
Although I don't know the MatLab sources,
but I bet MatLab used th? Fortran source code already known for many 
decades ;-)
For the record: The Remez algorithm (Remez 1934),
and Parks and McClellan (1972) were the first who described and 
implemented the algorithm.

cheers,
Stef Mientki




From lev at columbia.edu  Sun Jul 15 18:46:19 2007
From: lev at columbia.edu (Lev Givon)
Date: Sun, 15 Jul 2007 18:46:19 -0400
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <469A930D.4000509@gmail.com>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>
	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>
	<20070705001420.GB30029@localhost.ee.columbia.edu>
	<20070715193147.GA19666@avicenna.cc.columbia.edu>
	<469A930D.4000509@gmail.com>
Message-ID: <20070715224619.GA23329@avicenna.cc.columbia.edu>

Received from Robert Kern on Sun, Jul 15, 2007 at 05:35:09PM EDT:
> Lev Givon wrote:
> > Received from Lev Givon on Wed, Jul 04, 2007 at 08:14:21PM EDT:
> > 
> > (snip)
> >  
> >> Insofar as the item that prompted my first message is concerned, I
> >> desire a filter design function similar to others currently
> >> available in scipy to facilitate the construction of FIR filters
> >> with the Remez algorithm (i.e., something similar to the remezord or
> >> firpmord functions in Matlab) given some typical design parameters
> >> (e.g., band cutoff frequencies, ripple constraints, etc.). As the
> >> relevant algorithms needed by such a filter design function are not
> >> overly complicated, I could look them up an appropriate DSP text or
> >> paper and try implementing them completely from scratch (unless some
> >> generous soul reading this list has already done so :-)
> >>
> >>    	       	   	    		     	  L.G.
> > 
> > I finished implementing and testing the aforementioned filter design
> > algorithms. May I add the code directly to Lib/signal/filter_design.py
> > (where similar functions are defined)? Or can I create my own corner
> > in the sandbox and put the functions in there for folks to examine and
> > test before they are moved into the aforementioned file?
> 
> To clarify, did you implement them from scratch, or did you use the
> Matlab source? The final location should be filter_design.py, but
> please submit it in the form of a patch to the Trac. Thanks.
> 
> -- 
> Robert Kern

I implemented them by reading a relatively recent paper on FIR filter
length estimation that describes several such algorithms; the default
algorithm is ostensibly superior to that employed in Matlab. I'll
include a reference to the paper in the source code.

							L.G.


From lev at columbia.edu  Sun Jul 15 18:54:36 2007
From: lev at columbia.edu (Lev Givon)
Date: Sun, 15 Jul 2007 18:54:36 -0400
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <469A9994.5040503@gmail.com>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>
	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>
	<20070705001420.GB30029@localhost.ee.columbia.edu>
	<20070715193147.GA19666@avicenna.cc.columbia.edu>
	<469A930D.4000509@gmail.com> <469A9994.5040503@gmail.com>
Message-ID: <20070715225436.GB23329@avicenna.cc.columbia.edu>

Received from stef mientki on Sun, Jul 15, 2007 at 06:03:00PM EDT:
> Robert Kern wrote:
> > Lev Givon wrote:
> >   
> >> Received from Lev Givon on Wed, Jul 04, 2007 at 08:14:21PM EDT:
> >>
> >> (snip)
> >>  
> >>     
> >>> Insofar as the item that prompted my first message is concerned, I
> >>> desire a filter design function similar to others currently
> >>> available in scipy to facilitate the construction of FIR filters
> >>> with the Remez algorithm (i.e., something similar to the remezord or
> >>> firpmord functions in Matlab) given some typical design parameters
> >>> (e.g., band cutoff frequencies, ripple constraints, etc.). As the
> >>> relevant algorithms needed by such a filter design function are not
> >>> overly complicated, I could look them up an appropriate DSP text or
> >>> paper and try implementing them completely from scratch (unless some
> >>> generous soul reading this list has already done so :-)
> >>>
> >>>    	       	   	    		     	  L.G.
> >>>       
> very good, although not very efficient, FIR filters are my favorits.
> >> I finished implementing and testing the aforementioned filter design
> >> algorithms. May I add the code directly to Lib/signal/filter_design.py
> >> (where similar functions are defined)? Or can I create my own corner
> >> in the sandbox and put the functions in there for folks to examine and
> >> test before they are moved into the aforementioned file?
> >>     
> >
> > To clarify, did you implement them from scratch, or did you use the Matlab
> > source?
> Although I don't know the MatLab sources,
> but I bet MatLab used th? Fortran source code already known for many 
> decades ;-)
> For the record: The Remez algorithm (Remez 1934),
> and Parks and McClellan (1972) were the first who described and 
> implemented the algorithm.
> 
> cheers,
> Stef Mientki

Matlab uses an empirically determined approximation algorithm
developed by Herrmann and others in the 1970s. The paper I consulted
while writing my code describes Herrmann's method, a method developed
by Kaiser, and a third that supposedly is better than either. For the
sake of comparison, I implemented all three :-)

							L.G.






From robert.kern at gmail.com  Sun Jul 15 19:21:55 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Sun, 15 Jul 2007 18:21:55 -0500
Subject: [SciPy-user] scipy code contribution question
In-Reply-To: <20070715224619.GA23329@avicenna.cc.columbia.edu>
References: <20070704205625.GC28755@localhost.ee.columbia.edu>	<1b5a37350707041533q1db2d939oe95de76f01782455@mail.gmail.com>	<20070705001420.GB30029@localhost.ee.columbia.edu>	<20070715193147.GA19666@avicenna.cc.columbia.edu>	<469A930D.4000509@gmail.com>
	<20070715224619.GA23329@avicenna.cc.columbia.edu>
Message-ID: <469AAC13.4060909@gmail.com>

Lev Givon wrote:

> I implemented them by reading a relatively recent paper on FIR filter
> length estimation that describes several such algorithms; the default
> algorithm is ostensibly superior to that employed in Matlab. I'll
> include a reference to the paper in the source code.

Fantastic! I look forward to seeing 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 amcmorl at gmail.com  Mon Jul 16 01:45:27 2007
From: amcmorl at gmail.com (Angus McMorland)
Date: Mon, 16 Jul 2007 17:45:27 +1200
Subject: [SciPy-user] Numerical Recipes robust fit implementation
Message-ID: <e85f13c40707152245l57a23b99x9a13045be6c79d66@mail.gmail.com>

Hi all,

I'm in need of a robust linear-fit algorithm- Numerical Recipes (in C)
medfit routine will do nicely. I've had a quick look and don't see any
similar routines in scipy. Is there one I'm missing?

If not, I'd lke to implement the NR routine for use with my
numpy-arrayed data. What are my options? Obviously I could translate
the routine into python, but I realise a potentially easier
alternative is to interface the existing NR C code. Having never done
that before, can someone give me some pointers/tips on how to start -
there's quite a range of linking tools listed on
http://www.scipy.org/Topical_Software.

Thanks,

Angus.
-- 
AJC McMorland, PhD Student
Physiology, University of Auckland


From strawman at astraw.com  Mon Jul 16 02:37:03 2007
From: strawman at astraw.com (Andrew Straw)
Date: Sun, 15 Jul 2007 23:37:03 -0700
Subject: [SciPy-user] Numerical Recipes robust fit implementation
In-Reply-To: <e85f13c40707152245l57a23b99x9a13045be6c79d66@mail.gmail.com>
References: <e85f13c40707152245l57a23b99x9a13045be6c79d66@mail.gmail.com>
Message-ID: <469B120F.6000006@astraw.com>

Hi Angus,

I don't know if the MEDFIT algorithm is something you really want, but I
wrote a simple implementation of the RANSAC algorithm:
http://www.scipy.org/Cookbook/RANSAC

This is not to say I recommend the use of such algorithms as RANSAC! :)

-Andrew

PS There seems to be some criticism of the MEDFIT algorithm in general
and as implemented in NR in particular online.

Angus McMorland wrote:
> Hi all,
> 
> I'm in need of a robust linear-fit algorithm- Numerical Recipes (in C)
> medfit routine will do nicely. I've had a quick look and don't see any
> similar routines in scipy. Is there one I'm missing?
> 
> If not, I'd lke to implement the NR routine for use with my
> numpy-arrayed data. What are my options? Obviously I could translate
> the routine into python, but I realise a potentially easier
> alternative is to interface the existing NR C code. Having never done
> that before, can someone give me some pointers/tips on how to start -
> there's quite a range of linking tools listed on
> http://www.scipy.org/Topical_Software.
> 
> Thanks,
> 
> Angus.



From amcmorl at gmail.com  Mon Jul 16 07:01:16 2007
From: amcmorl at gmail.com (Angus McMorland)
Date: Mon, 16 Jul 2007 23:01:16 +1200
Subject: [SciPy-user] Numerical Recipes robust fit implementation
In-Reply-To: <469B120F.6000006@astraw.com>
References: <e85f13c40707152245l57a23b99x9a13045be6c79d66@mail.gmail.com>
	<469B120F.6000006@astraw.com>
Message-ID: <e85f13c40707160401l30543ff9qd1e67a4341d1d3ee@mail.gmail.com>

Hi Andrew et al.,

> Angus McMorland wrote:
> > I'm in need of a robust linear-fit algorithm- Numerical Recipes (in C)
> > medfit routine will do nicely. I've had a quick look and don't see any
> > similar routines in scipy. Is there one I'm missing?

On 16/07/07, Andrew Straw <strawman at astraw.com> wrote:
> I don't know if the MEDFIT algorithm is something you really want, but I
> wrote a simple implementation of the RANSAC algorithm:
> http://www.scipy.org/Cookbook/RANSAC
>
> This is not to say I recommend the use of such algorithms as RANSAC! :)
>
> -Andrew
>
> PS There seems to be some criticism of the MEDFIT algorithm in general
> and as implemented in NR in particular online.

Thanks for the heads up on those issues. I've had a quick google on
this myself now and found this
(http://www.lysator.liu.se/c/num-recipes-in-c.html) comment verbatim
several times on different websites  - someincluding mention that the
discussion refers to the first edition and some not. According to NR
(http://www.nr.com/bug-rebutt.htm - seems a little melodramatically
worded in places) in the second edition they've addressed issues
raised about medfit.

I'm going to have to think a bit more what I want to achieve to see if
RANSAC is useful. Ultimately I hope to determine the probability of a
a given data set being exponentially distributed, by comparing the raw
frequency distribution to an expected distribution based on a linear
fit to the log transform of the raw one. It seems a bit like basing my
'expected' distribution on a subset of data from which outliers have
been completely excluded is self-fulfilling, and having some other
criterion for weighting of the error term (as medfit does) seems more
appropriate. This is however very much just a gut feeling rather than
an educated assessment, so any other comments are welcome.

It certainly is useful to know the RANSAC algorithm exists, so thanks for that.

Thanks again,

Angus.
-- 
AJC McMorland, PhD Student
Physiology, University of Auckland


From matthieu.brucher at gmail.com  Mon Jul 16 07:30:32 2007
From: matthieu.brucher at gmail.com (Matthieu Brucher)
Date: Mon, 16 Jul 2007 13:30:32 +0200
Subject: [SciPy-user] Numerical Recipes robust fit implementation
In-Reply-To: <e85f13c40707160401l30543ff9qd1e67a4341d1d3ee@mail.gmail.com>
References: <e85f13c40707152245l57a23b99x9a13045be6c79d66@mail.gmail.com>
	<469B120F.6000006@astraw.com>
	<e85f13c40707160401l30543ff9qd1e67a4341d1d3ee@mail.gmail.com>
Message-ID: <e76aa17f0707160430t7d20eca7kbec2a8ff27238099@mail.gmail.com>

Hi,


Ultimately I hope to determine the probability of a
> a given data set being exponentially distributed,




In this case, you could use a non quadratic fit function, with a pseudo
norm1 cost ( for instance sqrt(eps + (y-f(x))?) with a small eps so that the
cost is derivable and a gradient descent should do the trick).

Matthieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070716/fb53a118/attachment.html>

From openopt at ukr.net  Mon Jul 16 07:41:06 2007
From: openopt at ukr.net (dmitrey)
Date: Mon, 16 Jul 2007 14:41:06 +0300
Subject: [SciPy-user] Numerical Recipes robust fit implementation
In-Reply-To: <e76aa17f0707160430t7d20eca7kbec2a8ff27238099@mail.gmail.com>
References: <e85f13c40707152245l57a23b99x9a13045be6c79d66@mail.gmail.com>	<469B120F.6000006@astraw.com>	<e85f13c40707160401l30543ff9qd1e67a4341d1d3ee@mail.gmail.com>
	<e76aa17f0707160430t7d20eca7kbec2a8ff27238099@mail.gmail.com>
Message-ID: <469B5952.6050100@ukr.net>

my 2 cents:
openopt ralg solver works good to (currently unconstrained only) 
problems like F(x) = sum(abs(f[i](x))) (or other non-smooth funcs; 
convexity is highly preferable)
and it can use F(x) 1st derive if provided by user

however, ralg is intended for medium-scale problems only (nVars up to 
1000), and it is missing patterns yet, I guess it could be nice for your 
problem

HTH, D.

Matthieu Brucher wrote:
> Hi,
>  
>
>     Ultimately I hope to determine the probability of a
>     a given data set being exponentially distributed, 
>
>
>
>
> In this case, you could use a non quadratic fit function, with a 
> pseudo norm1 cost ( for instance sqrt(eps + (y-f(x))?) with a small 
> eps so that the cost is derivable and a gradient descent should do the 
> trick).
>
> Matthieu
> ------------------------------------------------------------------------
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   



From peridot.faceted at gmail.com  Mon Jul 16 07:54:28 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Mon, 16 Jul 2007 07:54:28 -0400
Subject: [SciPy-user] Numerical Recipes robust fit implementation
In-Reply-To: <e85f13c40707160401l30543ff9qd1e67a4341d1d3ee@mail.gmail.com>
References: <e85f13c40707152245l57a23b99x9a13045be6c79d66@mail.gmail.com>
	<469B120F.6000006@astraw.com>
	<e85f13c40707160401l30543ff9qd1e67a4341d1d3ee@mail.gmail.com>
Message-ID: <ce557a360707160454m4ef59a33pe50ad72e8255583c@mail.gmail.com>

On 16/07/07, Angus McMorland <amcmorl at gmail.com> wrote:

> I'm going to have to think a bit more what I want to achieve to see if
> RANSAC is useful. Ultimately I hope to determine the probability of a
> a given data set being exponentially distributed, by comparing the raw
> frequency distribution to an expected distribution based on a linear
> fit to the log transform of the raw one. It seems a bit like basing my
> 'expected' distribution on a subset of data from which outliers have
> been completely excluded is self-fulfilling, and having some other
> criterion for weighting of the error term (as medfit does) seems more
> appropriate. This is however very much just a gut feeling rather than
> an educated assessment, so any other comments are welcome.

If what you're trying to do is test whether your data points are
likely to have been drawn from a given distribution, you may be able
to do much better by not putting them in a histogram first, and using
something like the Kolmogorov-Smirnov test (scipy.stats.kstest). If
you have outliers you may have a problem. (It's feasible to fit
parameters so they maximize the kstest p value, although of course the
p value you get at the end is no longer an actual probability.) I
suspect if you look in the statistical literature you'll find other
tricks for fitting distributions directly (rather than fitting
histograms).

Anne


From fdu.xiaojf at gmail.com  Mon Jul 16 08:18:10 2007
From: fdu.xiaojf at gmail.com (fdu.xiaojf at gmail.com)
Date: Mon, 16 Jul 2007 20:18:10 +0800
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
Message-ID: <469B6202.6020002@gmail.com>

Hi all,

I'm trying to minimize a function f(X) = f(x1,x2,...,xn) using cobyla.

In my minimization, all variables should be larger than 0, so I define
constraint functions like this:

cons = []
for i in range(n): # n is the number of variables
	c =lambda x: x[i] - 1e-10
	cons.append(c)

And then I use fmin_cobyla(func=f, x0=[0.1]*n, cons=cons) to minimize the
function.

There are two problems:

1) In the constrain functions, I expected the value of i to be bounded
to the specific constrain function, but it doesn't work as I expected.

Here is a demonstration:
In [14]: a = []

In [15]: for i in range(4):
     b = lambda : i**2
     a.append(b)
    ....:
    ....:

In [18]: for f in a:
     print f()
    ....:
    ....:
9
9
9
9

What I want is that every function in list a should print different
values. The value of a[0](), a[1](), a[2](), a[3]() should be 0, 1, 4,
9.

How to achieve this?

2) In function f(X), there are item of log(xi). Although I have
constrained the value of xi in the constrain functions, xi still may be
equal or less than 0 during minimization, then "math domain error"
occurred.

What should I do ?

My method is when "math domain error" occurred, catch it and set the
return value of f to a very large number. Should this work or not?



Thanks a lot!

Xiao Jianfeng


From openopt at ukr.net  Mon Jul 16 08:30:39 2007
From: openopt at ukr.net (dmitrey)
Date: Mon, 16 Jul 2007 15:30:39 +0300
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469B6202.6020002@gmail.com>
References: <469B6202.6020002@gmail.com>
Message-ID: <469B64EF.2080401@ukr.net>

fdu.xiaojf at gmail.com wrote:
> Hi all,
>
> I'm trying to minimize a function f(X) = f(x1,x2,...,xn) using cobyla.
>
> In my minimization, all variables should be larger than 0, so I define
> constraint functions like this:
>
> cons = []
> for i in range(n): # n is the number of variables
> 	c =lambda x: x[i] - 1e-10
> 	cons.append(c)
>
> And then I use fmin_cobyla(func=f, x0=[0.1]*n, cons=cons) to minimize the
> function.
>
> There are two problems:
>
> 1) In the constrain functions, I expected the value of i to be bounded
> to the specific constrain function, but it doesn't work as I expected.
>
> Here is a demonstration:
> In [14]: a = []
>
> In [15]: for i in range(4):
>      b = lambda : i**2
>      a.append(b)
>     ....:
>     ....:
>
> In [18]: for f in a:
>      print f()
>     ....:
>     ....:
> 9
> 9
> 9
> 9
>
> What I want is that every function in list a should print different
> values. The value of a[0](), a[1](), a[2](), a[3]() should be 0, 1, 4,
> 9.
>
> How to achieve this?
>
> 2) In function f(X), there are item of log(xi). Although I have
> constrained the value of xi in the constrain functions, xi still may be
> equal or less than 0 during minimization, then "math domain error"
> occurred.
>
> What should I do ?
>   
I would replace xi and log(xi) by exp(xi) and xi
You can try also free (BSD lic) OpenOpt solver lincher, it should handle 
your problem well
something like
from scikits.openopt import NLP
p = NLP(f, x0, lb = zeros(x0.size))
r = p.solve('lincher')
however, currently lincher requires QP solver, and the only one 
connected is CVXOPT one, so you should have cvxopt installed (GPL). BTW 
cvxopt is capable of solving your problem, it has appropriate  NLP solver
Also, take a look at tnc
http://www.scipy.org/doc/api_docs/scipy.optimize.tnc.html
 - afaik it allows to handle lb - ub bounds


> My method is when "math domain error" occurred, catch it and set the
> return value of f to a very large number. Should this work or not?
>   
This will not work with any NLP or Nonsmooth solver, that calls for 
gradient/subgradient
this will work for some global solvers like anneal, but they are capable 
of small-scaled problems only (nVars up to 10)
HTH, D.
>
>
> Thanks a lot!
>
> Xiao Jianfeng
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>
>   



From openopt at ukr.net  Mon Jul 16 08:52:14 2007
From: openopt at ukr.net (dmitrey)
Date: Mon, 16 Jul 2007 15:52:14 +0300
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469B6202.6020002@gmail.com>
References: <469B6202.6020002@gmail.com>
Message-ID: <469B69FE.4040305@ukr.net>

fdu.xiaojf at gmail.com wrote:
> There are two problems:
>
> 1) In the constrain functions, I expected the value of i to be bounded
> to the specific constrain function, but it doesn't work as I expected.
>
> Here is a demonstration:
> In [14]: a = []
>
> In [15]: for i in range(4):
>      b = lambda : i**2
>      a.append(b)
>     ....:
>     ....:
>
> In [18]: for f in a:
>      print f()
>     ....:
>     ....:
> 9
> 9
> 9
> 9
>
> What I want is that every function in list a should print different
> values. The value of a[0](), a[1](), a[2](), a[3]() should be 0, 1, 4,
> 9.
>
> How to achieve this?
>   
if your b[i] are funcs from x, you should use b = lambda x: x **2 (for 
example)
however, NLP solvers can't handle these constraints as lb-ub bounds 
(fixed float numbers are required), you should describe them as 
non-linear constraints
c[i](x)<=0
HTH, D.


From massimo.sandal at unibo.it  Mon Jul 16 09:04:15 2007
From: massimo.sandal at unibo.it (massimo sandal)
Date: Mon, 16 Jul 2007 15:04:15 +0200
Subject: [SciPy-user] [OT] advice about publishing a scientific software
Message-ID: <469B6CCF.8090506@unibo.it>

Hi,

I am currently developing a (Python-and-Scipy based, of course!) data 
analysis software (more precisely, a single molecule force spectroscopy 
data analysis sw). It has been thought to be not the usual 
quick-and-dirty in house script, but to be an actual application that 
can be used in general by different labs. In fact, it actually features 
an easy and working (even if slightly unpolished) plugin API, and most 
functionality is implemented by mean of independent plugin.

Now, I would like to release it as free software (probably LGPL, but 
still thinking about it) online, however there are two concerns blocking me:

- Major one: I would like to publish a peer-reviewed paper about this 
software. If I want this, do I need *first* to write the paper, get it 
accepted, *then* to publish the software online? Or, given that 
functionality is plugin-driven, can I publish the software backbone and 
later publish a paper that presents the software and some (new, to 
release) nifty plugin that does some clever kind of data analysis? Or 
what? What are the usual requirements of journals about this?

- Minor one: Where to publish it? I am used to think about SourceForge 
as THE place for publishing open source software, however a number of 
people told me about Google Code. Basically what I want, if possible, is 
no-nonsense SVN access and some integrated mailing list/forum. Google 
Code seems good at that -it also gives automatically Wiki space, which 
is a major plus for me. What do you think about? Drawbacks? Licence 
issues (i.e. something like "since you publish it on Google Code the 
code is also owned by Google" tiny-letters clause?) Is there some other 
alternative?

Moreover, I'd like to know your opinions about the degree of 
"polishness" and other requirements for initial publication of a 
scientific software. As for documentation, I am working on actual, user 
manual and hacking (mainly but not only focused on plugin writing) 
manual with tutorials etc.. The sw still lacks an installer but 
"installation" is braindead (provided that you have all the correct 
dependencies): put all the files in the same directory and run "python 
software.py" from where: do you think an installer is absolutely 
necessary or it is OK to let it go this way, and add the installer later?

Thanks a lot,
Massimo

-- 
Massimo Sandal
University of Bologna
Department of Biochemistry "G.Moruzzi"

snail mail:
Via Irnerio 48, 40126 Bologna, Italy

email:
massimo.sandal at unibo.it

tel: +39-051-2094388
fax: +39-051-2094387
-------------- next part --------------
A non-text attachment was scrubbed...
Name: massimo.sandal.vcf
Type: text/x-vcard
Size: 274 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070716/50b4aad8/attachment.vcf>

From gael.varoquaux at normalesup.org  Mon Jul 16 09:01:41 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Mon, 16 Jul 2007 15:01:41 +0200
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <469B6CCF.8090506@unibo.it>
References: <469B6CCF.8090506@unibo.it>
Message-ID: <20070716130140.GH23143@clipper.ens.fr>

On Mon, Jul 16, 2007 at 03:04:15PM +0200, massimo sandal wrote:
> The sw still lacks an installer but 
> "installation" is braindead (provided that you have all the correct 
> dependencies): put all the files in the same directory and run "python 
> software.py" from where: do you think an installer is absolutely 
> necessary or it is OK to let it go this way, and add the installer later?

My 2 cents:

Use eggs for distribution. It just makes life so much easier for
everbody, its standard in the Python world, and it deals with
dependancies.

Ga?l


From fdu.xiaojf at gmail.com  Mon Jul 16 09:02:15 2007
From: fdu.xiaojf at gmail.com (fdu.xiaojf at gmail.com)
Date: Mon, 16 Jul 2007 21:02:15 +0800
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469B69FE.4040305@ukr.net>
References: <469B6202.6020002@gmail.com> <469B69FE.4040305@ukr.net>
Message-ID: <469B6C57.4010603@gmail.com>

dmitrey wrote:
> fdu.xiaojf at gmail.com wrote:
>> There are two problems:
>>
>> 1) In the constrain functions, I expected the value of i to be bounded
>> to the specific constrain function, but it doesn't work as I expected.
>>
>> Here is a demonstration:
>> In [14]: a = []
>>
>> In [15]: for i in range(4):
>>      b = lambda : i**2
>>      a.append(b)
>>     ....:
>>     ....:
>>
>> In [18]: for f in a:
>>      print f()
>>     ....:
>>     ....:
>> 9
>> 9
>> 9
>> 9
>>
>> What I want is that every function in list a should print different
>> values. The value of a[0](), a[1](), a[2](), a[3]() should be 0, 1, 4,
>> 9.
>>
>> How to achieve this?
>>   
> if your b[i] are funcs from x, you should use b = lambda x: x **2 (for 
> example)
> however, NLP solvers can't handle these constraints as lb-ub bounds 
> (fixed float numbers are required), you should describe them as 
> non-linear constraints
> c[i](x)<=0
> HTH, D.

Yes, b[i] are functions from X, and X is a list.

I want to constrain every xi in X to be larger than 0, so I tried to create 
the constrain function like this:

cons = []
for i in range(N): # N is the length of X, or the number of variables
     f = lambda x: x[i] - 1e-10
     cons.append(f)

But it didn't work as expected :(

According to the demands of fmin_cobyla, the parameter of cons should be a 
list of functions that all must be >=0.

Regards



From massimo.sandal at unibo.it  Mon Jul 16 09:13:39 2007
From: massimo.sandal at unibo.it (massimo sandal)
Date: Mon, 16 Jul 2007 15:13:39 +0200
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <20070716130140.GH23143@clipper.ens.fr>
References: <469B6CCF.8090506@unibo.it> <20070716130140.GH23143@clipper.ens.fr>
Message-ID: <469B6F03.3070606@unibo.it>

Gael Varoquaux ha scritto:
> On Mon, Jul 16, 2007 at 03:04:15PM +0200, massimo sandal wrote:
>> The sw still lacks an installer but 
>> "installation" is braindead (provided that you have all the correct 
>> dependencies): put all the files in the same directory and run "python 
>> software.py" from where: do you think an installer is absolutely 
>> necessary or it is OK to let it go this way, and add the installer later?
> 
> My 2 cents:
> 
> Use eggs for distribution. It just makes life so much easier for
> everbody, its standard in the Python world, and it deals with
> dependancies.

I haven't heard nothing but harsh criticism around "eggs", but I must 
admit I have never used them. Most Python software I've seen is 
distributed via the usual setup.py / *nix packages / Windows installers. 
All three are easy and standard.

What are the advantages of eggs? Where can I find info about them? Can a 
software redistributed with eggs play well (i.e. be repackaged) with 
Linux packaging or Windows installers?

m.

-- 
Massimo Sandal
University of Bologna
Department of Biochemistry "G.Moruzzi"

snail mail:
Via Irnerio 48, 40126 Bologna, Italy

email:
massimo.sandal at unibo.it

tel: +39-051-2094388
fax: +39-051-2094387
-------------- next part --------------
A non-text attachment was scrubbed...
Name: massimo.sandal.vcf
Type: text/x-vcard
Size: 274 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070716/7975dd35/attachment.vcf>

From openopt at ukr.net  Mon Jul 16 09:18:24 2007
From: openopt at ukr.net (dmitrey)
Date: Mon, 16 Jul 2007 16:18:24 +0300
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469B6C57.4010603@gmail.com>
References: <469B6202.6020002@gmail.com> <469B69FE.4040305@ukr.net>
	<469B6C57.4010603@gmail.com>
Message-ID: <469B7020.3060803@ukr.net>

fdu.xiaojf at gmail.com wrote:
> Yes, b[i] are functions from X, and X is a list.
>
> I want to constrain every xi in X to be larger than 0, so I tried to create 
> the constrain function like this:
>
> cons = []
> for i in range(N): # N is the length of X, or the number of variables
>      f = lambda x: x[i] - 1e-10
>      cons.append(f)
>
> But it didn't work as expected :(
>
> According to the demands of fmin_cobyla, the parameter of cons should be a 
> list of functions that all must be >=0.
>
> Regards
>
>   
so if you have only x >= 0 constraints, it's better to use tnc or 
openopt lincher than fmin_cobyla

if you have to use cobyla anyway, try this:


a = []
for i in range(5):
    a.append(lambda x: x[i]**2)
x = [1,2,3,4,5]
for i in range(5):
    print a[i](x)

 1
4
9
16
25

If you will replace xi by exp(xi), don't forget to replace lb[i]=0 by 
lb[i]=1

HTH, d


From fdu.xiaojf at gmail.com  Mon Jul 16 09:16:30 2007
From: fdu.xiaojf at gmail.com (fdu.xiaojf at gmail.com)
Date: Mon, 16 Jul 2007 21:16:30 +0800
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469B64EF.2080401@ukr.net>
References: <469B6202.6020002@gmail.com> <469B64EF.2080401@ukr.net>
Message-ID: <469B6FAE.1070807@gmail.com>

dmitrey wrote:
 > fdu.xiaojf at gmail.com wrote:

 >>
 >> 2) In function f(X), there are item of log(xi). Although I have
 >> constrained the value of xi in the constrain functions, xi still may be
 >> equal or less than 0 during minimization, then "math domain error"
 >> occurred.
 >>
 >> What should I do ?
 >>
 > I would replace xi and log(xi) by exp(xi) and xi
 > You can try also free (BSD lic) OpenOpt solver lincher, it should handle
 > your problem well
 > something like
 > from scikits.openopt import NLP
 > p = NLP(f, x0, lb = zeros(x0.size))
 > r = p.solve('lincher')
 > however, currently lincher requires QP solver, and the only one
 > connected is CVXOPT one, so you should have cvxopt installed (GPL). BTW
 > cvxopt is capable of solving your problem, it has appropriate  NLP solver
 > Also, take a look at tnc
 > http://www.scipy.org/doc/api_docs/scipy.optimize.tnc.html
 >  - afaik it allows to handle lb - ub bounds
 >
cvxopt can only handle convex functions, but my function is too
complicated to get the expression of derivate easily, so according to
a previous post from Joachim Dahl(dahl.joachim at gmail.com) in this list,
it is probably non-convex.

Can openopt handle non-convex functions?

My function is too complex to compute derivate, so I can't use tnc to do
the job.

 >
 >> My method is when "math domain error" occurred, catch it and set the
 >> return value of f to a very large number. Should this work or not?
 >>
 > This will not work with any NLP or Nonsmooth solver, that calls for
 > gradient/subgradient this will work for some global solvers like
 > anneal, but they are capable of small-scaled problems only (nVars up
 > to 10) HTH, D.

The number of variables is less than 15, does this make any sense?


Regards,




From gael.varoquaux at normalesup.org  Mon Jul 16 09:20:14 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Mon, 16 Jul 2007 15:20:14 +0200
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <469B6F03.3070606@unibo.it>
References: <469B6CCF.8090506@unibo.it> <20070716130140.GH23143@clipper.ens.fr>
	<469B6F03.3070606@unibo.it>
Message-ID: <20070716132014.GI23143@clipper.ens.fr>

On Mon, Jul 16, 2007 at 03:13:39PM +0200, massimo sandal wrote:
> I haven't heard nothing but harsh criticism around "eggs",

I have never heard any criticism. I much just be hidding in a hole to
deep :->. 

> Most Python software I've seen is 
> distributed via the usual setup.py / *nix packages / Windows installers. 
> All three are easy and standard.

Well, I would call Unix package easy to develop. And they are
distribution specific. Windows installers... I must admit I don't have a
clue how to develop one of these, and I can't learn, I don't have a
Windows box lying around (and I am not to interested). setup.py... well,
yes, that's pretty much what eggs are, with some sugar over it.

> What are the advantages of eggs?

Do you know apt, or yum ? Well aggs are pretty much that for Python (like
debs or rpm): an egg is versions, the "easy_install" script used to
install eggs knows where to place them automatically, it can put your
scripts in the path, deal with dependencies (eg by downloading eggs
from pypi). Of course eggs are not perfect, they are no where as nice as
debs, but they are easier to create.

> Where can I find info about them? Can a software redistributed with
> eggs play well (i.e. be repackaged) with Linux packaging or Windows
> installers?

There are projects doing this (http://stdeb.python-hosting.com to makes
debs. I don't use windows, so I never tried this out, but it seems
possible to make a windows installer (using the "bdist_wininst" to the
setup.py script, that also can be used to generate the eggs).

Info can be found at http://peak.telecommunity.com/DevCenter/setuptools .
Try it out, it feels a bit strange in the beginning, as it looks like
you have written no code, so nothing can happen, but its quite handy.

Ga?l


From fdu.xiaojf at gmail.com  Mon Jul 16 09:25:42 2007
From: fdu.xiaojf at gmail.com (fdu.xiaojf at gmail.com)
Date: Mon, 16 Jul 2007 21:25:42 +0800
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469B7020.3060803@ukr.net>
References: <469B6202.6020002@gmail.com>
	<469B69FE.4040305@ukr.net>	<469B6C57.4010603@gmail.com>
	<469B7020.3060803@ukr.net>
Message-ID: <469B71D6.7030601@gmail.com>

Hi dmitrey,

Thank you so much for your kindness!

dmitrey wrote:

 > so if you have only x >= 0 constraints, it's better to use tnc or
 > openopt lincher than fmin_cobyla

Unfortunately, the function I want to minimize still has equality
constraints. The reason why I chose cobyla is that cobyla can handle
inequality and equality constraints, and it doesn't require derivate
information.

 >
 > if you have to use cobyla anyway, try this:
 >
 >
 > a = []
 > for i in range(5):
 >     a.append(lambda x: x[i]**2)
 > x = [1,2,3,4,5]
 > for i in range(5):
 >     print a[i](x)
 >
 >  1
 > 4
 > 9
 > 16
 > 25

The constrain functions are called by fmin_cobyla, so I couldn't make
sure "for i ...." are  used when the constrain functions are called.

In your example, if I use "for j in range(5): print a[j](x)", it will
not work as expected.

 >
 > If you will replace xi by exp(xi), don't forget to replace lb[i]=0 by
 > lb[i]=1

I'm trying this.
 >
 > HTH, d


From openopt at ukr.net  Mon Jul 16 09:35:00 2007
From: openopt at ukr.net (dmitrey)
Date: Mon, 16 Jul 2007 16:35:00 +0300
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469B6FAE.1070807@gmail.com>
References: <469B6202.6020002@gmail.com> <469B64EF.2080401@ukr.net>
	<469B6FAE.1070807@gmail.com>
Message-ID: <469B7404.6040507@ukr.net>

fdu.xiaojf at gmail.com wrote:
> dmitrey wrote:
>  > fdu.xiaojf at gmail.com wrote:
>
>  >>
>  >> 2) In function f(X), there are item of log(xi). Although I have
>  >> constrained the value of xi in the constrain functions, xi still may be
>  >> equal or less than 0 during minimization, then "math domain error"
>  >> occurred.
>  >>
>  >> What should I do ?
>  >>
>  > I would replace xi and log(xi) by exp(xi) and xi
>  > You can try also free (BSD lic) OpenOpt solver lincher, it should handle
>  > your problem well
>  > something like
>  > from scikits.openopt import NLP
>  > p = NLP(f, x0, lb = zeros(x0.size))
>  > r = p.solve('lincher')
>  > however, currently lincher requires QP solver, and the only one
>  > connected is CVXOPT one, so you should have cvxopt installed (GPL). BTW
>  > cvxopt is capable of solving your problem, it has appropriate  NLP solver
>  > Also, take a look at tnc
>  > http://www.scipy.org/doc/api_docs/scipy.optimize.tnc.html
>  >  - afaik it allows to handle lb - ub bounds
>  >
> cvxopt can only handle convex functions, but my function is too
> complicated to get the expression of derivate easily, so according to
> a previous post from Joachim Dahl(dahl.joachim at gmail.com) in this list,
> it is probably non-convex.
>
> Can openopt handle non-convex functions?
>   
this is incorrect question. For any NLP solver (w/o global ones of 
course) I can construct a non-convex func that will failed to be solved. 
Moreover, I can construct *convex* func, which one will be very, very 
hard to solve (for the solver chose). So I can only talk about "this 
solver is more or less suitable for solving non-convex funcs". 
Fortunately, lincher is rather suitable for solving non-convex funcs, 
but let me remember you once again - it requires cvxopt qp solver yet, 
no other ones are connected or are written (I hope till the end of 
summer openopt will have it own QP/QPCP solver).
> My function is too complex to compute derivate, so I can't use tnc to do
> the job.
>   

if derivatives are not provided to tnc, it will calculate that ones by 
itself via finite-differences, as any other NLP solver.
>  >
>  >> My method is when "math domain error" occurred, catch it and set the
>  >> return value of f to a very large number. Should this work or not?
>  >>
>  > This will not work with any NLP or Nonsmooth solver, that calls for
>  > gradient/subgradient this will work for some global solvers like
>  > anneal, but they are capable of small-scaled problems only (nVars up
>  > to 10) HTH, D.
>
> The number of variables is less than 15, does this make any sense?
>   
Very unlikely.
10 is already a big problem.

>The reason why I chose cobyla is that cobyla can handle inequality and equality constraints, and it doesn't require derivate information.

lincher is capable of handling equality constraints. However, there can 
be some difficult cases (for example, when constraints form a 
linear-dependent system in xk - point from iter k)
But for rather small nVars = 15 I guess no problems will be obtained.
HTH, D.
>
> Regards,
>
>   



From c.j.lee at tnw.utwente.nl  Mon Jul 16 09:54:26 2007
From: c.j.lee at tnw.utwente.nl (Chris Lee)
Date: Mon, 16 Jul 2007 15:54:26 +0200
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <469B6CCF.8090506@unibo.it>
References: <469B6CCF.8090506@unibo.it>
Message-ID: <F9457D37-EAA8-4E98-8948-E9E22ABFC2B4@tnw.utwente.nl>


On Jul 16, 2007, at 3:04 PM, massimo sandal wrote:

> I am currently developing a (Python-and-Scipy based, of course!)  
> data analysis software (more precisely, a single molecule force  
> spectroscopy data analysis sw). It has been thought to be not the  
> usual quick-and-dirty in house script, but to be an actual  
> application that can be used in general by different labs. In fact,  
> it actually features an easy and working (even if slightly  
> unpolished) plugin API, and most functionality is implemented by  
> mean of independent plugin.

I would think that you could write a paper to the appropriate methods  
journal and its review would be independent of the release to open  
source as long as you clearly stated in the paper that the authors  
were the sole developers up to the x.yz release.

Which also answers the question about which Journal. Every field has  
either specialist methods/instrumentation journals or journals with  
sections devoted to that.  I would aim for one of those.

Cheers
Chris


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070716/6889b870/attachment.html>

From fdu.xiaojf at gmail.com  Mon Jul 16 10:00:36 2007
From: fdu.xiaojf at gmail.com (fdu.xiaojf at gmail.com)
Date: Mon, 16 Jul 2007 22:00:36 +0800
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469B7404.6040507@ukr.net>
References: <469B6202.6020002@gmail.com>
	<469B64EF.2080401@ukr.net>	<469B6FAE.1070807@gmail.com>
	<469B7404.6040507@ukr.net>
Message-ID: <469B7A04.50102@gmail.com>

dmitrey wrote:
 > fdu.xiaojf at gmail.com wrote:
 >>  >
 >> cvxopt can only handle convex functions, but my function is too
 >> complicated to get the expression of derivate easily, so according to
 >> a previous post from Joachim Dahl(dahl.joachim at gmail.com) in this list,
 >> it is probably non-convex.
 >>
 >> Can openopt handle non-convex functions?
 >>
 > this is incorrect question. For any NLP solver (w/o global ones of
 > course) I can construct a non-convex func that will failed to be solved.
 > Moreover, I can construct *convex* func, which one will be very, very
 > hard to solve (for the solver chose). So I can only talk about "this
 > solver is more or less suitable for solving non-convex funcs".
 > Fortunately, lincher is rather suitable for solving non-convex funcs,
 > but let me remember you once again - it requires cvxopt qp solver yet,
 > no other ones are connected or are written (I hope till the end of
 > summer openopt will have it own QP/QPCP solver).

Thanks.

 >> My function is too complex to compute derivate, so I can't use tnc to do
 >> the job.
 >>
 >
 > if derivatives are not provided to tnc, it will calculate that ones by
 > itself via finite-differences, as any other NLP solver.
Oh, yes, tnc can calculate derivatives if the parameter approx_grad is
True.

 >>  >
 >>  >> My method is when "math domain error" occurred, catch it and set the
 >>  >> return value of f to a very large number. Should this work or not?
 >>  >>
 >>  > This will not work with any NLP or Nonsmooth solver, that calls for
 >>  > gradient/subgradient this will work for some global solvers like
 >>  > anneal, but they are capable of small-scaled problems only (nVars up
 >>  > to 10) HTH, D.
 >>
 >> The number of variables is less than 15, does this make any sense?
 >>
 > Very unlikely.
 > 10 is already a big problem.
 >
 >> The reason why I chose cobyla is that cobyla can handle inequality and 
equality constraints, and it doesn't require derivate information.
 >
 > lincher is capable of handling equality constraints. However, there can
 > be some difficult cases (for example, when constraints form a
 > linear-dependent system in xk - point from iter k)
 > But for rather small nVars = 15 I guess no problems will be obtained.

Can linear avoid the "math domain error" prolbem in cobyla ? In my
function, there are items like log(xi*R*T/V) (R,T,V are positive
variables too).

Regards



From openopt at ukr.net  Mon Jul 16 10:09:49 2007
From: openopt at ukr.net (dmitrey)
Date: Mon, 16 Jul 2007 17:09:49 +0300
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469B7A04.50102@gmail.com>
References: <469B6202.6020002@gmail.com>	<469B64EF.2080401@ukr.net>	<469B6FAE.1070807@gmail.com>	<469B7404.6040507@ukr.net>
	<469B7A04.50102@gmail.com>
Message-ID: <469B7C2D.9090805@ukr.net>

fdu.xiaojf at gmail.com wrote:
> Can linear avoid the "math domain error" prolbem in cobyla ? In my
> function, there are items like log(xi*R*T/V) (R,T,V are positive
> variables too).
>   
do you mean "lincher" instead of "linear"?
no, at least for now (but some solvers, usually commercial, can handle 
NaN in objfun).

if R, T, V are constants I would replace

zi = log(xi*R*T/V) hence
xi = exp(xi)*V/(R*T)

hth, D

> Regards
>
>
>   



From aisaac at american.edu  Mon Jul 16 10:23:27 2007
From: aisaac at american.edu (Alan G Isaac)
Date: Mon, 16 Jul 2007 10:23:27 -0400
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <F9457D37-EAA8-4E98-8948-E9E22ABFC2B4@tnw.utwente.nl>
References: <469B6CCF.8090506@unibo.it><F9457D37-EAA8-4E98-8948-E9E22ABFC2B4@tnw.utwente.nl>
Message-ID: <Mahogany-0.67.0-708-20070716-102327.00@american.edu>

This is the kind of thing that would be good for the 
projected SciPy journal.  It might be appropriate for
JStatSoft http://www.jstatsoft.org/, although that has
traditionally had a bias toward R.

fwiw,
Alan Isaac




From massimo.sandal at unibo.it  Mon Jul 16 10:50:33 2007
From: massimo.sandal at unibo.it (massimo sandal)
Date: Mon, 16 Jul 2007 16:50:33 +0200
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <Mahogany-0.67.0-708-20070716-102327.00@american.edu>
References: <469B6CCF.8090506@unibo.it><F9457D37-EAA8-4E98-8948-E9E22ABFC2B4@tnw.utwente.nl>
	<Mahogany-0.67.0-708-20070716-102327.00@american.edu>
Message-ID: <469B85B9.9010502@unibo.it>

Alan G Isaac ha scritto:
> This is the kind of thing that would be good for the 
> projected SciPy journal.  It might be appropriate for
> JStatSoft http://www.jstatsoft.org/, although that has
> traditionally had a bias toward R.

Well, its peculiarity is surely not in being a strong statistical 
package, but in being one of the few software packages for force 
spectroscopy analysis (and most probably the first one being both open 
source and plugin-capable). I don't feel J.Stat.Soft. would be right, 
but maybe I'm wrong.

m.

-- 
Massimo Sandal
University of Bologna
Department of Biochemistry "G.Moruzzi"

snail mail:
Via Irnerio 48, 40126 Bologna, Italy

email:
massimo.sandal at unibo.it

tel: +39-051-2094388
fax: +39-051-2094387
-------------- next part --------------
A non-text attachment was scrubbed...
Name: massimo.sandal.vcf
Type: text/x-vcard
Size: 274 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070716/edbb959a/attachment.vcf>

From robert.kern at gmail.com  Mon Jul 16 11:51:30 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Mon, 16 Jul 2007 10:51:30 -0500
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <469B6CCF.8090506@unibo.it>
References: <469B6CCF.8090506@unibo.it>
Message-ID: <469B9402.5070403@gmail.com>

massimo sandal wrote:

> - Minor one: Where to publish it? I am used to think about SourceForge
> as THE place for publishing open source software, however a number of
> people told me about Google Code. Basically what I want, if possible, is
> no-nonsense SVN access and some integrated mailing list/forum. Google
> Code seems good at that -it also gives automatically Wiki space, which
> is a major plus for me. What do you think about? Drawbacks? Licence
> issues (i.e. something like "since you publish it on Google Code the
> code is also owned by Google" tiny-letters clause?) Is there some other
> alternative?

If you want "no-nonsense," use Google Code. Sourceforge is a lot of nonsense. I
get incredibly frustrated trying to use the Sourceforge website or any of its
spawn on either side (downloader or developer). I don't believe there are any IP
issues; you own your code, not Google.

> Moreover, I'd like to know your opinions about the degree of
> "polishness" and other requirements for initial publication of a
> scientific software.

Meh. Release early and often. Having the code public is a good way to motivate
polish. Your code may crash or miss features, but it's code that wasn't
available before. You are making a positive contribution regardless.

> As for documentation, I am working on actual, user
> manual and hacking (mainly but not only focused on plugin writing)
> manual with tutorials etc.. The sw still lacks an installer but
> "installation" is braindead (provided that you have all the correct
> dependencies): put all the files in the same directory and run "python
> software.py" from where: do you think an installer is absolutely
> necessary or it is OK to let it go this way, and add the installer later?

For a Python project, you really should be using distutils, even if it is just
the simplest possible setup.py.

-- 
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 massimo.sandal at unibo.it  Mon Jul 16 12:33:09 2007
From: massimo.sandal at unibo.it (massimo sandal)
Date: Mon, 16 Jul 2007 18:33:09 +0200
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <469B9402.5070403@gmail.com>
References: <469B6CCF.8090506@unibo.it> <469B9402.5070403@gmail.com>
Message-ID: <469B9DC5.7050301@unibo.it>

Robert Kern ha scritto:
> massimo sandal wrote:
> If you want "no-nonsense," use Google Code. Sourceforge is a lot of nonsense. I
> get incredibly frustrated trying to use the Sourceforge website or any of its
> spawn on either side (downloader or developer). I don't believe there are any IP
> issues; you own your code, not Google.

OK, thanks.

>> Moreover, I'd like to know your opinions about the degree of
>> "polishness" and other requirements for initial publication of a
>> scientific software.
> 
> Meh. Release early and often. Having the code public is a good way to motivate
> polish. Your code may crash or miss features, but it's code that wasn't
> available before. You are making a positive contribution regardless.

Yes, but as stated before, I want to publish something about this code. 
So I expect that it must have some minimum standard of quality.

m.

-- 
Massimo Sandal
University of Bologna
Department of Biochemistry "G.Moruzzi"

snail mail:
Via Irnerio 48, 40126 Bologna, Italy

email:
massimo.sandal at unibo.it

tel: +39-051-2094388
fax: +39-051-2094387
-------------- next part --------------
A non-text attachment was scrubbed...
Name: massimo.sandal.vcf
Type: text/x-vcard
Size: 274 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070716/2754e433/attachment.vcf>

From robert.kern at gmail.com  Mon Jul 16 12:42:37 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Mon, 16 Jul 2007 11:42:37 -0500
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <469B9DC5.7050301@unibo.it>
References: <469B6CCF.8090506@unibo.it> <469B9402.5070403@gmail.com>
	<469B9DC5.7050301@unibo.it>
Message-ID: <469B9FFD.5070008@gmail.com>

massimo sandal wrote:
> Robert Kern ha scritto:
>> massimo sandal wrote:

>>> Moreover, I'd like to know your opinions about the degree of
>>> "polishness" and other requirements for initial publication of a
>>> scientific software.
>>
>> Meh. Release early and often. Having the code public is a good way to
>> motivate
>> polish. Your code may crash or miss features, but it's code that wasn't
>> available before. You are making a positive contribution regardless.
> 
> Yes, but as stated before, I want to publish something about this code.
> So I expect that it must have some minimum standard of quality.

I would think that the only time that would matter is on submission of your
paper, not before.

-- 
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 massimo.sandal at unibo.it  Mon Jul 16 13:32:18 2007
From: massimo.sandal at unibo.it (massimo sandal)
Date: Mon, 16 Jul 2007 19:32:18 +0200
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <469B9FFD.5070008@gmail.com>
References: <469B6CCF.8090506@unibo.it>
	<469B9402.5070403@gmail.com>	<469B9DC5.7050301@unibo.it>
	<469B9FFD.5070008@gmail.com>
Message-ID: <469BABA2.4030608@unibo.it>

Robert Kern ha scritto:
> massimo sandal wrote:
>> Robert Kern ha scritto:
> I would think that the only time that would matter is on submission of your
> paper, not before.

That's exactly what I feared.

m.

-- 
Massimo Sandal
University of Bologna
Department of Biochemistry "G.Moruzzi"

snail mail:
Via Irnerio 48, 40126 Bologna, Italy

email:
massimo.sandal at unibo.it

tel: +39-051-2094388
fax: +39-051-2094387
-------------- next part --------------
A non-text attachment was scrubbed...
Name: massimo.sandal.vcf
Type: text/x-vcard
Size: 274 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070716/28d2f44c/attachment.vcf>

From aisaac at american.edu  Mon Jul 16 13:51:58 2007
From: aisaac at american.edu (Alan G Isaac)
Date: Mon, 16 Jul 2007 13:51:58 -0400
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <469BABA2.4030608@unibo.it>
References: <469B6CCF.8090506@unibo.it><469B9402.5070403@gmail.com>	<469B9DC5.7050301@unibo.it><469B9FFD.5070008@gmail.com><469BABA2.4030608@unibo.it>
Message-ID: <Mahogany-0.67.0-708-20070716-135158.00@american.edu>

> Robert Kern ha scritto:
>> I would think that the only time that would matter is on 
>> submission of your paper, not before. 

On Mon, 16 Jul 2007, massimo sandal apparently wrote:
> That's exactly what I feared. 

I think Robert was suggesting that early release is a good
way to quickly get to the code quality you need for a
useful pulbication, not that it would count against you
at the time of publication.

Cheers,
Alan Isaac




From peridot.faceted at gmail.com  Mon Jul 16 14:00:24 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Mon, 16 Jul 2007 14:00:24 -0400
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469B6202.6020002@gmail.com>
References: <469B6202.6020002@gmail.com>
Message-ID: <ce557a360707161100t437166ael190142f77564b68c@mail.gmail.com>

On 16/07/07, fdu.xiaojf at gmail.com <fdu.xiaojf at gmail.com> wrote:

> 2) In function f(X), there are item of log(xi). Although I have
> constrained the value of xi in the constrain functions, xi still may be
> equal or less than 0 during minimization, then "math domain error"
> occurred.
>
> What should I do ?
>
> My method is when "math domain error" occurred, catch it and set the
> return value of f to a very large number. Should this work or not?

This problem happens because while fmin_cobyla is able to handle
constraints, it nevertheless evaluates the objective function outside
those constraints. I think this is a very serious bug, which limits
the usefulness of fmin_cobyla. Unfortunately, I do not have a general
solution.

In your specific case, where you have positive numbers, I would
recommend (as another poster did) that you reparameterize: any
positive number xi can be written as exp(yi). If you write it this
way, yi may take any value and xi will still be positive. If you're
lucky, this may allow you to dispense with constraints entirely. For a
simple example:

Suppose we want to do least-squares fitting of an exponential to some
data points:

yvals = A*exp(b*xvals) + noise

Suppose further that we want to ensure A is positive. We could minimize
sum((yvals-A*exp(b*xvals))**2)
subject to the constraint A>0. But this will be tricky. Instead, write
A=exp(a), and minimize
sum((yvals-exp(a)*exp(b*xvals))**2).

Anne


From dominique.orban at gmail.com  Mon Jul 16 14:14:20 2007
From: dominique.orban at gmail.com (Dominique Orban)
Date: Mon, 16 Jul 2007 14:14:20 -0400
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <ce557a360707161100t437166ael190142f77564b68c@mail.gmail.com>
References: <469B6202.6020002@gmail.com>
	<ce557a360707161100t437166ael190142f77564b68c@mail.gmail.com>
Message-ID: <469BB57C.6010809@gmail.com>

Anne Archibald wrote:
> On 16/07/07, fdu.xiaojf at gmail.com <fdu.xiaojf at gmail.com> wrote:
> 
>> 2) In function f(X), there are item of log(xi). Although I have
>> constrained the value of xi in the constrain functions, xi still may be
>> equal or less than 0 during minimization, then "math domain error"
>> occurred.
>>
>> What should I do ?
>>
>> My method is when "math domain error" occurred, catch it and set the
>> return value of f to a very large number. Should this work or not?
> 
> This problem happens because while fmin_cobyla is able to handle
> constraints, it nevertheless evaluates the objective function outside
> those constraints. I think this is a very serious bug, which limits
> the usefulness of fmin_cobyla. Unfortunately, I do not have a general
> solution.

This behavior of Cobyla is not a bug.

There are two kinds of methods for constrained problems in optimization: 
feasible and infeasible methods. The iterates generated in a feasible method 
will all satisfy the constraints. In an infeasible method, they may or may not 
satisfy the constraints but, if convergence happens and if all goes well, the 
iterates converge to a point which is (approximately) feasible. Cobyla is an 
infeasible method.

The fact that your objective function is not defined everywhere (because of the 
log terms) suggests that you should not be using an infeasible method. Instead, 
you could look into feasible interior-point methods (beware, there also are 
infeasible interior-point methods... in fact, most of them are).

NLPy features one such method, albeit a simple one, under the name TRIP (for 
'trust-region interior-point method'). For those who care about this sort of 
thing, it is a purely primal method at the moment. I mostly implemented it as a 
proof of concept in the NLPy framework, and plan to upgrade it in the future. It 
will not treat equality constraints, but the inequalities can be as nonlinear as 
you please, as long as they are twice differentiable. You can give it a try and 
let me know what happens.

In addition to the constraints you already have, add new constraints stating 
that the arguments of the logs must be >= 0. Interior-point methods ensure that 
at each iteration, those quantities remain > 0 (strictly).

If the constraints you end up with are *only bound constraints*, you can also 
look into L-BFGS-B (which I believe is interfaced in SciPy), a projected 
quasi-Newton method, or TRON (www.mcs.anl.gov/~more/tron which I don't think is 
interfaced in SciPy), a projected Newton method. NLPy also contains a 
gradient-projection method which I find surprisingly efficient, but also only 
applies to bound constraints.

Re-parametrizing your problem is, of course, another option.

Good luck,
Dominique


From robert.kern at gmail.com  Mon Jul 16 17:03:55 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Mon, 16 Jul 2007 16:03:55 -0500
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <Mahogany-0.67.0-708-20070716-135158.00@american.edu>
References: <469B6CCF.8090506@unibo.it><469B9402.5070403@gmail.com>	<469B9DC5.7050301@unibo.it><469B9FFD.5070008@gmail.com><469BABA2.4030608@unibo.it>
	<Mahogany-0.67.0-708-20070716-135158.00@american.edu>
Message-ID: <469BDD3B.3060008@gmail.com>

Alan G Isaac wrote:
>> Robert Kern ha scritto:
>>> I would think that the only time that would matter is on 
>>> submission of your paper, not before. 
> 
> On Mon, 16 Jul 2007, massimo sandal apparently wrote:
>> That's exactly what I feared. 
> 
> I think Robert was suggesting that early release is a good
> way to quickly get to the code quality you need for a
> useful pulbication, not that it would count against you
> at the time of publication.

Indeed. To rephrase a little bit more clearly: "I would think that only the
quality of the code at the time of submission of your paper would be important."
I don't think reviewers will be going back through your SVN logs. Sharing code
before publication is expected, I think. As Chris Lee suggests, though, you have
to be very careful about keeping track of contributions, more so than with other
open source projects.

-- 
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  Mon Jul 16 17:07:54 2007
From: wbaxter at gmail.com (Bill Baxter)
Date: Tue, 17 Jul 2007 06:07:54 +0900
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <469B9402.5070403@gmail.com>
References: <469B6CCF.8090506@unibo.it> <469B9402.5070403@gmail.com>
Message-ID: <e86a5fd00707161407md9ce204j10dc41ac2ecb9d19@mail.gmail.com>

On 7/17/07, Robert Kern <robert.kern at gmail.com> wrote:
> massimo sandal wrote:
>
> If you want "no-nonsense," use Google Code.

The only issue I've heard with Google Code is that they only allow you
to choose from a fixed list of licenses.  But most of the usual
suspects are there, so that probably won't be a problem for you,
unless you are determined to use some obscure/homebrew license.  What
it does stop is mature projects (like wxPython) with pre-existing
licenses that happen not to be on Google's list.

--bb


From amcmorl at gmail.com  Mon Jul 16 18:07:22 2007
From: amcmorl at gmail.com (Angus McMorland)
Date: Tue, 17 Jul 2007 10:07:22 +1200
Subject: [SciPy-user] Numerical Recipes robust fit implementation
In-Reply-To: <ce557a360707160454m4ef59a33pe50ad72e8255583c@mail.gmail.com>
References: <e85f13c40707152245l57a23b99x9a13045be6c79d66@mail.gmail.com>
	<469B120F.6000006@astraw.com>
	<e85f13c40707160401l30543ff9qd1e67a4341d1d3ee@mail.gmail.com>
	<ce557a360707160454m4ef59a33pe50ad72e8255583c@mail.gmail.com>
Message-ID: <e85f13c40707161507h2c03864ch71291f3d4d819d5c@mail.gmail.com>

Hi all,

On 16/07/07, Anne Archibald <peridot.faceted at gmail.com> wrote:
> On 16/07/07, Angus McMorland <amcmorl at gmail.com> wrote:
>
> > I'm going to have to think a bit more what I want to achieve to see if
> > RANSAC is useful. Ultimately I hope to determine the probability of a
> > a given data set being exponentially distributed, by comparing the raw
> > frequency distribution to an expected distribution based on a linear
> > fit to the log transform of the raw one. It seems a bit like basing my
> > 'expected' distribution on a subset of data from which outliers have
> > been completely excluded is self-fulfilling, and having some other
> > criterion for weighting of the error term (as medfit does) seems more
> > appropriate. This is however very much just a gut feeling rather than
> > an educated assessment, so any other comments are welcome.
>
> If what you're trying to do is test whether your data points are
> likely to have been drawn from a given distribution, you may be able
> to do much better by not putting them in a histogram first, and using
> something like the Kolmogorov-Smirnov test (scipy.stats.kstest). If
> you have outliers you may have a problem. (It's feasible to fit
> parameters so they maximize the kstest p value, although of course the
> p value you get at the end is no longer an actual probability.) I
> suspect if you look in the statistical literature you'll find other
> tricks for fitting distributions directly (rather than fitting
> histograms).

Kolmogorov-Smirnov is the way I had intended originally to go (there's
a variant that approximates the probability for grouped (read:
frequency) data). But, as Anne rightly points out I can apply it to
the individual variates since I have them--- my earlier approach arose
mainly from how I was looking at the data, and I need to get away from
that.

I still need an expected cdf based on my hypothesized distribution,
parameters for which I want to estimate from my data: that's where the
fitting comes in. My only remaining decision is whether a robust or
least-squares fitting approach is more appropriate for deriving the
expected distribution. The former is inherently self-fulfilling, in
that it excludes from the estimation of the expected distribution the
outliers that are likely the important deviations, and the latter will
include all the data, but fit none of it very well. Time to play
around and see how much difference there is, I think.

Thanks for all your suggestions, they've been very useful.

Angus.
-- 
AJC McMorland, PhD Student
Physiology, University of Auckland


From peridot.faceted at gmail.com  Mon Jul 16 19:32:06 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Mon, 16 Jul 2007 19:32:06 -0400
Subject: [SciPy-user] Numerical Recipes robust fit implementation
In-Reply-To: <e85f13c40707161507h2c03864ch71291f3d4d819d5c@mail.gmail.com>
References: <e85f13c40707152245l57a23b99x9a13045be6c79d66@mail.gmail.com>
	<469B120F.6000006@astraw.com>
	<e85f13c40707160401l30543ff9qd1e67a4341d1d3ee@mail.gmail.com>
	<ce557a360707160454m4ef59a33pe50ad72e8255583c@mail.gmail.com>
	<e85f13c40707161507h2c03864ch71291f3d4d819d5c@mail.gmail.com>
Message-ID: <ce557a360707161632w51f78bb7wfa0bf4aa66b67aa1@mail.gmail.com>

On 16/07/07, Angus McMorland <amcmorl at gmail.com> wrote:

> Kolmogorov-Smirnov is the way I had intended originally to go (there's
> a variant that approximates the probability for grouped (read:
> frequency) data). But, as Anne rightly points out I can apply it to
> the individual variates since I have them--- my earlier approach arose
> mainly from how I was looking at the data, and I need to get away from
> that.
>
> I still need an expected cdf based on my hypothesized distribution,
> parameters for which I want to estimate from my data: that's where the
> fitting comes in. My only remaining decision is whether a robust or
> least-squares fitting approach is more appropriate for deriving the
> expected distribution. The former is inherently self-fulfilling, in
> that it excludes from the estimation of the expected distribution the
> outliers that are likely the important deviations, and the latter will
> include all the data, but fit none of it very well. Time to play
> around and see how much difference there is, I think.
>
> Thanks for all your suggestions, they've been very useful.

I don't know why I didn't think of this before, I've been working with
them, but if you want to estimate a PDF (and therefore a CDF), kernel
density estimators are a very reasonable approach. Scipy implements
one, but if you want to include outliers you may find using a kernel
with bigger tails than a Gaussian useful. I don't know of a robust
kernel density estimator, but they've seen extensive work in the
statistical literature.

Anne


From amcmorl at gmail.com  Mon Jul 16 23:41:14 2007
From: amcmorl at gmail.com (Angus McMorland)
Date: Tue, 17 Jul 2007 15:41:14 +1200
Subject: [SciPy-user] Numerical Recipes robust fit implementation
In-Reply-To: <ce557a360707161632w51f78bb7wfa0bf4aa66b67aa1@mail.gmail.com>
References: <e85f13c40707152245l57a23b99x9a13045be6c79d66@mail.gmail.com>
	<469B120F.6000006@astraw.com>
	<e85f13c40707160401l30543ff9qd1e67a4341d1d3ee@mail.gmail.com>
	<ce557a360707160454m4ef59a33pe50ad72e8255583c@mail.gmail.com>
	<e85f13c40707161507h2c03864ch71291f3d4d819d5c@mail.gmail.com>
	<ce557a360707161632w51f78bb7wfa0bf4aa66b67aa1@mail.gmail.com>
Message-ID: <e85f13c40707162041r2fe3930crb1e2ed376c8f029d@mail.gmail.com>

Hi Anne et al.,

On 17/07/07, Anne Archibald <peridot.faceted at gmail.com> wrote:
> On 16/07/07, Angus McMorland <amcmorl at gmail.com> wrote:
> I don't know why I didn't think of this before, I've been working with
> them, but if you want to estimate a PDF (and therefore a CDF), kernel
> density estimators are a very reasonable approach. Scipy implements
> one, but if you want to include outliers you may find using a kernel
> with bigger tails than a Gaussian useful. I don't know of a robust
> kernel density estimator, but they've seen extensive work in the
> statistical literature.

I may be misunderstanding, but it looks like KDEs generate generic
pdfs based on the distribution of the observed values and not any
particular standard distribution form. I specifically need an
estimated exponential distribution to test against using the KS, since
this addresses the underlying mechanism generating the data.

Regardless, thanks for the suggestion. I'm picking up new tools left,
right and centre.

Angus.
-- 
AJC McMorland, PhD Student
Physiology, University of Auckland


From fdu.xiaojf at gmail.com  Tue Jul 17 02:13:57 2007
From: fdu.xiaojf at gmail.com (fdu.xiaojf at gmail.com)
Date: Tue, 17 Jul 2007 14:13:57 +0800
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469BB57C.6010809@gmail.com>
References: <469B6202.6020002@gmail.com>	<ce557a360707161100t437166ael190142f77564b68c@mail.gmail.com>
	<469BB57C.6010809@gmail.com>
Message-ID: <469C5E25.4060709@gmail.com>

Dominique Orban wrote:
 >> This problem happens because while fmin_cobyla is able to handle
 >> constraints, it nevertheless evaluates the objective function outside
 >> those constraints. I think this is a very serious bug, which limits
 >> the usefulness of fmin_cobyla. Unfortunately, I do not have a general
 >> solution.
 >
 > This behavior of Cobyla is not a bug.
 >
 > There are two kinds of methods for constrained problems in optimization:
 > feasible and infeasible methods. The iterates generated in a feasible method
 > will all satisfy the constraints. In an infeasible method, they may or may not
 > satisfy the constraints but, if convergence happens and if all goes well, the
 > iterates converge to a point which is (approximately) feasible. Cobyla is an
 > infeasible method.
 >
 > The fact that your objective function is not defined everywhere (because of 
the
 > log terms) suggests that you should not be using an infeasible method. 
Instead,
 > you could look into feasible interior-point methods (beware, there also are
 > infeasible interior-point methods... in fact, most of them are).
 >
 > NLPy features one such method, albeit a simple one, under the name TRIP (for
 > 'trust-region interior-point method'). For those who care about this sort of
 > thing, it is a purely primal method at the moment. I mostly implemented it 
as a
 > proof of concept in the NLPy framework, and plan to upgrade it in the 
future. It
 > will not treat equality constraints, but the inequalities can be as 
nonlinear as
 > you please, as long as they are twice differentiable. You can give it a try 
and
 > let me know what happens.

I couldn't find proper documents for TRIP method from NLPY. There are
only 3 small examples in the website of NLPY, and when I tried to get
the source code of NLPY:

$ svn co https://nlpy.svn.sourceforge.net/svnroot/nlpy/trunk/nlpy ./
svn: Can't convert string from native encoding to 'UTF-8':
svn: ?\208?\194?\189?\168 ?\206?\196?\177?\190?\206?\196?\181?\181.txt

I have tried to fix this problem, but with no success.

So could you point me some links that give a detailed explanation of how
to use TRIP?

 >
 > In addition to the constraints you already have, add new constraints stating
 > that the arguments of the logs must be >= 0. Interior-point methods ensure 
that
 > at each iteration, those quantities remain > 0 (strictly).
 >
 > If the constraints you end up with are *only bound constraints*, you can also
 > look into L-BFGS-B (which I believe is interfaced in SciPy), a projected
 > quasi-Newton method, or TRON (www.mcs.anl.gov/~more/tron which I don't 
think is
 > interfaced in SciPy), a projected Newton method. NLPy also contains a
 > gradient-projection method which I find surprisingly efficient, but also only
 > applies to bound constraints.
 >
 > Re-parametrizing your problem is, of course, another option.
 >
 > Good luck,
 > Dominique

Thanks a lot.


From fdu.xiaojf at gmail.com  Tue Jul 17 03:08:01 2007
From: fdu.xiaojf at gmail.com (fdu.xiaojf at gmail.com)
Date: Tue, 17 Jul 2007 15:08:01 +0800
Subject: [SciPy-user] Questions about scipy.optimize.fmin_cobyla
In-Reply-To: <469BB57C.6010809@gmail.com>
References: <469B6202.6020002@gmail.com>	<ce557a360707161100t437166ael190142f77564b68c@mail.gmail.com>
	<469BB57C.6010809@gmail.com>
Message-ID: <469C6AD1.4050603@gmail.com>

Hi,
Dominique Orban wrote:
 > Anne Archibald wrote:

 >
 > Re-parametrizing your problem is, of course, another option.
 >

I have reparameterized my problem, fmin_cobyla can run with out "math
domain error". But I get following message during minimization:

#--------begin of message---------------
Tolerance of 3.5527136788e-015 reached
Tolerance of 3.5527136788e-015 reached
Tolerance of 3.5527136788e-015 reached
Tolerance of 3.5527136788e-015 reached
Tolerance of -3.5527136788e-015 reached
Tolerance of -3.5527136788e-015 reached

    Return from subroutine COBYLA because the MAXFUN limit has been reached.

    NFVALS = 1000   F =-2.111828E+06    MAXCV = 6.311133E-07
    X = 8.497382E-01  -5.764801E+00  -7.701660E+00   5.986411E-01  -3.959199E+00
       -9.925761E-01  -6.380951E+00   1.034551E+00  -2.976923E+00   1.046551E-01
#--------end of message---------------

What does "Tolerance of 3.5527136788e-015 reached" mean ? Does it mean
some error occured or fmin_cobyla doesn't perform minimization well ?

Regards,

Xiao Jianfeng



From koepsell at gmail.com  Tue Jul 17 03:21:46 2007
From: koepsell at gmail.com (killian koepsell)
Date: Tue, 17 Jul 2007 00:21:46 -0700
Subject: [SciPy-user] (Mac) Close a plot window, crash IPython?
In-Reply-To: <3F008727-2F29-43D2-9E9E-B9A830211787@koepsell.de>
References: <88e473830707131310w57f2c3d3lce09a2f481c48c80@mail.gmail.com>
	<3F008727-2F29-43D2-9E9E-B9A830211787@koepsell.de>
Message-ID: <ca22c9920707170021p28c638c9n79e42a9907362a3@mail.gmail.com>

hi,


i observed a similar problem that persists in matplotlib 0.90.1 (python 2.5,
numpy 1.0.1, ipython 0.7.3).
the problem occurs when a window is closed and it seems to be specific to
the non-interactive mode using the GTK or GTKAgg backend. the following
short script runs ok once, but when i try to run it a second time, python
hangs:


# start script
import pylab as P
P.ioff()
P.figure()
P.close()
# stop script

this is probably a problem of matplotlib and we should continue this thread
on the matplotlib email list.


cheers,
  kilian






> From: "John Hunter" <jdh2358 at gmail.com>
> > Date: July 13, 2007 1:10:35 PM PDT
> > To: "SciPy Users List" < scipy-user at scipy.org>
> > Subject: Re: [SciPy-user] (Mac) Close a plot window, crash IPython?
> > Reply-To: SciPy Users List < scipy-user at scipy.org>
> >
> > On 7/13/07, David Warde-Farley <david.warde.farley at utoronto.ca> wrote:
> >
> >> It does seem as though the *exact* same bug was reported today on
> >> matplotlib-devel (what are the odds?), I shall checkout the svn
> >> version and see if that fixes it, and make them aware of the other
> >> WX-
> >> related bug as well.
> >
> > I hope this does fix your problem, but the bug that was fixed was also
> > recently introduced (in a svn commit after the 0.90.1 release) so it
> > may not be your problem.  But if you can update from svn and see if
> > the problem is still there, that would be a great start.  See you on
> > the matplotlib-devel side :-)
> >
> > JDH
> > _______________________________________________
> > 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/20070717/fb25f651/attachment.html>

From massimo.sandal at unibo.it  Tue Jul 17 08:11:33 2007
From: massimo.sandal at unibo.it (massimo sandal)
Date: Tue, 17 Jul 2007 14:11:33 +0200
Subject: [SciPy-user] [OT] advice about publishing a scientific software
In-Reply-To: <469BDD3B.3060008@gmail.com>
References: <469B6CCF.8090506@unibo.it><469B9402.5070403@gmail.com>	<469B9DC5.7050301@unibo.it><469B9FFD.5070008@gmail.com><469BABA2.4030608@unibo.it>	<Mahogany-0.67.0-708-20070716-135158.00@american.edu>
	<469BDD3B.3060008@gmail.com>
Message-ID: <469CB1F5.7050907@unibo.it>

Robert Kern ha scritto:
> Indeed. To rephrase a little bit more clearly: "I would think that only the
> quality of the code at the time of submission of your paper would be important."
> I don't think reviewers will be going back through your SVN logs. Sharing code
> before publication is expected, I think. As Chris Lee suggests, though, you have
> to be very careful about keeping track of contributions, more so than with other
> open source projects.

Oh, ok. This makes me happier. Of course yes, every contributor should 
give me name and address etc. for name in publication.

Thanks for your answers, they're clarifying a lot of things to me.

m.

-- 
Massimo Sandal
University of Bologna
Department of Biochemistry "G.Moruzzi"

snail mail:
Via Irnerio 48, 40126 Bologna, Italy

email:
massimo.sandal at unibo.it

tel: +39-051-2094388
fax: +39-051-2094387
-------------- next part --------------
A non-text attachment was scrubbed...
Name: massimo.sandal.vcf
Type: text/x-vcard
Size: 274 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070717/c9cf9365/attachment.vcf>

From mathias.wagner at physik.tu-darmstadt.de  Tue Jul 17 09:27:30 2007
From: mathias.wagner at physik.tu-darmstadt.de (Mathias Wagner)
Date: Tue, 17 Jul 2007 15:27:30 +0200
Subject: [SciPy-user] Finding a point where a function becomes constant
Message-ID: <200707171527.32059.mathias.wagner@physik.tu-darmstadt.de>

Hi,

I have a numerical function defined f(x) that is zero for x < x_0 and f(x)>0 
for x > x_0. I have to find x_0.
The problem with conventional root finding is that if the algorithm evaluates 
the function for any x < x_0 it will return this as a solution.
Does anyone know whether there is some suitable algorithm for this problem or 
do I have to modify an existing method like bisection for my needs?

Mathias

-- 
// ***************************************************************
// ** 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 aisaac at american.edu  Tue Jul 17 09:46:17 2007
From: aisaac at american.edu (Alan G Isaac)
Date: Tue, 17 Jul 2007 09:46:17 -0400
Subject: [SciPy-user] Finding a point where a function becomes constant
In-Reply-To: <200707171527.32059.mathias.wagner@physik.tu-darmstadt.de>
References: <200707171527.32059.mathias.wagner@physik.tu-darmstadt.de>
Message-ID: <Mahogany-0.67.0-816-20070717-094617.00@american.edu>

On Tue, 17 Jul 2007, Mathias Wagner apparently wrote:
> I have a numerical function defined f(x) that is zero for x < x_0 and f(x)>0 
> for x > x_0. I have to find x_0. 
> The problem with conventional root finding is that if the algorithm evaluates 
> the function for any x < x_0 it will return this as a solution. 
> Does anyone know whether there is some suitable algorithm for this problem or 
> do I have to modify an existing method like bisection for my needs? 

This one does not require any modification at all, I think.
(License: MIT.)

Cheers,
Alan Isaac


def bisect(f, x1, x2, eps=1e-8):
    f1, f2 = f(x1), f(x2)
    if f1*f2 > 0:
        raise ValueError
    #initialize xneg, xpos
    xneg, xpos = (x1,x2) if(f2>0) else (x2,x1)
    while xpos-xneg > eps:
        xmid = (xneg+xpos)/2
        if f(xmid) > 0:
            xpos = xmid
        else:
            xneg = xmid
    return (xneg+xpos)/2






From fie.pye at atlas.cz  Tue Jul 17 16:53:31 2007
From: fie.pye at atlas.cz (fie.pye at atlas.cz)
Date: Tue, 17 Jul 2007 20:53:31 GMT
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
Message-ID: <96133d6a05de4caa883a0724251e98a5@b68193df7bb044b79da5716a3035062c>

Hi David, hi Vincent,

thank you for your response. 
On my computer I have three gcc compilers: gcc4.1, gcc4.2 and gcc3.4.6. Since you mentioned possible problem with gfortran I decided to compile BLAS, LAPACK, python2.5 and numpy again with gcc3.4.6 and f77. 
For numpy installation I used both
python2.5 setup.py build
python2.5 setup.py install
or
python2.5 setup.py build build_ext -libg2c install
but when importing numpy I got the same message

>>> import numpy
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/__init__.py", line 43, in <module>
    import linalg
  File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in <module>
    from linalg import *
  File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in <module>
    from numpy.linalg import lapack_lite
ImportError: /usr/local/lib64/lapack/liblapack.so: undefined symbol: slamch_

I included all installation scripts and log files to enclosed file. 

I also began to install from source rpm but I am not familiar with that installation therefore I stopped after building. What does it mean refblas. 
http://software.opensuse.org/download/home:/ashigabou/Fedora_Extras_6/src/


Installation information
ldd /usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so
        liblapack.so => /usr/local/lib64/lapack/liblapack.so (0x00002aaaaacb4000)
        libblas.so => /usr/local/lib64/blas/libblas.so (0x00002aaaab636000)
        libg2c.so.0 => /usr/lib64/libg2c.so.0 (0x00002aaaab88a000)
        libm.so.6 => /lib64/libm.so.6 (0x00002aaaabada000)
        libgcc_s.so.1 => /usr/local/lib64/libgcc_s.so.1 (0x00002aaaabd5d000)
        libc.so.6 => /lib64/libc.so.6 (0x00002aaaabf6a000)
        /lib64/ld-linux-x86-64.so.2 (0x0000555555554000)

Best regards.
Fie Pye

PS: I always replay to e-mail. Why my respons does not append a tree and starts new contribution? 



>---------------------------------------------------------
>Od: David Cournapeau
>P?ijato: 12.7.2007 6:12:47
>P?edm?t: Re: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
>
>
>
>> Hi David.
>
>>
>
>> Thank you for your response. Based on your notes I recompiled BLAS and LAPACK libraries and build and installed python2.5.1 and numpy again. Unfortunately the error appears again
>
>>
>
>>>>> import numpy
>
>> Traceback (most recent call last):
>
>>   File "<stdin>", line 1, in <module>
>
>>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/__init__.py", line 43, in <module>
>
>>     import linalg
>
>>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in <module>
>
>>     from linalg import *
>
>>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in <module>
>
>>     from numpy.linalg import lapack_lite
>
>> ImportError: /usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so: undefined symbol: _gfortran_concat_string
>
>>
>
>This is typically because you did not use the same fortran compiler 
>
>everywhere (g77 vs gfortran): the default fortran compiler on fedora 
>
>core (at least since 5) is gfortran, but this is not the default for 
>
>numpy or scipy. Again, this is exactly for this kind of things that I 
>
>packaged blas/lapack/atlas/numpy/scipy. Did you try building the source 
>
>rpms instead of building everything from sources ?
>
>
>
>cheers,
>
>
>
>David
>
>_______________________________________________
>
>SciPy-user mailing list
>
>SciPy-user at scipy.org
>
>http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>
>
>
>
>
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: numpy_install.tar.gz
Type: application/x-gzip
Size: 54364 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070717/f1dc31b5/attachment.bin>

From david at ar.media.kyoto-u.ac.jp  Tue Jul 17 21:47:26 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Wed, 18 Jul 2007 10:47:26 +0900
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
In-Reply-To: <96133d6a05de4caa883a0724251e98a5@b68193df7bb044b79da5716a3035062c>
References: <96133d6a05de4caa883a0724251e98a5@b68193df7bb044b79da5716a3035062c>
Message-ID: <469D712E.6060303@ar.media.kyoto-u.ac.jp>

fie.pye at atlas.cz wrote:
> Hi David, hi Vincent,
>
> thank you for your response. 
> On my computer I have three gcc compilers: gcc4.1, gcc4.2 and gcc3.4.6. Since you mentioned possible problem with gfortran I decided to compile BLAS, LAPACK, python2.5 and numpy again with gcc3.4.6 and f77.
To be precise, what is problematic is mixing fortran compilers (for 
example using gfortran for BLAS and g77 for LAPACK). gfortran by itself 
is not that problematic.
>  
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/__init__.py", line 43, in <module>
>     import linalg
>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in <module>
>     from linalg import *
>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in <module>
>     from numpy.linalg import lapack_lite
> ImportError: /usr/local/lib64/lapack/liblapack.so: undefined symbol: slamch_
The problem is the way you build the shared library is wrong: you cannot 
just use g77 -shared -o soname *.o, because LAPACK object files are in 
several directories.

One correct way is to use the static library .a, uncompress it in a 
directory, and build the shared library from that. There is a problem 
when several object files have the same name (because you cannot have 
several files with the same name in a directory), though, so extra care 
must be taken.
>
> I included all installation scripts and log files to enclosed file. 
>
> I also began to install from source rpm but I am not familiar with that installation therefore I stopped after building. What does it mean refblas. 
I cannot use blas and lapack from officiel repository because they do 
not work very well, and I use a different name to avoid name clashes. 
You should build the following srpms in this order:

- refblas
- lapack3
- python-numpy
- python-scipy

Building them is not really different than any other source rpms. Since 
you are using a distribution I've never tested, there is a good chance 
it does not work out of the box, but I would rather help you solving 
those problems so that everybody can benefit from them rather than 
explaining how to build blas and lapack properly (which is really error 
prone and non trivial).

David


From sransom at nrao.edu  Tue Jul 17 22:18:41 2007
From: sransom at nrao.edu (Scott Ransom)
Date: Tue, 17 Jul 2007 22:18:41 -0400
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
In-Reply-To: <469D712E.6060303@ar.media.kyoto-u.ac.jp>
References: <96133d6a05de4caa883a0724251e98a5@b68193df7bb044b79da5716a3035062c>
	<469D712E.6060303@ar.media.kyoto-u.ac.jp>
Message-ID: <20070718021840.GA28479@ssh.cv.nrao.edu>

> The problem is the way you build the shared library is wrong: you cannot 
> just use g77 -shared -o soname *.o, because LAPACK object files are in 
> several directories.
> 
> One correct way is to use the static library .a, uncompress it in a 
> directory, and build the shared library from that. There is a problem 
> when several object files have the same name (because you cannot have 
> several files with the same name in a directory), though, so extra care 
> must be taken.

You might be able to do something like this.  It has worked for me
in the past (with GNU ld):

ld -shared -o libfoo.so --whole-archive libfoo.a

Scott

-- 
Scott M. Ransom            Address:  NRAO
Phone:  (434) 296-0320               520 Edgemont Rd.
email:  sransom at nrao.edu             Charlottesville, VA 22903 USA
GPG Fingerprint: 06A9 9553 78BE 16DB 407B  FFCA 9BFA B6FF FFD3 2989


From openopt at ukr.net  Wed Jul 18 05:11:48 2007
From: openopt at ukr.net (dmitrey)
Date: Wed, 18 Jul 2007 12:11:48 +0300
Subject: [SciPy-user] changes in scipy.optimization.tnc (v 1.2 to 1.3)
Message-ID: <469DD954.4030208@ukr.net>

Hi all,
(excuse my English)
I was asked to connect tnc 1.3 instead of the v 1.2 that is present in 
scipy svn for now (ticket 296).
Let me remember you that tnc is a routine for solving box-bounded 
non-linear optimization problems

f(x)->min
subjected to
lb <=x <= ub
(you should use None if some lb-ub coords are absent)

Before closing the ticket I want to note that it will not provide 
backward compability (and hear your suggestions if they will be 
obtained), because some things had been changed in tnc, see below some 
statements from /tnc/HISTORY (please pay attention to the lines 
"Warning: API changes"):

# TNC Release History
# $Jeannot: HISTORY,v 1.15 2005/01/28 18:27:31 js Exp $

01/28/2005, V1.3   : Fix a bug in the anti-zigzaging mechanism (many thanks
                     to S.G. NASH).
                     Warning: API changes: refined stopping criterions 
(xtol,
                     pgtol). ftol is no more relative to the value of f.
                     new parameter offset to translate the coordinates
04/18/2004, V1.2.5 : n==0 is now valid
04/14/2004, V1.2.4 : Fix a potential bug in the Python interface 
(infinity==0)
04/14/2004, V1.2.3 : Fix a bug in the Python interface (reference counting)
04/13/2004, V1.2.2 : Fix a bug in the Python interface (memory allocation)
04/13/2004, V1.2.1 : Fix a bug in the Python interface (scaling ignored)
04/03/2004, V1.2   : Memory allocation checks

HTH, Dmitrey.



From aisaac at american.edu  Wed Jul 18 08:22:38 2007
From: aisaac at american.edu (Alan G Isaac)
Date: Wed, 18 Jul 2007 08:22:38 -0400
Subject: [SciPy-user] changes in scipy.optimization.tnc (v 1.2 to 1.3)
In-Reply-To: <469DD954.4030208@ukr.net>
References: <469DD954.4030208@ukr.net>
Message-ID: <Mahogany-0.67.0-1532-20070718-082238.00@american.edu>

On Wed, 18 Jul 2007, dmitrey apparently wrote:
> I want to note that it will not provide backward 
> compability (and hear your suggestions
>                      Warning: API changes: refined 
>                      stopping criterions
> (xtol, 
>                      pgtol). ftol is no more relative to the value of f.
>                      new parameter offset to translate the coordinates


Hi Dmitrey,

Regarding http://projects.scipy.org/scipy/scipy/ticket/296
Thanks for working on this.

Is your question the following?
        Should you implement any warnings etc. if someone 
        appears to be using the old API?
Jarrod, any commments?
Anyone else?

Cheers,
Alan Isaac






From scipy-pma at agapow.net  Wed Jul 18 11:09:45 2007
From: scipy-pma at agapow.net (Paul-Michael Agapow)
Date: Wed, 18 Jul 2007 16:09:45 +0100
Subject: [SciPy-user] Weave woes
Message-ID: <A9D0A7C0-C04A-45E1-B021-53322A1B363E@agapow.net>


I've got errors installing and using weave, that persist across  
different installations and Python versions.  My google-fu has failed  
me and an identical error reported on the mailing list some time ago  
went unanswered.

Symptoms: First I installed weave from svn into Python2.5 and ran  
weave.test():

      Found 1 tests for weave.ast_tools
      [...snip...]
      Found 26 tests for weave.catalog
    building extensions here: /Users/agapow/.python25_compiled/m3
      [...snip...]
      Found 3 tests for weave.standard_array_spec
    /Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/ 
site-packages/weave/tests/test_wx_spec.py:16: DeprecationWarning: The  
wxPython compatibility package is no longer automatically generated  
or activly maintained.  Please switch to the wx package as soon as  
possible.
      import wxPython
      Found 0 tests for weave.wx_spec
      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
    ............................removing '/tmp/ 
tmptBN1Qxcat_test' (and everything under it)
    .removing '/tmp/tmpY2WiLfcat_test' (and everything under it)
    ..............................F..F.................................. 
...........................
     
======================================================================
    FAIL: check_1d_3  
(weave.tests.test_size_check.test_dummy_array_indexing)
     
----------------------------------------------------------------------
    Traceback (most recent call last):
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/tests/test_size_check.py", line 168, in  
check_1d_3
        self.generic_1d('a[-11:]')
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/tests/test_size_check.py", line 135, in  
generic_1d
        self.generic_wrap(a,expr)
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/tests/test_size_check.py", line 127, in  
generic_wrap
        self.generic_test(a,expr,desired)
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/tests/test_size_check.py", line 123, in  
generic_test
        assert_array_equal(actual,desired, expr)
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/numpy/testing/utils.py", line 223, in  
assert_array_equal
        verbose=verbose, header='Arrays are not equal')
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/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  
(weave.tests.test_size_check.test_dummy_array_indexing)
     
----------------------------------------------------------------------
    Traceback (most recent call last):
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/tests/test_size_check.py", line 174, in  
check_1d_6
        self.generic_1d('a[:-11]')
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/tests/test_size_check.py", line 135, in  
generic_1d
        self.generic_wrap(a,expr)
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/tests/test_size_check.py", line 127, in  
generic_wrap
        self.generic_test(a,expr,desired)
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/tests/test_size_check.py", line 123, in  
generic_test
        assert_array_equal(actual,desired, expr)
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/numpy/testing/utils.py", line 223, in  
assert_array_equal
        verbose=verbose, header='Arrays are not equal')
      File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/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])

I'm uncertain if the "__bad_path__" message is important, but the two  
errors may relfect issues with numpy (v1.0.1). I installed weave into  
Python2.4 for identical symptoms. I then installed the whole Scipy  
package just to be sure (weave 0.4.9, numpy 1.0.4.dev3882). No change.

Along the way - mindful that maybe the tests were broken - I tried  
out a simple line of weave::

    >>> a = 1; weave.inline('printf("%d\\n",a);',['a'])

which gave::

Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
   File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/inline_tools.py", line 325, in inline
     results = attempt_function_call(code,local_dict,global_dict)
   File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/inline_tools.py", line 375, in  
attempt_function_call
     function_list = function_catalog.get_functions(code,module_dir)
   File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/catalog.py", line 611, in get_functions
     function_list = self.get_cataloged_functions(code)
   File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/catalog.py", line 524, in  
get_cataloged_functions
     cat = get_catalog(path,mode)
   File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/site-packages/weave/catalog.py", line 294, in get_catalog
     or os.path.exists(catalog_file):
   File "/Library/Frameworks/Python.framework/Versions/2.5/lib/ 
python2.5/posixpath.py", line 171, in exists
     st = os.stat(path)
TypeError: coercing to Unicode: need string or buffer, NoneType found


Any ideas on what to try next? (Technical details: MacOSX 10.4.10  
Intel macBook,  gcc 4.0.1.)

--
Dr Paul-Michael Agapow: VieDigitale / Inst. for Animal Health
pma at viedigitale.com / paul-michael.agapow at bbsrc.ac.uk



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070718/b9a0b80b/attachment.html>

From nwagner at iam.uni-stuttgart.de  Wed Jul 18 11:15:13 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Wed, 18 Jul 2007 17:15:13 +0200
Subject: [SciPy-user] Weave woes
In-Reply-To: <A9D0A7C0-C04A-45E1-B021-53322A1B363E@agapow.net>
References: <A9D0A7C0-C04A-45E1-B021-53322A1B363E@agapow.net>
Message-ID: <469E2E81.9060900@iam.uni-stuttgart.de>

Paul-Michael Agapow wrote:
>
> I've got errors installing and using weave, that persist across
> different installations and Python versions.  My google-fu has failed
> me and an identical error reported on the mailing list some time ago
> went unanswered.
>
> Symptoms: First I installed weave from svn into Python2.5 and ran
> weave.test():
>
>      Found 1 tests for weave.ast_tools
>      [...snip...]
>      Found 26 tests for weave.catalog
>    building extensions here: /Users/agapow/.python25_compiled/m3
>      [...snip...]
>      Found 3 tests for weave.standard_array_spec
>   
> /Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/tests/test_wx_spec.py:16:
> DeprecationWarning: The wxPython compatibility package is no longer
> automatically generated or activly maintained.  Please switch to the
> wx package as soon as possible.
>      import wxPython
>      Found 0 tests for weave.wx_spec
>      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
>    ............................removing '/tmp/tmptBN1Qxcat_test' (and
> everything under it)
>    .removing '/tmp/tmpY2WiLfcat_test' (and everything under it)
>   
> ..............................F..F.............................................................
>    ======================================================================
>    FAIL: check_1d_3
> (weave.tests.test_size_check.test_dummy_array_indexing)
>    ----------------------------------------------------------------------
>    Traceback (most recent call last):
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/tests/test_size_check.py",
> line 168, in check_1d_3
>        self.generic_1d('a[-11:]')
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/tests/test_size_check.py",
> line 135, in generic_1d
>        self.generic_wrap(a,expr)
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/tests/test_size_check.py",
> line 127, in generic_wrap
>        self.generic_test(a,expr,desired)
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/tests/test_size_check.py",
> line 123, in generic_test
>        assert_array_equal(actual,desired, expr)
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/numpy/testing/utils.py",
> line 223, in assert_array_equal
>        verbose=verbose, header='Arrays are not equal')
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/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
> (weave.tests.test_size_check.test_dummy_array_indexing)
>    ----------------------------------------------------------------------
>    Traceback (most recent call last):
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/tests/test_size_check.py",
> line 174, in check_1d_6
>        self.generic_1d('a[:-11]')
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/tests/test_size_check.py",
> line 135, in generic_1d
>        self.generic_wrap(a,expr)
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/tests/test_size_check.py",
> line 127, in generic_wrap
>        self.generic_test(a,expr,desired)
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/tests/test_size_check.py",
> line 123, in generic_test
>        assert_array_equal(actual,desired, expr)
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/numpy/testing/utils.py",
> line 223, in assert_array_equal
>        verbose=verbose, header='Arrays are not equal')
>      File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/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])
>
> I'm uncertain if the "__bad_path__" message is important, but the two
> errors may relfect issues with numpy (v1.0.1). I installed weave into
> Python2.4 for identical symptoms. I then installed the whole Scipy
> package just to be sure (weave 0.4.9, numpy 1.0.4.dev3882). No change.
>
> Along the way - mindful that maybe the tests were broken - I tried out
> a simple line of weave::
>
>    >>> a = 1; weave.inline('printf("%d\\n",a);',['a'])
>
> which gave::
>
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/inline_tools.py",
> line 325, in inline
>     results = attempt_function_call(code,local_dict,global_dict)
>   File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/inline_tools.py",
> line 375, in attempt_function_call
>     function_list = function_catalog.get_functions(code,module_dir)
>   File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/catalog.py",
> line 611, in get_functions
>     function_list = self.get_cataloged_functions(code)
>   File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/catalog.py",
> line 524, in get_cataloged_functions
>     cat = get_catalog(path,mode)
>   File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/weave/catalog.py",
> line 294, in get_catalog
>     or os.path.exists(catalog_file):
>   File
> "/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/posixpath.py",
> line 171, in exists
>     st = os.stat(path)
> TypeError: coercing to Unicode: need string or buffer, NoneType found
>
>
> Any ideas on what to try next? (Technical details: MacOSX 10.4.10
> Intel macBook,  gcc 4.0.1.)
>
> --
> Dr Paul-Michael Agapow: VieDigitale / Inst. for Animal Health
> pma at viedigitale.com <mailto:pma at viedigitale.com> /
> paul-michael.agapow at bbsrc.ac.uk <mailto:paul-michael.agapow at bbsrc.ac.uk>
>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   
I can confirm failures on SuSE 64-bit (latest svn)

python -i
/usr/lib64/python2.4/site-packages/scipy/weave/tests/test_size_check.py
  Found 74 tests for weave.size_check
  Found 0 tests for __main__
............................F..F..........................................
======================================================================
FAIL: check_1d_3 (weave.tests.test_size_check.test_dummy_array_indexing)
----------------------------------------------------------------------
Traceback (most recent call last):
  File
"/usr/lib64/python2.4/site-packages/scipy/weave/tests/test_size_check.py",
line 168, in check_1d_3
    self.generic_1d('a[-11:]')
  File
"/usr/lib64/python2.4/site-packages/scipy/weave/tests/test_size_check.py",
line 135, in generic_1d
    self.generic_wrap(a,expr)
  File
"/usr/lib64/python2.4/site-packages/scipy/weave/tests/test_size_check.py",
line 127, in generic_wrap
    self.generic_test(a,expr,desired)
  File
"/usr/lib64/python2.4/site-packages/scipy/weave/tests/test_size_check.py",
line 123, in generic_test
    assert_array_equal(actual,desired, expr)
  File "/usr/lib64/python2.4/site-packages/numpy/testing/utils.py", line
223, in assert_array_equal
    verbose=verbose, header='Arrays are not equal')
  File "/usr/lib64/python2.4/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 (weave.tests.test_size_check.test_dummy_array_indexing)
----------------------------------------------------------------------
Traceback (most recent call last):
  File
"/usr/lib64/python2.4/site-packages/scipy/weave/tests/test_size_check.py",
line 174, in check_1d_6
    self.generic_1d('a[:-11]')
  File
"/usr/lib64/python2.4/site-packages/scipy/weave/tests/test_size_check.py",
line 135, in generic_1d
    self.generic_wrap(a,expr)
  File
"/usr/lib64/python2.4/site-packages/scipy/weave/tests/test_size_check.py",
line 127, in generic_wrap
    self.generic_test(a,expr,desired)
  File
"/usr/lib64/python2.4/site-packages/scipy/weave/tests/test_size_check.py",
line 123, in generic_test
    assert_array_equal(actual,desired, expr)
  File "/usr/lib64/python2.4/site-packages/numpy/testing/utils.py", line
223, in assert_array_equal
    verbose=verbose, header='Arrays are not equal')
  File "/usr/lib64/python2.4/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 74 tests in 0.248s

FAILED (failures=2)

Nils


 



From rmay at ou.edu  Wed Jul 18 14:14:50 2007
From: rmay at ou.edu (Ryan May)
Date: Wed, 18 Jul 2007 13:14:50 -0500
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
Message-ID: <469E589A.9080202@ou.edu>

Hi,

I've been using Travis's pynetcdf package (0.7) for doing NetCDF IO for 
a lot of my code.  I recently updated my Gentoo install to use Python 
2.5, and now previously working code crashes (segmentation fault, no 
python exception) when I try to access an array of data.  A GDB 
backtrace puts the error inside _netcdf.so. I haven't recompiled with 
debugging symbols to try to track this down any further.

Has anyone else had any problems with pynetcdf and python 2.5?

I'm aware of the other options (scipy.sandbox.netcdf, 
Scientific.IO.NetCDF, and pycdf).  Is there an accepted package for 
NetCDF in python?

Thanks,

Ryan

-- 
Ryan May
Graduate Research Assistant
School of Meteorology
University of Oklahoma


From wjdandreta at att.net  Wed Jul 18 15:34:15 2007
From: wjdandreta at att.net (Bill Dandreta)
Date: Wed, 18 Jul 2007 15:34:15 -0400
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
In-Reply-To: <469E589A.9080202@ou.edu>
References: <469E589A.9080202@ou.edu>
Message-ID: <469E6B37.6030602@att.net>

I don't know if this has any bearing on your problem but Python 2.5 is 
hard masked on Gentoo do to problems on 64 bit systems with packages 
depending on numeric.

Ryan May wrote:
> Hi,
>
> I've been using Travis's pynetcdf package (0.7) for doing NetCDF IO for 
> a lot of my code.  I recently updated my Gentoo install to use Python 
> 2.5, and now previously working code crashes (segmentation fault, no 
> python exception) when I try to access an array of data.  A GDB 
> backtrace puts the error inside _netcdf.so. I haven't recompiled with 
> debugging symbols to try to track this down any further.
>
> Has anyone else had any problems with pynetcdf and python 2.5?
>
> I'm aware of the other options (scipy.sandbox.netcdf, 
> Scientific.IO.NetCDF, and pycdf).  Is there an accepted package for 
> NetCDF in python?
>
> Thanks,
>
> Ryan
>   
-- 
Bill

wjdandreta at att.net

Gentoo Linux X86_64 2.6.20-gentoo-r8

Reclaim Your Inbox with http://www.mozilla.org/products/thunderbird/

All things cometh to he who waiteth as long as he who waiteth worketh like hell while he waiteth.



From rmay at ou.edu  Wed Jul 18 15:41:39 2007
From: rmay at ou.edu (Ryan May)
Date: Wed, 18 Jul 2007 14:41:39 -0500
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
In-Reply-To: <469E6B37.6030602@att.net>
References: <469E589A.9080202@ou.edu> <469E6B37.6030602@att.net>
Message-ID: <469E6CF3.1070800@ou.edu>

Bill Dandreta wrote:
> I don't know if this has any bearing on your problem but Python 2.5 is 
> hard masked on Gentoo do to problems on 64 bit systems with packages 
> depending on numeric.
> 

Thanks, but, yeah, I am aware of that. (I did my research before 
unmasking it.  It's driven me nuts that a stable python release has been 
waiting 10+ months to get into my "cutting edge" distro.)  There weren't 
any bugs against the packages I use, though I have found that plugins 
for gedit 2.16 don't like python, but this has been fixed in 2.18.  One 
of the reasons I originally chose pynetcdf was its support for numpy, 
not Numeric, so that's not the issue here.

Ryan

-- 
Ryan May
Graduate Research Assistant
School of Meteorology
University of Oklahoma


From emanuelez at gmail.com  Wed Jul 18 15:44:03 2007
From: emanuelez at gmail.com (Emanuele Zattin)
Date: Wed, 18 Jul 2007 21:44:03 +0200
Subject: [SciPy-user] leastsq and fitting quality
Message-ID: <d7b942cd0707181244u7427d260g2897a3f50250baf9@mail.gmail.com>

Hello,
i've been looking over the net and in the archives of this mailing
list, but without any conclusive result so here i am asking it again
:-P
i'm using leastsq many times to perform star detection over an
astronomical image. i can see that usually the result is very
satisfying, but sometimes some bad result comes out. i would like to
be able to find those cases so that i can handle the situation.
i tried analysing the residuals, or their sum, but it does not seem to
be a particularly good solution. maybe the covariance matrix in the
extended output contains some useful information, but so far i was not
able to use it properly. do u have any hint?


From dwf at cs.toronto.edu  Wed Jul 18 16:41:55 2007
From: dwf at cs.toronto.edu (David Warde-Farley)
Date: Wed, 18 Jul 2007 16:41:55 -0400
Subject: [SciPy-user] Fastest way to save/load fairly large (sparse) matrices
Message-ID: <CA0C8056-5184-445D-B2CE-C7B1E308B8D4@cs.toronto.edu>

Hi folks,

I've been wrestling with this problem for some time now. Essentially  
I'm using Numpy/Scipy as the backend in a python-based web  
application and need a way of serializing the contents of a fairly  
large, but sparsely populated, symmetric matrix to disk. So far I've  
experimented with just ASCII dumps and pickling, but neither really  
has the kind of speed that I'm looking for (and is kind of necessary  
for an interactive web application). Is there some better way?

To give you an idea of the scale, I'm hoping to get this to work with  
20000x20000 matrices that are about 1% full.

Thanks,

David


From rmay at ou.edu  Wed Jul 18 16:45:37 2007
From: rmay at ou.edu (Ryan May)
Date: Wed, 18 Jul 2007 15:45:37 -0500
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
In-Reply-To: <469E589A.9080202@ou.edu>
References: <469E589A.9080202@ou.edu>
Message-ID: <469E7BF1.7020409@ou.edu>

Ryan May wrote:
> Hi,
> 
> I've been using Travis's pynetcdf package (0.7) for doing NetCDF IO for 
> a lot of my code.  I recently updated my Gentoo install to use Python 
> 2.5, and now previously working code crashes (segmentation fault, no 
> python exception) when I try to access an array of data.  A GDB 
> backtrace puts the error inside _netcdf.so. I haven't recompiled with 
> debugging symbols to try to track this down any further.
>

If anyone knows a bit about this code, GDB points the crash down to a 
SIGFPE at line 1435:

dims[d] = (indices[i].stop-indices[i].start-1)/indices[i].stride+1;

According to GDB, indices[i].stride is 0, which causes the crash. 
Anybody have a suggestion of what to look at next?

Ryan

-- 
Ryan May
Graduate Research Assistant
School of Meteorology
University of Oklahoma


From strawman at astraw.com  Wed Jul 18 16:59:06 2007
From: strawman at astraw.com (Andrew Straw)
Date: Wed, 18 Jul 2007 13:59:06 -0700
Subject: [SciPy-user] Fastest way to save/load fairly large (sparse)
	matrices
In-Reply-To: <CA0C8056-5184-445D-B2CE-C7B1E308B8D4@cs.toronto.edu>
References: <CA0C8056-5184-445D-B2CE-C7B1E308B8D4@cs.toronto.edu>
Message-ID: <469E7F1A.9010800@astraw.com>

Dear David, this code works for me. It's not the most elegant, but at 
least you get the idea. See the test() function for the example usage.

David Warde-Farley wrote:
> Hi folks,
> 
> I've been wrestling with this problem for some time now. Essentially  
> I'm using Numpy/Scipy as the backend in a python-based web  
> application and need a way of serializing the contents of a fairly  
> large, but sparsely populated, symmetric matrix to disk. So far I've  
> experimented with just ASCII dumps and pickling, but neither really  
> has the kind of speed that I'm looking for (and is kind of necessary  
> for an interactive web application). Is there some better way?
> 
> To give you an idea of the scale, I'm hoping to get this to work with  
> 20000x20000 matrices that are about 1% full.
> 
> Thanks,
> 
> David
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user

-------------- next part --------------
A non-text attachment was scrubbed...
Name: ashelf.py
Type: text/x-python
Size: 3201 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070718/8bb776b9/attachment.py>

From strawman at astraw.com  Wed Jul 18 17:03:36 2007
From: strawman at astraw.com (Andrew Straw)
Date: Wed, 18 Jul 2007 14:03:36 -0700
Subject: [SciPy-user] Fastest way to save/load fairly large
	(sparse)	matrices
In-Reply-To: <469E7F1A.9010800@astraw.com>
References: <CA0C8056-5184-445D-B2CE-C7B1E308B8D4@cs.toronto.edu>
	<469E7F1A.9010800@astraw.com>
Message-ID: <469E8028.1020502@astraw.com>

Whoops the file I sent said "all rights reserved". I hereby un-reserve 
some rights and release that module under the MIT license. I have 
updated my copy of the file to include the following:

> Copyright (c) 2005-2007, California Institute of Technology
> 
> Permission is hereby granted, free of charge, to any person obtaining a copy
> of this software and associated documentation files (the "Software"), to deal
> in the Software without restriction, including without limitation the rights
> to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
> copies of the Software, and to permit persons to whom the Software is
> furnished to do so, subject to the following conditions:
> 
> The above copyright notice and this permission notice shall be included in
> all copies or substantial portions of the Software.
> 
> THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
> IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
> AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
> LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
> OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
> THE SOFTWARE.



From strawman at astraw.com  Wed Jul 18 17:06:34 2007
From: strawman at astraw.com (Andrew Straw)
Date: Wed, 18 Jul 2007 14:06:34 -0700
Subject: [SciPy-user] Fastest way to save/load fairly large
	(sparse)	matrices
In-Reply-To: <469E7F1A.9010800@astraw.com>
References: <CA0C8056-5184-445D-B2CE-C7B1E308B8D4@cs.toronto.edu>
	<469E7F1A.9010800@astraw.com>
Message-ID: <469E80DA.305@astraw.com>

I'm not doing so well today. The file I attached was outdated (with no 
test() function). I attach the correct file here.

-Andrew "working on too many computers at once for his own good" Straw
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ashelf.py
Type: text/x-python
Size: 5178 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070718/0ef8c28c/attachment.py>

From dwf at cs.toronto.edu  Wed Jul 18 17:16:19 2007
From: dwf at cs.toronto.edu (David Warde-Farley)
Date: Wed, 18 Jul 2007 17:16:19 -0400
Subject: [SciPy-user] Fastest way to save/load fairly large
	(sparse)	matrices
In-Reply-To: <469E80DA.305@astraw.com>
References: <CA0C8056-5184-445D-B2CE-C7B1E308B8D4@cs.toronto.edu>
	<469E7F1A.9010800@astraw.com> <469E80DA.305@astraw.com>
Message-ID: <BC037701-2743-4116-B96F-3669AA59CDA3@cs.toronto.edu>

Hi Andrew,

This is exactly what I needed. Thanks so much!

David


On 18-Jul-07, at 5:06 PM, Andrew Straw wrote:

> I'm not doing so well today. The file I attached was outdated (with  
> no test() function). I attach the correct file here.
>
> -Andrew "working on too many computers at once for his own good" Straw
> # Author: Andrew D. Straw
> # Copyright (C) 2005-2007, California Institute of Technology
> #
> # Permission is hereby granted, free of charge, to any person  
> obtaining a copy
> # of this software and associated documentation files (the  
> "Software"), to deal
> # in the Software without restriction, including without limitation  
> the rights
> # to use, copy, modify, merge, publish, distribute, sublicense, and/ 
> or sell
> # copies of the Software, and to permit persons to whom the  
> Software is
> # furnished to do so, subject to the following conditions:
> #
> # The above copyright notice and this permission notice shall be  
> included in
> # all copies or substantial portions of the Software.
> #
> # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,  
> EXPRESS OR
> # IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF  
> MERCHANTABILITY,
> # FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT  
> SHALL THE
> # AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR  
> OTHER
> # LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,  
> ARISING FROM,
> # OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER  
> DEALINGS IN
> # THE SOFTWARE.
>
>
>
> import os
> import numpy
> import scipy
> import scipy.sparse
> from cgtypes import vec3 # HACK
>
> # include this in files:
> # import os
> # ashelf_datadir = os.path.split(__file__)[0]
>
> def save_as_python(fd, var, varname, fname_extra=None):
>     if fname_extra is None:
>         fname_extra = ''
>     fname_prefix = varname + fname_extra
>     buf = get_code_for_var( varname, fname_prefix, var)
>     fd.write(buf)
>
> def get_code_for_var( name, fname_prefix, var):
>     if type(var)==numpy.ndarray:
>         fname = fname_prefix + '.ashelf'
>         var.tofile( fname )
>
>         shape = var.shape
>
>         bufs = []
>         bufs.append(
>             '%s = numpy.fromfile(file=os.path.join(ashelf_datadir,"% 
> s"),dtype=numpy.dtype(%s))'%(name,fname,repr(var.dtype.str)))
>         bufs.append(
>             '%s.shape = %s'%(name,repr(shape,)))
>         return '\n'.join(bufs)+'\n'
>
>     if isinstance(var,scipy.sparse.csc_matrix):
>         bufs = []
>         bufs.append(
>             get_code_for_var( '%s_tmp_sparse_data'%name,  
> fname_prefix+'_data', var.data )[:-1])
>         bufs.append(
>             get_code_for_var( '%s_tmp_sparse_indices'%name,  
> fname_prefix+'_indices', var.indices )[:-1])
>         bufs.append(
>             get_code_for_var( '%s_tmp_sparse_indptr'%name,  
> fname_prefix+'_indptr', var.indptr )[:-1])
>         bufs.append(
>             '%s = scipy.sparse.csc_matrix((%s,%s,%s))'%(
>             name,
>             '%s_tmp_sparse_data'%name,
>             '%s_tmp_sparse_indices'%name,
>             '%s_tmp_sparse_indptr'%name,
>             ))
>         bufs.append(
>             '%s.shape = %s'%(name,repr(var.shape)))
>         bufs.append(
>             'del %s_tmp_sparse_data'%name)
>         bufs.append(
>             'del %s_tmp_sparse_indices'%name)
>         bufs.append(
>             'del %s_tmp_sparse_indptr'%name)
>         return '\n'.join(bufs)+'\n'
>
>     if isinstance(var,scipy.sparse.csr_matrix):
>         bufs = []
>         bufs.append(
>             get_code_for_var( '%s_tmp_sparse_data'%name,  
> fname_prefix+'_data', var.data )[:-1])
>         bufs.append(
>             get_code_for_var( '%s_tmp_sparse_colind'%name,  
> fname_prefix+'_colind', var.colind )[:-1])
>         bufs.append(
>             get_code_for_var( '%s_tmp_sparse_indptr'%name,  
> fname_prefix+'_indptr', var.indptr )[:-1])
>         bufs.append(
>             '%s = scipy.sparse.csr_matrix((%s,%s,%s))'%(
>             name,
>             '%s_tmp_sparse_data'%name,
>             '%s_tmp_sparse_colind'%name,
>             '%s_tmp_sparse_indptr'%name,
>             ))
>         bufs.append(
>             '%s.shape = %s'%(name,repr(var.shape)))
>         bufs.append(
>             'del %s_tmp_sparse_data'%name)
>         bufs.append(
>             'del %s_tmp_sparse_colind'%name)
>         bufs.append(
>             'del %s_tmp_sparse_indptr'%name)
>         return '\n'.join(bufs)+'\n'
>     if 1:
>         ra = repr(var)
>         try:
>             cmp = eval(ra)
>         except:
>             raise RuntimeError("eval failed")
>         else:
>             if cmp==var:
>                 return '%s = '%(name,)+ra+'\n'
>             else:
>                 raise RuntimeError("failed conversion")
>
> def test():
>     bigmat = numpy.zeros( (2000, 20000), dtype=numpy.float64 )
>     for i in range(20):
>         for j in range(200,300):
>             bigmat[i,j]=i*j
>     spmat = scipy.sparse.csc_matrix(bigmat)
>     fname = 'test_ashelf_data.py'
>     fd = open(fname,'wb')
>     fd.write( '# Automatically generated by ashelf.py\n')
>     fd.write( 'import numpy\n')
>     fd.write( 'import scipy.sparse\n')
>     fd.write( 'import os\n')
>     fd.write( 'ashelf_datadir = os.path.split(__file__)[0]\n')
>     save_as_python(fd, spmat, 'test_spmat')
>     fd.close()
>
>     locals = {'__file__':fname}
>     execfile(fname,{},locals) # loads test_spmat
>     assert numpy.allclose( spmat.data, locals['test_spmat'].data )
>     assert numpy.allclose( spmat.indices, locals 
> ['test_spmat'].indices )
>     assert numpy.allclose( spmat.indptr, locals['test_spmat'].indptr )
>     print 'sparse matrix saved and loaded OK'
>
> if __name__=='__main__':
>     test()
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user



From cmac at MIT.EDU  Wed Jul 18 17:26:58 2007
From: cmac at MIT.EDU (Christopher W. MacMinn)
Date: Wed, 18 Jul 2007 17:26:58 -0400
Subject: [SciPy-user] SciPy install error (MacPorts)
Message-ID: <5191A80C-4EA0-4D43-AFBA-3D21EBA8A4F8@mit.edu>

Hey folks -

I'm trying to install SciPy via MacPorts, and I'm getting a weird  
error (below).  Anyone have any idea what might be causing this?

I'm using MacPorts 1.5 on an Intel MacBook running OS X 10.4.10.  I'm  
using Python 2.4.4 (although I have others on my system) and I  
installed NumPy 1.0.3 via MacPorts last night without a problem.

Any help or suggestions would be greatly appreciated.

Thanks!

Best
- Chris


--->  Building py-scipy with target build
Error: Target org.macports.build returned: shell command " cd "/opt/ 
local/var/macports/build/ 
_opt_local_var_macports_sources_rsync.macports.org_release_ports_python_ 
py-scipy/work/scipy-0.5.2" && /opt/local/bin/python2.4 setup.py build  
" returned error 1
Command output: _PyImport_ImportModule
_PyInt_Type
_PyModule_GetDict
_PyNumber_Int
_PyObject_GetAttrString
_PySequence_Check
_PySequence_GetItem
_PyString_FromString
_PyString_Type
_PyType_IsSubtype
_PyType_Type
_Py_BuildValue
_Py_InitModule4
__Py_NoneStruct
_PyCObject_FromVoidPtr
_PyDict_DelItemString
_PyDict_GetItemString
_PyDict_New
_PyExc_AttributeError
_PyExc_TypeError
_PyExc_ValueError
_PyObject_Free
_PyObject_Str
_PyObject_Type
_PyString_AsString
_PyString_ConcatAndDel
_Py_FindMethod
__PyObject_New
_MAIN_
error: Command "g95 -shared build/temp.macosx-10.3-i386-2.4/build/ 
src.macosx-10.3-i386-2.4/Lib/fftpack/_fftpackmodule.o build/ 
temp.macosx-10.3-i386-2.4/Lib/fftpack/src/zfft.o build/ 
temp.macosx-10.3-i386-2.4/Lib/fftpack/src/drfft.o build/ 
temp.macosx-10.3-i386-2.4/Lib/fftpack/src/zrfft.o build/ 
temp.macosx-10.3-i386-2.4/Lib/fftpack/src/zfftnd.o build/ 
temp.macosx-10.3-i386-2.4/build/src.macosx-10.3-i386-2.4/ 
fortranobject.o -L/opt/local/lib -Lbuild/temp.macosx-10.3-i386-2.4 - 
ldfftpack -lfftw3 -o build/lib.macosx-10.3-i386-2.4/scipy/fftpack/ 
_fftpack.so" failed with exit status 1

Error: Status 1 encountered during processing.


From haley at ucar.edu  Wed Jul 18 19:02:44 2007
From: haley at ucar.edu (Mary Haley)
Date: Wed, 18 Jul 2007 17:02:44 -0600 (MDT)
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
Message-ID: <Pine.SOC.4.61.0707181655350.5424@windom.ucar.edu>


Hi Ryan,

Somebody told me about this thread, and I thought I'd jump in and
offer another solution.  We have a software package called PyNIO that
is based on Konrad Hinsen's Scientific.IO.netCDF package, only it also
handles GRIB1, GRIB2, HDF SDS 4, and netCDF 4 classic files.

I just finished building a set of PyNIO V1.2.0 binaries under Python
2.4.4 and Python 2.5, supporting both NumPy 1.0.3 and Numeric 24-2. If
you are interested in trying PyNIO, drop me a line.

[I haven't gone public with this release yet, because I'm having
difficulty trying to build NumPy 1.0.3 on a Linux box that has a newer
version of the SSL libraries than what I think NumPy 1.0.3 is
expecting.]

You can see some PyNIO documentation at:

   http://www.pyngl.ucar.edu/Nio.shtml

Cheers,

--Mary

> 
> Bill Dandreta wrote:
> > I don't know if this has any bearing on your problem but Python 2.5 is 
> > hard masked on Gentoo do to problems on 64 bit systems with packages 
> > depending on numeric.
> > 
> 
> Thanks, but, yeah, I am aware of that. (I did my research before 
> unmasking it.  It's driven me nuts that a stable python release has been 
> waiting 10+ months to get into my "cutting edge" distro.)  There weren't 
> any bugs against the packages I use, though I have found that plugins 
> for gedit 2.16 don't like python, but this has been fixed in 2.18.  One 
> of the reasons I originally chose pynetcdf was its support for numpy, 
> not Numeric, so that's not the issue here.
> 
> Ryan
> 
> -- 
> Ryan May
> Graduate Research Assistant
> School of Meteorology
> University of Oklahoma
>


From robert.kern at gmail.com  Wed Jul 18 19:04:58 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 18 Jul 2007 18:04:58 -0500
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
In-Reply-To: <Pine.SOC.4.61.0707181655350.5424@windom.ucar.edu>
References: <Pine.SOC.4.61.0707181655350.5424@windom.ucar.edu>
Message-ID: <469E9C9A.3070707@gmail.com>

Mary Haley wrote:

> [I haven't gone public with this release yet, because I'm having
> difficulty trying to build NumPy 1.0.3 on a Linux box that has a newer
> version of the SSL libraries than what I think NumPy 1.0.3 is
> expecting.]

SSL? numpy has nothing to do with SSL.

-- 
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 haley at ucar.edu  Wed Jul 18 19:17:14 2007
From: haley at ucar.edu (Mary Haley)
Date: Wed, 18 Jul 2007 17:17:14 -0600 (MDT)
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
In-Reply-To: <469E9C9A.3070707@gmail.com>
References: <Pine.SOC.4.61.0707181655350.5424@windom.ucar.edu>
	<469E9C9A.3070707@gmail.com>
Message-ID: <Pine.SOC.4.61.0707181707300.5424@windom.ucar.edu>


Robert, sorry, you're right, this isn't a NumPy issue, but it
manifested itself when I tried to build NumPy. I built Python 2.5 (I
thought successfully), and then when I tried to build NumPy 1.0.3 I
got the error:

    ImportError: No module named _md5

I googled this left and right, and it turns out that when I built
Python 2.5, some module (hashlib?) didn't get built, possibly because
it's expecting some particular version of SSL, and I happen to have a
different version. Python 2.5 is working for me, but I can't build
NumPy because it depends on "_md5" which I believe is dependent
on "hashlib".

I'm still trying to figure thins out. I would be great if I could 
have NumPy bypass this somehow, because I don't think there's
a fix for this on the Python 2.5 end.

--Mary



On Wed, 18 Jul 2007, Robert Kern wrote:

> Mary Haley wrote:
>
>> [I haven't gone public with this release yet, because I'm having
>> difficulty trying to build NumPy 1.0.3 on a Linux box that has a newer
>> version of the SSL libraries than what I think NumPy 1.0.3 is
>> expecting.]
>
> SSL? numpy has nothing to do with SSL.
>
> -- 
> 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 wjdandreta at att.net  Wed Jul 18 21:10:07 2007
From: wjdandreta at att.net (Bill Dandreta)
Date: Wed, 18 Jul 2007 21:10:07 -0400
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
In-Reply-To: <Pine.SOC.4.61.0707181707300.5424@windom.ucar.edu>
References: <Pine.SOC.4.61.0707181655350.5424@windom.ucar.edu>	<469E9C9A.3070707@gmail.com>
	<Pine.SOC.4.61.0707181707300.5424@windom.ucar.edu>
Message-ID: <469EB9EF.1000903@att.net>

A shot in the dark. Maybe python has an optional compile time switch to 
include ssl support?

Bill

Mary Haley wrote:
> Robert, sorry, you're right, this isn't a NumPy issue, but it
> manifested itself when I tried to build NumPy. I built Python 2.5 (I
> thought successfully), and then when I tried to build NumPy 1.0.3 I
> got the error:
>
>     ImportError: No module named _md5
>
> I googled this left and right, and it turns out that when I built
> Python 2.5, some module (hashlib?) didn't get built, possibly because
> it's expecting some particular version of SSL, and I happen to have a
> different version. Python 2.5 is working for me, but I can't build
> NumPy because it depends on "_md5" which I believe is dependent
> on "hashlib".
>
> I'm still trying to figure thins out. I would be great if I could 
> have NumPy bypass this somehow, because I don't think there's
> a fix for this on the Python 2.5 end.
>
> --Mary
>
>
>
> On Wed, 18 Jul 2007, Robert Kern wrote:
>
>   
>> Mary Haley wrote:
>>
>>     
>>> [I haven't gone public with this release yet, because I'm having
>>> difficulty trying to build NumPy 1.0.3 on a Linux box that has a newer
>>> version of the SSL libraries than what I think NumPy 1.0.3 is
>>> expecting.]
>>>       
>> SSL? numpy has nothing to do with SSL.
>>
>> -- 
>> 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
>>
>>     
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>   


-- 
Bill

wjdandreta at att.net

Gentoo Linux X86_64 2.6.20-gentoo-r8

Reclaim Your Inbox with http://www.mozilla.org/products/thunderbird/

All things cometh to he who waiteth as long as he who waiteth worketh like hell while he waiteth.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070718/9a160f26/attachment.html>

From david at ar.media.kyoto-u.ac.jp  Thu Jul 19 01:29:26 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Thu, 19 Jul 2007 14:29:26 +0900
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
In-Reply-To: <Pine.SOC.4.61.0707181707300.5424@windom.ucar.edu>
References: <Pine.SOC.4.61.0707181655350.5424@windom.ucar.edu>	<469E9C9A.3070707@gmail.com>
	<Pine.SOC.4.61.0707181707300.5424@windom.ucar.edu>
Message-ID: <469EF6B6.4030600@ar.media.kyoto-u.ac.jp>

Mary Haley wrote:
> Robert, sorry, you're right, this isn't a NumPy issue, but it
> manifested itself when I tried to build NumPy. I built Python 2.5 (I
> thought successfully), and then when I tried to build NumPy 1.0.3 I
> got the error:
>
>     ImportError: No module named _md5
>   
I got a similar problem when building python on a solaris machine. On 
Linux, I would strongly recommend using the python available in your 
distribution, though. What is the reason you built your own ?

cheers,

David


From robert.kern at gmail.com  Thu Jul 19 01:58:58 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Thu, 19 Jul 2007 00:58:58 -0500
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
In-Reply-To: <Pine.SOC.4.61.0707181707300.5424@windom.ucar.edu>
References: <Pine.SOC.4.61.0707181655350.5424@windom.ucar.edu>	<469E9C9A.3070707@gmail.com>
	<Pine.SOC.4.61.0707181707300.5424@windom.ucar.edu>
Message-ID: <469EFDA2.4000109@gmail.com>

Mary Haley wrote:
> Robert, sorry, you're right, this isn't a NumPy issue, but it
> manifested itself when I tried to build NumPy. I built Python 2.5 (I
> thought successfully), and then when I tried to build NumPy 1.0.3 I
> got the error:
> 
>     ImportError: No module named _md5
> 
> I googled this left and right, and it turns out that when I built
> Python 2.5, some module (hashlib?) didn't get built, possibly because
> it's expecting some particular version of SSL, and I happen to have a
> different version. Python 2.5 is working for me, but I can't build
> NumPy because it depends on "_md5" which I believe is dependent
> on "hashlib".
> 
> I'm still trying to figure thins out. I would be great if I could 
> have NumPy bypass this somehow, because I don't think there's
> a fix for this on the Python 2.5 end.

Ah, okay. We do use the md5 module in part of the build process. See
numpy/core/code_generators/genapi.py . If you just need to get something built,
you can modify that file to use a different hash. You could write your own that
does something extremely simple like keep the strings in a dict mapping to
numbers that you increment.

However, your Python installation is broken. hashlib is a standard module that
should exist in all Python installations.

-- 
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 lbolla at gmail.com  Thu Jul 19 03:34:14 2007
From: lbolla at gmail.com (lorenzo bolla)
Date: Thu, 19 Jul 2007 09:34:14 +0200
Subject: [SciPy-user] scipy.sparse and dot product
Message-ID: <80c99e790707190034j4fd0ff5fkff73b9101668139a@mail.gmail.com>

does this seem weird to you, too?

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

In [34]: A = scipy.sparse.spdiags([[1,2,3]],0,3,3)

In [35]: x = scipy.array([1,2,3])

In [36]: scipy.dot(A,x)
Out[36]:
array([  (0, 0) 1.0
  (1, 1)        2.0
  (2, 2)        3.0,
         (0, 0) 2.0
  (1, 1)        4.0
  (2, 2)        6.0,
         (0, 0) 3.0
  (1, 1)        6.0
  (2, 2)        9.0], dtype=object) # <--- a matrix with repeated entries???
In [37]: scipy.dot(A.todense(),x)
Out[37]: matrix([[ 1.,  4.,  9.]]) # <--- ok

In [38]: A*x
Out[38]: array([ 1.,  4.,  9.]) # <--- ok

In [39]: A.todense()*x
Out[39]: matrix([[ 1.,  4.,  9.]]) <--- ok
 -----------------------------------------------------------------

L.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070719/fdbf63df/attachment.html>

From stefan at sun.ac.za  Thu Jul 19 06:33:46 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Thu, 19 Jul 2007 12:33:46 +0200
Subject: [SciPy-user] scipy.sparse and dot product
In-Reply-To: <80c99e790707190034j4fd0ff5fkff73b9101668139a@mail.gmail.com>
References: <80c99e790707190034j4fd0ff5fkff73b9101668139a@mail.gmail.com>
Message-ID: <20070719103346.GQ7290@mentat.za.net>

On Thu, Jul 19, 2007 at 09:34:14AM +0200, lorenzo bolla wrote:
> does this seem weird to you, too?
>  
> -----------------------------------------------------------------
>  
> In [34]: A = scipy.sparse.spdiags([[1,2,3]],0,3,3)
> 
> In [35]: x = scipy.array([1,2,3])
> 
> In [36]: scipy.dot(A,x)
> Out[36]:
> array([  (0, 0) 1.0
>   (1, 1)        2.0
>   (2, 2)        3.0,
>          (0, 0) 2.0
>   (1, 1)        4.0
>   (2, 2)        6.0,
>          (0, 0) 3.0
>   (1, 1)        6.0
>   (2, 2)        9.0], dtype=object) # <--- a matrix with repeated
> entries???

The numpy 'dot'-function does not know about sparse matrices.  You
might want to try

A.dot(x) or A*x

I don't think the elementwise-product has been implemented yet.

Regards
St?fan


From haley at ucar.edu  Thu Jul 19 08:52:27 2007
From: haley at ucar.edu (Mary Haley)
Date: Thu, 19 Jul 2007 06:52:27 -0600
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
In-Reply-To: <469EF6B6.4030600@ar.media.kyoto-u.ac.jp>
References: <Pine.SOC.4.61.0707181655350.5424@windom.ucar.edu>	<469E9C9A.3070707@gmail.com>
	<Pine.SOC.4.61.0707181707300.5424@windom.ucar.edu>
	<469EF6B6.4030600@ar.media.kyoto-u.ac.jp>
Message-ID: <d015a4c0f095d119a75a8e95f22187b6@ucar.edu>


On Jul 18, 2007, at 11:29 PM, David Cournapeau wrote:

> Mary Haley wrote:
>> Robert, sorry, you're right, this isn't a NumPy issue, but it
>> manifested itself when I tried to build NumPy. I built Python 2.5 (I
>> thought successfully), and then when I tried to build NumPy 1.0.3 I
>> got the error:
>>
>>     ImportError: No module named _md5
>>
> I got a similar problem when building python on a solaris machine. On
> Linux, I would strongly recommend using the python available in your
> distribution, though. What is the reason you built your own ?
>

I'm building software for other users, and I like to test it against 
the latest
versions of Python and NumPy, if possible. Also, we provide these 
precompiled
binaries to users, and like to have it available for a variety of 
software versions.

The other thing is that our systems tend to be a few versions behind in 
python
and it's usually just easier to "roll our own".

I ended up finding another Linux system that had an older version of 
SSL installed,
and I was able to get Python 2.5 fully built on it.

Thanks for everybody's suggestions.

--Mary



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



From fie.pye at atlas.cz  Thu Jul 19 18:33:13 2007
From: fie.pye at atlas.cz (fie.pye at atlas.cz)
Date: Thu, 19 Jul 2007 22:33:13 GMT
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
Message-ID: <f849b8daec4746f685fdc7c3b0d2980e@45f24f68e5c24903876e8c374cb5b660>

>---------------------------------------------------------
>Od: Scott Ransom
>P?ijato: 18.7.2007 4:15:24
>P?edm?t: Re: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
>
>> The problem is the way you build the shared library is wrong: you cannot 
>
>> just use g77 -shared -o soname *.o, because LAPACK object files are in 
>
>> several directories.
>
>> 
>
>> One correct way is to use the static library .a, uncompress it in a 
>
>> directory, and build the shared library from that. There is a problem 
>
>> when several object files have the same name (because you cannot have 
>
>> several files with the same name in a directory), though, so extra care 
>
>> must be taken.
>
>
>
>You might be able to do something like this.  It has worked for me
>
>in the past (with GNU ld):
>
>
>
>ld -shared -o libfoo.so --whole-archive libfoo.a
>
>
>
>Scott
>
>
>
>-- 
>
>Scott M. Ransom            Address:  NRAO
>
>Phone:  (434) 296-0320               520 Edgemont Rd.
>
>email:  sransom at nrao.edu             Charlottesville, VA 22903 USA
>
>GPG Fingerprint: 06A9 9553 78BE 16DB 407B  FFCA 9BFA B6FF FFD3 2989
>
>_______________________________________________
>
>SciPy-user mailing list
>
>SciPy-user at scipy.org
>
>http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>
>
Hello.

Thank you all for responses. Finally I installed BLAS, LAPACK, python2.5.1, numpy and scipy. Scot was right. I had to change the way I built the shared libraries. The command

ld -shared -o libfoo.so --whole-archive libfoo.a

helped but installation stopped on other error. 

After that I had to change numpy and scipy packages. The *.tar.gz numpy and scipy packages I downloaded via www.scipy.org did not work. I followed the recommendation from  
http://permalink.gmane.org/gmane.comp.python.scientific.user/12133
It works.

Best regards.
Fie Pye

From fie.pye at atlas.cz  Thu Jul 19 18:40:42 2007
From: fie.pye at atlas.cz (fie.pye at atlas.cz)
Date: Thu, 19 Jul 2007 22:40:42 GMT
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
Message-ID: <2e7ee8776fc740ca8cd618d8d0d1d2fc@2271d759aab94160bbd6a2bf37e42b2e>

>---------------------------------------------------------
>Od: David Cournapeau
>P?ijato: 18.7.2007 3:54:48
>P?edm?t: Re: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
>
>fie.pye at atlas.cz wrote:
>
>> Hi David, hi Vincent,
>
>>
>
>> thank you for your response. 
>
>> On my computer I have three gcc compilers: gcc4.1, gcc4.2 and gcc3.4.6. Since you mentioned possible problem with gfortran I decided to compile BLAS, LAPACK, python2.5 and numpy again with gcc3.4.6 and f77.
>
>To be precise, what is problematic is mixing fortran compilers (for 
>
>example using gfortran for BLAS and g77 for LAPACK). gfortran by itself 
>
>is not that problematic.
>
>>  
>
>> Traceback (most recent call last):
>
>>   File "<stdin>", line 1, in <module>
>
>>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/__init__.py", line 43, in <module>
>
>>     import linalg
>
>>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in <module>
>
>>     from linalg import *
>
>>   File "/usr/local/python/python2.5.1/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in <module>
>
>>     from numpy.linalg import lapack_lite
>
>> ImportError: /usr/local/lib64/lapack/liblapack.so: undefined symbol: slamch_
>
>The problem is the way you build the shared library is wrong: you cannot 
>
>just use g77 -shared -o soname *.o, because LAPACK object files are in 
>
>several directories.
>
>
>
>One correct way is to use the static library .a, uncompress it in a 
>
>directory, and build the shared library from that. There is a problem 
>
>when several object files have the same name (because you cannot have 
>
>several files with the same name in a directory), though, so extra care 
>
>must be taken.
>
>>
>
>> I included all installation scripts and log files to enclosed file. 
>
>>
>
>> I also began to install from source rpm but I am not familiar with that installation therefore I stopped after building. What does it mean refblas. 
>
>I cannot use blas and lapack from officiel repository because they do 
>
>not work very well, and I use a different name to avoid name clashes. 
>
>You should build the following srpms in this order:
>
>
>
>- refblas
>
>- lapack3
>
>- python-numpy
>
>- python-scipy
>
>
>
>Building them is not really different than any other source rpms. Since 
>
>you are using a distribution I've never tested, there is a good chance 
>
>it does not work out of the box, but I would rather help you solving 
>
>those problems so that everybody can benefit from them rather than 
>
>explaining how to build blas and lapack properly (which is really error 
>
>prone and non trivial).
>
>
>
>David
>
>_______________________________________________
>
>SciPy-user mailing list
>
>SciPy-user at scipy.org
>
>http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>
>
Hello David.

Thank you for responses. I managed to install numpy and scipy with BLAS and LAPACK libraries. Now I am going to install ATLAS and again python2.5.1, numpy and scipy. I will follow two ways. The first one is to build it from *.tar.gz sources. The second one is to build it from *.src.rpm as you recommended.  I think that the closer one *.src.rpm from http://software.opensuse.org/download/home:/ashigabou to CentOS 5.0 is Fedoea 6
http://software.opensuse.org/download/home:/ashigabou/Fedora_Extras_6/src/

The reason why I started with the *.tar.gz is that I have the full control over the version of libraries and places where the libraries are located. I can easily remove them when installation fails. In case of *.src.rmp I will have to study the installation and removing procedure first. I am going to start with rpmbuild. If you have some suggestions let me know.

If everything goes well I will install BLAS, LAPACK, ATLAS, python2.5.1, numpy and scipy also on HP workstation and server with AMD Opteron 2220 and 2218 with SUSE Linux Enterprise 10. 

Best regards.
Fie Pye


From david at ar.media.kyoto-u.ac.jp  Thu Jul 19 22:27:04 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Fri, 20 Jul 2007 11:27:04 +0900
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
In-Reply-To: <2e7ee8776fc740ca8cd618d8d0d1d2fc@2271d759aab94160bbd6a2bf37e42b2e>
References: <2e7ee8776fc740ca8cd618d8d0d1d2fc@2271d759aab94160bbd6a2bf37e42b2e>
Message-ID: <46A01D78.7010009@ar.media.kyoto-u.ac.jp>

fie.pye at atlas.cz wrote:
> Hello David
>
> Thank you for responses. I managed to install numpy and scipy with BLAS and LAPACK libraries. Now I am going to install ATLAS and again python2.5.1, numpy and scipy. I will follow two ways. The first one is to build it from *.tar.gz sources. The second one is to build it from *.src.rpm as you recommended.  I think that the closer one *.src.rpm from http://software.opensuse.org/download/home:/ashigabou to CentOS 5.0 is Fedoea 6
> http://software.opensuse.org/download/home:/ashigabou/Fedora_Extras_6/src/
>
> The reason why I started with the *.tar.gz is that I have the full control over the version of libraries and places where the libraries are located. I can easily remove them when installation fails. In case of *.src.rmp I will have to study the installation and removing procedure first. I am going to start with rpmbuild. If you have some suggestions let me know.
>
> If everything goes well I will install BLAS, LAPACK, ATLAS, python2.5.1, numpy and scipy also on HP workstation and server with AMD Opteron 2220 and 2218 with SUSE Linux Enterprise 10. 
>   
I can try to generate rpm for this distribution if you want (CentOS is 
not supported by the Suse build system, but Suse Linux enterprise is).

If you manage to generate rpms for Centos, I would appreciate to get the 
modifications you had to do (if any, but I think they do not work out of 
the box for Centos).

cheers,

David


From david at ar.media.kyoto-u.ac.jp  Thu Jul 19 22:53:39 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Fri, 20 Jul 2007 11:53:39 +0900
Subject: [SciPy-user] BLAS, LAPACK, ATLAS libraries.
In-Reply-To: <46A01D78.7010009@ar.media.kyoto-u.ac.jp>
References: <2e7ee8776fc740ca8cd618d8d0d1d2fc@2271d759aab94160bbd6a2bf37e42b2e>
	<46A01D78.7010009@ar.media.kyoto-u.ac.jp>
Message-ID: <46A023B3.9000809@ar.media.kyoto-u.ac.jp>

David Cournapeau wrote:
> fie.pye at atlas.cz wrote:
>> Hello David
>>
>> Thank you for responses. I managed to install numpy and scipy with BLAS and LAPACK libraries. Now I am going to install ATLAS and again python2.5.1, numpy and scipy. I will follow two ways. The first one is to build it from *.tar.gz sources. The second one is to build it from *.src.rpm as you recommended.  I think that the closer one *.src.rpm from http://software.opensuse.org/download/home:/ashigabou to CentOS 5.0 is Fedoea 6
>> http://software.opensuse.org/download/home:/ashigabou/Fedora_Extras_6/src/
>>
>> The reason why I started with the *.tar.gz is that I have the full control over the version of libraries and places where the libraries are located. I can easily remove them when installation fails. In case of *.src.rmp I will have to study the installation and removing procedure first. I am going to start with rpmbuild. If you have some suggestions let me know.
>>
>> If everything goes well I will install BLAS, LAPACK, ATLAS, python2.5.1, numpy and scipy also on HP workstation and server with AMD Opteron 2220 and 2218 with SUSE Linux Enterprise 10. 
>>   
> I can try to generate rpm for this distribution if you want (CentOS is 
> not supported by the Suse build system, but Suse Linux enterprise is).
Ok, it looks like SUSE Linux Enterprise 10 is similar enough to opensuse 
such as I won't have to do modifications, so the packages should be 
built soon and made available in 
http://software.opensuse.org/download/home:/ashigabou (a couple of hours 
maximum depending on the availability of the compiler farm).

David


From nwagner at iam.uni-stuttgart.de  Fri Jul 20 07:27:37 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Fri, 20 Jul 2007 13:27:37 +0200
Subject: [SciPy-user] How to build a sparse block arrowhead matrix in scipy
Message-ID: <46A09C29.5070703@iam.uni-stuttgart.de>

Hi all,

what is the best way to build a sparse block arrowhead matrix A in scipy ?

The diagonal blocks are given by the matrices B1,B2,...,B8
The entries in the last block row/column are given by the matrix C.
The example is taken from [1].

A short python script would be appreciated.

Nils
 
[1] A. Sameh, J. Lermit and K. Noh,
On the intermediate eigenvalues of symmetric sparse matrices, BIT Vol.
15 (1975) pp. 185-191


From cimrman3 at ntc.zcu.cz  Fri Jul 20 08:27:31 2007
From: cimrman3 at ntc.zcu.cz (Robert Cimrman)
Date: Fri, 20 Jul 2007 14:27:31 +0200
Subject: [SciPy-user] ANN: SFE-00.26.01 20.07.2007 release
Message-ID: <46A0AA33.7080603@ntc.zcu.cz>

Version 00.26.01 is released 20.07.2007, featuring new testing
framework, new terms (linear spring, volume), improved handling of
periodic boundary conditions and many other gradual updates and bug
fixes, see http://ui505p06-mbs.ntc.zcu.cz/sfe

cheers,
r.


From david at ar.media.kyoto-u.ac.jp  Sat Jul 21 03:56:55 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Sat, 21 Jul 2007 16:56:55 +0900
Subject: [SciPy-user] [scikits] svm and pyem now available in scikits.learn
Message-ID: <46A1BC47.2010104@ar.media.kyoto-u.ac.jp>

Hi,

    In short: for people who need to use svm (SVM) and pyem (Gaussian 
Mixture Models) from the scipy.sandbox, they are now available and 
working in scikits.learn package.

    - Getting the code : svn co http://svn.scipy.org/svn/scikits/trunk/learn
    - Importing the toolboxes : they both reside in the 
scikits.learn.machine namespace, that is "from scipy.sandbox import svm" 
becomes "from scikits.learn.machine import svm" and so on.

Anything which does not work as before is a bug, and should be filled as 
such on the scikits trac system (http://projects.scipy.org/scipy/scikits/).

    For the curious, the learn namespace will soon contain some code to 
load/pre process/manipulate datasets and some basic learner based on the 
above algoritms,

    cheers,

    David



From josh8912 at yahoo.com  Sun Jul 22 20:16:13 2007
From: josh8912 at yahoo.com (JJ)
Date: Sun, 22 Jul 2007 17:16:13 -0700 (PDT)
Subject: [SciPy-user] missing libacml.so when using cgi script
Message-ID: <186675.64366.qm@web54011.mail.re2.yahoo.com>

Hello all:
I am developing a website, initially on my local
machine using the apache httpd server.  My linux box
has scipy/numpy installed, and they work fine from a
terminal.  However, I need to use the programs in a
cgi script and am having problems getting them to
import.  In my bashrc file (for a terminal) I have an
export LD_LIBRARY_PATH statement that leads to my
libacml.so file.  However, when I try to import scipy
from a cgi script I get the following error:

File
'/usr/lib64/python2.4/site-packages/scipy/optimize/lbfgsb.py",
line 30, in ?   import _lbfgsb   ImportError:
libacml.so: cannot open shared object file: No such
file or directory

Numpy does import just fine.  I tried using the
following in the cgi script but it did not correct the
problem:

if os.environ.has_key('LD_LIBRARY_PATH')==False:
  os.environ['LD_LIBRARY_PATH'] = path to lib
else:
  os.environ['LD_LIBRARY_PATH'] =
os.environ['LD_LIBRARY_PATH'] + path to lib

I also tried adding similar statments for LD_RUN_PATH
and LAPACK to no avail.  Also, file permissions for
the folder with libacml.so are set to allow others to
read.  Does anyone know how I might get scipy working
from a cgi script?

John






       
____________________________________________________________________________________
Take the Internet to Go: Yahoo!Go puts the Internet in your pocket: mail, news, photos & more. 
http://mobile.yahoo.com/go?refer=1GNXIC


From robert.kern at gmail.com  Sun Jul 22 20:33:11 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Sun, 22 Jul 2007 19:33:11 -0500
Subject: [SciPy-user] missing libacml.so when using cgi script
In-Reply-To: <186675.64366.qm@web54011.mail.re2.yahoo.com>
References: <186675.64366.qm@web54011.mail.re2.yahoo.com>
Message-ID: <46A3F747.9030301@gmail.com>

JJ wrote:
> Hello all:
> I am developing a website, initially on my local
> machine using the apache httpd server.  My linux box
> has scipy/numpy installed, and they work fine from a
> terminal.  However, I need to use the programs in a
> cgi script and am having problems getting them to
> import.  In my bashrc file (for a terminal) I have an
> export LD_LIBRARY_PATH statement that leads to my
> libacml.so file.  However, when I try to import scipy
> from a cgi script I get the following error:
> 
> File
> '/usr/lib64/python2.4/site-packages/scipy/optimize/lbfgsb.py",
> line 30, in ?   import _lbfgsb   ImportError:
> libacml.so: cannot open shared object file: No such
> file or directory
> 
> Numpy does import just fine.  I tried using the
> following in the cgi script but it did not correct the
> problem:
> 
> if os.environ.has_key('LD_LIBRARY_PATH')==False:
>   os.environ['LD_LIBRARY_PATH'] = path to lib
> else:
>   os.environ['LD_LIBRARY_PATH'] =
> os.environ['LD_LIBRARY_PATH'] + path to lib

This won't work. The LD_LIBRARY_PATH environment variable needs to be set before
the process starts, not just before the library is loaded. Read the Apache
documentation for how to set up the environment for your CGI scripts.

-- 
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 lists at benair.net  Mon Jul 23 02:51:41 2007
From: lists at benair.net (BK)
Date: Mon, 23 Jul 2007 08:51:41 +0200
Subject: [SciPy-user] (no subject)
Message-ID: <1185173501.4656.1.camel@iagpc71.iag.uni-stuttgart.de>

Hi everybody,

since I discovered scipy I am constantly using python for all my little
data handling jobs. However, there is one problem bugging me:

my data often comes in Tecplot ascii-files that contain several sets of
(float-)data. The principle file layout ist something like

TITLE = "some name"
VARIABLES = "x", "y", "z", "abc"
ZONE T="zone1", I=    65, J=    97
[float data]
ZONE T="zone2", I=    65, J=    49
[float data]
...

I open the file using infile=open("filename",'r') to read the header
information and extract the array sizes (I,J) for the first ZONE. Then I
use io.array_import.read_array(infile,lines=(0,(1,I*J))) to read the
data for the first ZONE. Then I would like to read the header for the
second zone "zone2", extract (I,J) the same way as before and read the
data for the second zone and so on. However, I get the error 
'ValueError: I/O operation on closed file'
which seems to indicate that io.array_import.read_array closed the file
object after extracting the first data set. Looking into
'array_import.py' confirmes, that the class 'ascii_stream' closes the
file-object upon destruction of the class object.

Just for testing I changed 'array_import.py' such that it keeps the file
object open if it was not opened by array_import itself. But then I had
to learn that after importing an array the cursor is not positioned
where I expected at the end of the data set, but somewhere arbitrary in
between.

Is there any way to change this behaviour? If I re-open the file I have
to search
the starting position of the next zone before reading the next data.
While this is doable it's nevertheless annoying. Can anyone tell me how
to keep the file object open after using read_array and continuing read
operations after the last data set?

Thank you!
bene



From robert.kern at gmail.com  Mon Jul 23 02:59:09 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Mon, 23 Jul 2007 01:59:09 -0500
Subject: [SciPy-user] (no subject)
In-Reply-To: <1185173501.4656.1.camel@iagpc71.iag.uni-stuttgart.de>
References: <1185173501.4656.1.camel@iagpc71.iag.uni-stuttgart.de>
Message-ID: <46A451BD.6040300@gmail.com>

BK wrote:

> Is there any way to change this behaviour?

I just checked in a change to allow this. If you pass it a file object instead
of a filename, it will not close the file object.

-- 
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 lists at benair.net  Mon Jul 23 04:36:01 2007
From: lists at benair.net (BK)
Date: Mon, 23 Jul 2007 10:36:01 +0200
Subject: [SciPy-user] (no subject)
In-Reply-To: <46A451BD.6040300@gmail.com>
References: <1185173501.4656.1.camel@iagpc71.iag.uni-stuttgart.de>
	<46A451BD.6040300@gmail.com>
Message-ID: <1185179762.5436.10.camel@iagpc71.iag.uni-stuttgart.de>

Thank you, that was a fast reply!

The file object is indeed kept open with your fix. However, the read
cursor is still at eof after calling array_import and not at the end of
the data array just read. I have to manually seek this position.
Considering I specify the number of lines to be read by array_import, is
it possible to leave the read position at the end of these lines?

Regards,
bene

Am Montag, den 23.07.2007, 01:59 -0500 schrieb Robert Kern:
> BK wrote:
> 
> > Is there any way to change this behaviour?
> 
> I just checked in a change to allow this. If you pass it a file object instead
> of a filename, it will not close the file object.
> 



From chanley at stsci.edu  Mon Jul 23 12:42:53 2007
From: chanley at stsci.edu (Christopher Hanley)
Date: Mon, 23 Jul 2007 12:42:53 -0400
Subject: [SciPy-user] SciPy 2007 Conference BOF
Message-ID: <46A4DA8D.6060109@stsci.edu>

Greetings,

I have a question for those planning on attending the SciPy 2007 
conference in a few weeks.  I have been approached regarding the 
possibility about having a BOF meeting to discuss PyFITS.  I was 
wondering if there would be an interest in this, or more generally, 
about astronomy software development.  We at the Space Telescope Science 
Institute could talk about our future plans.  Please let me know if you 
are interested in having a BOF meeting.  If there is enough interest I 
will arrange a session for Thursday evening.

Cheers,
Chris Hanley

-- 
Christopher Hanley
Systems Software Engineer
Space Telescope Science Institute
3700 San Martin Drive
Baltimore MD, 21218
(410) 338-4338



From shupe at ipac.caltech.edu  Mon Jul 23 14:21:26 2007
From: shupe at ipac.caltech.edu (David Shupe)
Date: Mon, 23 Jul 2007 18:21:26 +0000 (UTC)
Subject: [SciPy-user] SciPy 2007 Conference BOF
References: <46A4DA8D.6060109@stsci.edu>
Message-ID: <loom.20070723T200522-273@post.gmane.org>


Christopher Hanley <chanley <at> stsci.edu> writes:

> I have a question for those planning on attending the SciPy 2007 
> conference in a few weeks.  I have been approached regarding the 
> possibility about having a BOF meeting to discuss PyFITS.  I was 
> wondering if there would be an interest in this, or more generally, 
> about astronomy software development.  We at the Space Telescope Science 
> Institute could talk about our future plans.  Please let me know if you 
> are interested in having a BOF meeting.  If there is enough interest I 
> will arrange a session for Thursday evening.

I'm very interested in this BOF session, especially if the topic is broadened
beyond PyFITS to include development of utilities for astronomy.  

The Herschel Observatory is basing its data processing software on Jython.  It
might be useful to discuss these plans at the BOF.

Regards,
David Shupe 
NASA Herschel Science Center







From chanley at stsci.edu  Mon Jul 23 21:24:11 2007
From: chanley at stsci.edu (Christopher Hanley)
Date: Mon, 23 Jul 2007 21:24:11 -0400
Subject: [SciPy-user] Astronomy / PyFITS BOF at SciPy
Message-ID: <46A554BB.6090908@stsci.edu>

Hi,

It looks like we have more than enough interest for an astronomy BOF 
Thursday evening after dinner.  I look forward to seeing everyone.

Cheers,
Chris


From ryanlists at gmail.com  Mon Jul 23 23:27:47 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Mon, 23 Jul 2007 22:27:47 -0500
Subject: [SciPy-user] filtering without phase shift
Message-ID: <c5b438120707232027w32c8a564v346f9292501616e8@mail.gmail.com>

I have a vector of data from an impact test.  The data has some
"noise" in it that is really ringing in the mass used in the test.  I
have spent a fair amount of effort getting rid of the ringing or
trying to design the test so that the ringing is at as high a
frequency as possible.  I now must resort to filtering the data.

I can get fairly decent results with a 10 kHz filter (see attached).
But I can get really nice results with a 1 kHz filter and then
shifting the vector back in time - somewhat arbitrarily.  This feels
slightly unscientific.  I googled for zero phase filters, but don't
really understand what I was reading.  Is there a way to get magtitude
attenuation without phase lag that is easy to implement and accepted
as good practice among signal processing experts?

Thanks,

Ryan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: filtering.png
Type: image/png
Size: 64927 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070723/bd5e0e20/attachment.png>

From j.merritt at pgrad.unimelb.edu.au  Mon Jul 23 23:43:10 2007
From: j.merritt at pgrad.unimelb.edu.au (Jonathan Merritt)
Date: Tue, 24 Jul 2007 13:43:10 +1000
Subject: [SciPy-user] filtering without phase shift
In-Reply-To: <c5b438120707232027w32c8a564v346f9292501616e8@mail.gmail.com>
References: <c5b438120707232027w32c8a564v346f9292501616e8@mail.gmail.com>
Message-ID: <46A5754E.6020600@pgrad.unimelb.edu.au>

Hi Ryan,

Ryan Krauss wrote:
> Is there a way to get magtitude
> attenuation without phase lag that is easy to implement and accepted
> as good practice among signal processing experts?

This is a very common problem in Biomechanics, where we often want to 
take the second derivative of noisy signals.

Matlab has a filtfilt() function for applying a zero phase delay filter, 
and I have seen a SciPy implementation here:
    http://www.scipy.org/Cookbook/FiltFilt

I can't comment on whether this implementation is "good".  I'd like to 
hear from any signal processing experts too. :-)

Jonathan Merritt.



From bryan at cole.uklinux.net  Tue Jul 24 03:14:25 2007
From: bryan at cole.uklinux.net (Bryan Cole)
Date: Tue, 24 Jul 2007 08:14:25 +0100
Subject: [SciPy-user] filtering without phase shift
In-Reply-To: <c5b438120707232027w32c8a564v346f9292501616e8@mail.gmail.com>
References: <c5b438120707232027w32c8a564v346f9292501616e8@mail.gmail.com>
Message-ID: <1185261264.5678.9.camel@pc1.cole.uklinux.net>

The easiest way to do non-phase-shifted filtering is to use FFTs i.e.

FFT -> Low Pass -> iFFT

The only reason to move to time-domain filters is if you must do
filtering in real-time where the complete dataset is (yet) unknown and
you must obey the laws of causality.

Using a frequency domain filter, you can have whatever cutoff you like.

HTH

Bryan

On Mon, 2007-07-23 at 22:27 -0500, Ryan Krauss wrote:
> I have a vector of data from an impact test.  The data has some
> "noise" in it that is really ringing in the mass used in the test.  I
> have spent a fair amount of effort getting rid of the ringing or
> trying to design the test so that the ringing is at as high a
> frequency as possible.  I now must resort to filtering the data.
> 
> I can get fairly decent results with a 10 kHz filter (see attached).
> But I can get really nice results with a 1 kHz filter and then
> shifting the vector back in time - somewhat arbitrarily.  This feels
> slightly unscientific.  I googled for zero phase filters, but don't
> really understand what I was reading.  Is there a way to get magtitude
> attenuation without phase lag that is easy to implement and accepted
> as good practice among signal processing experts?
> 
> Thanks,
> 
> Ryan
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user




From s.mientki at ru.nl  Tue Jul 24 04:24:44 2007
From: s.mientki at ru.nl (Stef Mientki)
Date: Tue, 24 Jul 2007 10:24:44 +0200
Subject: [SciPy-user] filtering without phase shift
In-Reply-To: <1185261264.5678.9.camel@pc1.cole.uklinux.net>
References: <c5b438120707232027w32c8a564v346f9292501616e8@mail.gmail.com>
	<1185261264.5678.9.camel@pc1.cole.uklinux.net>
Message-ID: <46A5B74C.2060304@ru.nl>



Bryan Cole wrote:
> The easiest way to do non-phase-shifted filtering is to use FFTs i.e.
>
> FFT -> Low Pass -> iFFT
>
> The only reason to move to time-domain filters is if you must do
> filtering in real-time where the complete dataset is (yet) unknown and
> you must obey the laws of causality.
>
> Using a frequency domain filter, you can have whatever cutoff you like.
>   
I'm a bit rusty, but isn't there a real danger by filtering in the 
frequency domain,
creating a rectangle cutoff in the frequency domain,
leads to an infinite ringing signal in the time-domain.

The normal way of getting zero phaseshift,
is to apply the filter twice, once normally and once in reversed order.

Stef Mientki



From david at ar.media.kyoto-u.ac.jp  Tue Jul 24 04:15:21 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Tue, 24 Jul 2007 17:15:21 +0900
Subject: [SciPy-user] filtering without phase shift
In-Reply-To: <46A5B74C.2060304@ru.nl>
References: <c5b438120707232027w32c8a564v346f9292501616e8@mail.gmail.com>	<1185261264.5678.9.camel@pc1.cole.uklinux.net>
	<46A5B74C.2060304@ru.nl>
Message-ID: <46A5B519.5090505@ar.media.kyoto-u.ac.jp>

Stef Mientki wrote:
> Bryan Cole wrote:
>   
>> The easiest way to do non-phase-shifted filtering is to use FFTs i.e.
>>
>> FFT -> Low Pass -> iFFT
>>
>> The only reason to move to time-domain filters is if you must do
>> filtering in real-time where the complete dataset is (yet) unknown and
>> you must obey the laws of causality.
>>
>> Using a frequency domain filter, you can have whatever cutoff you like.
>>   
>>     
> I'm a bit rusty, but isn't there a real danger by filtering in the 
> frequency domain,
> creating a rectangle cutoff in the frequency domain,
> leads to an infinite ringing signal in the time-domain.
>   
I agree. It depends of course of the type of signals you have to filter 
and what you want to get from them, but I would not advise FFT filtering 
as THE method for filtering.
> The normal way of getting zero phaseshift,
> is to apply the filter twice, once normally and once in reversed order.
>   
That's what I learnt too, but I am a bit rusty on those too :)

David


From ded.espaze at laposte.net  Tue Jul 24 04:51:06 2007
From: ded.espaze at laposte.net (Dede)
Date: Tue, 24 Jul 2007 10:51:06 +0200
Subject: [SciPy-user] vectorizing methods?
In-Reply-To: <20070715093244.GA28109@clipper.ens.fr>
References: <ad0d4fcf0707141631r1d96702ej3586242912d47347@mail.gmail.com>
	<20070715093244.GA28109@clipper.ens.fr>
Message-ID: <20070724105106.1ebcddc0@localhost>

Sorry to reply a little late, I have else that solution:


def vectorize(bar_function):
    def modified_bar_function(self, x):
        return bar_function(self, x) + self.a
    modified_bar_function.__doc__ = "Be carrefull, I am not who you think"
    return modified_bar_function


class VectorizeBar(type):
    def __init__(cls, name, bases, dct):
        setattr(cls, "bar", vectorize(dct["bar"]))


class Bar:
    __metaclass__ = VectorizeBar
    
    def bar(self, x):
        return x
    
    a = 2.

It brings some magic but it's rather funny in some cases.

Cheers,

Dede


On Sun, 15 Jul 2007 11:32:44 +0200
Gael Varoquaux <gael.varoquaux at normalesup.org> wrote:

> On Sat, Jul 14, 2007 at 04:31:59PM -0700, Greg Novak wrote:
> > Is there a preferred way to vectorize methods, rather than
> > functions? The obvious thing doesn't work:
> 
> > class foo:
> >     def bar(self, x): pass
> >     bar = vectorize(bar)
> 
> This cannot work, as you are applying "vectorize" before the class is
> built, so the function is still unbound. The binding process, which
> happens during the building of the class, will not like what you have
> done to the function to vectorize it, mainly the fact that it is no
> longer a plain function.
> 
> 
> > class foo:
> >     __vectorMethods = ('bar', )
> 
> >     def __init__(self, n):
> >         for name in self.__vectorMethods:
> >             setattr(self, name, vectorize(getattr(self, name)))
> 
> >     def bar(self, x): pass
> 
> There you are vectorizing after the construction of the instance, so
> the method is already bound.
> 
> > which works fine but it doesn't seem like it should be necessary to
> > do it in the initialization of every instance.
> 
> Well, if your methods don't really refer to the object (they don't use
> self at all in the body of the function), you could do something like:
> 
> class Foo(object):
>     @staticmethod
>     @vectorize
>     def bar(x):
>         return x
> 
> 
> In the static method there is no reference to the object. The problem
> is that when the class is binding the method (making a call in which
> the first argument is already assigned to the instance) it is really
> modifying it. If you could apply vectorize to only part of the
> argument, and get a curry of the original function you could than
> transform bar(self, *args) in baz(self, *args), where *args can be
> vectors, and this might work.
> 
> Is can see two ways out of this:
> 
>    1) Do some heavy magic with a metaclass and modify the binding
> process (I am talking about things I don't know, so it may not be
>       feasible).
> 
>  2.a) Use a static method that you vectorize, and pass it the
> instance as the first argument using a convention bound method:
> 
> 	class Foo(object):
> 	    @staticmethod
> 	    @vectorize
> 	    def bar(self, x):
> 		return self.a + x
> 		
> 	    def baz(self, x):
> 		return self.bar(self, x)
> 
> 	    a = 1
> 
>  2.b) IMHO it would be even cleaner to only pass interesting vectors
> to the vectorize staticmethod, as it has been vectorized relatively to
>       "self", but self cannot be a vector:
> 
> 	class Foo(object):
> 	    @staticmethod
> 	    @vectorize
> 	    def bar(a, x):
> 		return a + x
> 		
> 	    def baz(self, x):
> 		return self.bar(self.a, x)
> 
> 	    a = 1
> 
>       Basically this is similar to using a standard function instead
> of a method, but lets face it, the notions of bound method and
>       vectorized function seem mutually incompatible as they both
> want to modify a function.
> 
> It might be possible to implement a decorator similar to staticmethod
> that does the proces presented in 2.a) but hiddes it from the user,
> but to do this, you would need to understand how staticmethod works,
> and I suspect there is some cooperation of the class building
> mechanisme (so we are back to 1), I think).
> 
> I hope 2.a) or 2.b) suit you, elsewhere maybe some one has a better
> idea ?
> 
> Cheers,
> 
> Ga?l
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user


From frank.horowitz at csiro.au  Tue Jul 24 06:59:01 2007
From: frank.horowitz at csiro.au (Frank Horowitz)
Date: Tue, 24 Jul 2007 18:59:01 +0800
Subject: [SciPy-user] Matplotlib transformations?
Message-ID: <924B9651-3C11-4658-A976-E0C2B48DD862@csiro.au>

Hi All,

I have a visualization problem that I need to solve,  I think that  
the matplotlib transforms module might do the trick, but want the  
opinion of any matplotlib experts out there whether I'm going down a  
blind alley before investing too much time into hacking...

I need to convert a "few time step" but multiple value visualization  
from an animation to a single figure for publication. (It's to do  
with flow along a network, so spatial layout is quite important.)

The best guess I currently have as to an acceptable visual display is  
something like a "skewed" bar graph for each component of the time  
series. In other words, take a normal rectangular bar graph/time- 
series and affine transform the bounding box so that it is deformed  
to a parallelogram.

Is a strategy of building the bar-graph, then applying an affine  
transform, then plotting a viable strategy in the context of MPL 0.87.7?

If not, I'm quite open to suggestions!

TIA for any sage advice you might have!

Cheers,
	Frank Horowitz


From fredmfp at gmail.com  Tue Jul 24 11:09:50 2007
From: fredmfp at gmail.com (fred)
Date: Tue, 24 Jul 2007 17:09:50 +0200
Subject: [SciPy-user] mgrid/ogrid...
Message-ID: <46A6163E.7080501@gmail.com>

Hi all,

What's difference between ogrid() and mgrid ?

help() or ogrid?/mgrid? in ipython
gives the same answer.

Cheers,

-- 
http://scipy.org/FredericPetit



From franckm at aims.ac.za  Tue Jul 24 11:45:36 2007
From: franckm at aims.ac.za (Kalala Mutombo Franck)
Date: Tue, 24 Jul 2007 17:45:36 +0200 (SAST)
Subject: [SciPy-user] download and install python with all modules
In-Reply-To: <46A6163E.7080501@gmail.com>
References: <46A6163E.7080501@gmail.com>
Message-ID: <59416.41.194.14.11.1185291936.squirrel@webmail.aims.ac.za>

Hi

I want to download python and install it on my computer, how can I
download python and all the needed modules? for examples
scipy,numeric,etc.
-- 
Franck Kalala Mutombo
African Institute for Mathematical Sciences
www.aims.ac.za

On Tue, July 24, 2007 5:09 pm, fred wrote:
> Hi all,
>
> What's difference between ogrid() and mgrid ?
>
> help() or ogrid?/mgrid? in ipython
> gives the same answer.
>
> Cheers,
>
> --
> http://scipy.org/FredericPetit
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>




From ryanlists at gmail.com  Tue Jul 24 12:56:18 2007
From: ryanlists at gmail.com (Ryan Krauss)
Date: Tue, 24 Jul 2007 11:56:18 -0500
Subject: [SciPy-user] download and install python with all modules
In-Reply-To: <59416.41.194.14.11.1185291936.squirrel@webmail.aims.ac.za>
References: <46A6163E.7080501@gmail.com>
	<59416.41.194.14.11.1185291936.squirrel@webmail.aims.ac.za>
Message-ID: <c5b438120707240956y6091d5ds45d1c601f62ed142@mail.gmail.com>

Scipy does not depend on numeric.  You are better off installing numpy.

This is a platform specific question.  Have a look at
http://scipy.org/Installing_SciPy

If you just want a quick windows install this should do it:
http://prdownloads.sourceforge.net/numpy/numpy-1.0.3.win32-py2.5.exe?download
http://prdownloads.sourceforge.net/scipy/scipy-0.5.2.win32-py2.5.exe?download

But exactly what you mean by all modules is slightly vague.  You
probably also want IPython and Matplotlib at a minimum.

On 7/24/07, Kalala Mutombo Franck <franckm at aims.ac.za> wrote:
> Hi
>
> I want to download python and install it on my computer, how can I
> download python and all the needed modules? for examples
> scipy,numeric,etc.
> --
> Franck Kalala Mutombo
> African Institute for Mathematical Sciences
> www.aims.ac.za
>
> On Tue, July 24, 2007 5:09 pm, fred wrote:
> > Hi all,
> >
> > What's difference between ogrid() and mgrid ?
> >
> > help() or ogrid?/mgrid? in ipython
> > gives the same answer.
> >
> > Cheers,
> >
> > --
> > http://scipy.org/FredericPetit
> >
> > _______________________________________________
> > 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 berthe.loic at gmail.com  Tue Jul 24 12:58:03 2007
From: berthe.loic at gmail.com (=?ISO-8859-1?Q?BERTHE_Lo=EFc?=)
Date: Tue, 24 Jul 2007 18:58:03 +0200
Subject: [SciPy-user] Installing numpy and scipy on IBM AIX
Message-ID: <f042e17a0707240958j29ed19c3ga178fc4daa610c46@mail.gmail.com>

   Hi,

I'm having trouble to install numpy and scipy on IBM AIX.

As the website didn't recommend to install  the stable version of
numpy-1.0.3 and as I don't have any svn access on this AIX station, I tried
to install numpy-1.0.2


Here are the logs of my install :

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

> python setup.py build

Running from numpy source directory.
F2PY Version 2_3649
[ ...]
compile options: '-c'
xlc: _configtest.c
xlc _configtest.o -L/rdm/berthe/tmp/AIX/lib/ATLAS -lptf77blas -lptcblas
-latlas -o _configtest
ATLAS version 3.6.0 built by on Tue Jul 17 13:46:16 DFT 2007:
   UNAME    : AIX btmcsx1 1 5 005F623A4C00
   INSTFLG  :
   MMDEF    : /rdm/berthe/tmp/AIX/src/ATLAS/CONFIG/ARCHS/POWER4/xlc/gemm
   ARCHDEF  : /rdm/berthe/tmp/AIX/src/ATLAS/CONFIG/ARCHS/POWER4/xlc/misc
   F2CDEFS  : -DNoChange -DStringSunStyle
   CACHEEDGE: 524288
   F77      : /usr/bin/xlf_r, version UNKNOWN
   F77FLAGS : -qtune=pwr3 -qarch=pwr3 -O3 -qmaxmem=-1 -qfloat=hsflt
   CC       : /usr/vac/bin/xlc_r, version UNKNOWN
   CC FLAGS : -qtune=pwr3 -qarch=pwr3 -O3 -qmaxmem=-1 -qfloat=hsflt
   MCC      : /usr/vac/bin/xlc_r, version UNKNOWN
   MCCFLAGS : -qtune=pwr3 -qarch=pwr3 -O3 -qmaxmem=-1 -qfloat=hsflt
-qalias=allp
success!
removing: _configtest.c _configtest.o _configtest
  FOUND:
    libraries = ['ptf77blas', 'ptcblas', 'atlas']
    library_dirs = ['/rdm/berthe/tmp//AIX/lib/ATLAS']
    language = c
    define_macros = [('ATLAS_INFO', '"\\"3.6.0\\""')]
[ ... ]
creating build/temp.aix-5.1-2.5
creating build/temp.aix-5.1-2.5/numpy
creating build/temp.aix-5.1-2.5/numpy/core
creating build/temp.aix-5.1-2.5/numpy/core/src
compile options: '-Ibuild/src.aix-5.1-2.5/numpy/core/src
-Inumpy/core/include -Ibuild/src.aix- 5.1-2.5/numpy/core -Inumpy/core/src
-Inumpy/core/include -I/rdm/berthe/tmp/AIX/include/python2.5 -c'
xlc: numpy/core/src/multiarraymodule.c
"numpy/core/include/numpy/ndarrayobject.h", line 191.24: 1506-275 (S)
Unexpected text ',' encountered.
"numpy/core/include/numpy/ndarrayobject.h", line 198.26: 1506-275 (S)
Unexpected text ',' encountered.
"numpy/core/include/numpy/ndarrayobject.h", line 210.26: 1506-275 (S)
Unexpected text ',' encountered.
"numpy/core/src/arrayobject.c", line 4489.17: 1506-068 (E) Operation between
types "int(*)(void*,void*,int,int)" and "void*" is not allowed.
"numpy/core/src/arrayobject.c", line 4594.61: 1506-280 (E) Function argument
assignment between types "void*" and "int(*)(unsigned long*,unsigned
long*,int,int)" is not allowed.
"numpy/core/src/arrayobject.c", line 4598.61: 1506-280 (E) Function argument
assignment between types "void*" and "int(*)(char*,char*,int,int)" is not
allowed.
"numpy/core/src/arrayobject.c", line 5472.39: 1506-068 (E) Operation between
types "int(*)(struct PyArrayObject*,struct _object*)" and "void*" is not
allowed.
"numpy/core/src/arrayobject.c", line 7528.42: 1506-068 (E) Operation between
types "void(*)(void*,void*,int,void*,void*)" and "void*" is not allowed.
"numpy/core/include/numpy/ndarrayobject.h", line 191.24: 1506-275 (S)
Unexpected text ',' encountered.
"numpy/core/include/numpy/ndarrayobject.h", line 198.26: 1506-275 (S)
Unexpected text ',' encountered.
"numpy/core/include/numpy/ndarrayobject.h", line 210.26: 1506-275 (S)
Unexpected text ',' encountered.
"numpy/core/src/arrayobject.c", line 4489.17: 1506-068 (E) Operation between
types "int(*)(void*,void*,int,int)" and "void*" is not allowed.
"numpy/core/src/arrayobject.c", line 4594.61: 1506-280 (E) Function argument
assignment between types "void*" and "int(*)(unsigned long*,unsigned
long*,int,int)" is not allowed.
"numpy/core/src/arrayobject.c", line 4598.61: 1506-280 (E) Function argument
assignment between types "void*" and "int(*)(char*,char*,int,int)" is not
allowed.
"numpy/core/src/arrayobject.c", line 5472.39: 1506-068 (E) Operation between
types "int(*)(struct PyArrayObject*,struct _object*)" and "void*" is not
allowed.
"numpy/core/src/arrayobject.c", line 7528.42: 1506-068 (E) Operation between
types "void(*)(void*,void*,int,void*,void*)" and "void*" is not allowed.
error: Command "xlc -DNDEBUG -O -Ibuild/src.aix-5.1-2.5/numpy/core/src
-Inumpy/core/include -Ibuild/src.aix- 5.1-2.5/numpy/core -Inumpy/core/src
-Inumpy/core/include -I/rdm/berthe/tmp/AIX/include/python2.5 -c
numpy/core/src/multiarraymodule.c -o
build/temp.aix-5.1-2.5/numpy/core/src/multiarraymodule.o"
failed with exit status 1

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

I'm using the ActivePython-2.5.1.1-aix5-powerpc version of python, available
on http://www.activestate.com/Products/activepython/ .

Have you got any clue to fix this ?

Furthermore, I didn't find any SVN snapshot archive for numpy or scipy. Is
there any ?
As numpy-1.0.3 is not recommended for installing scipy and as I 'm probably
not the only one who cannot use svn, I think this could be very handy.

Regards,

-- 
LB
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070724/e09b1bf2/attachment.html>

From robert.kern at gmail.com  Tue Jul 24 14:36:04 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 24 Jul 2007 13:36:04 -0500
Subject: [SciPy-user] mgrid/ogrid...
In-Reply-To: <46A6163E.7080501@gmail.com>
References: <46A6163E.7080501@gmail.com>
Message-ID: <46A64694.9080307@gmail.com>

fred wrote:
> Hi all,
> 
> What's difference between ogrid() and mgrid ?

In [1]: from numpy import *

In [2]: mgrid[0:5,0:5]
Out[2]:
array([[[0, 0, 0, 0, 0],
        [1, 1, 1, 1, 1],
        [2, 2, 2, 2, 2],
        [3, 3, 3, 3, 3],
        [4, 4, 4, 4, 4]],

       [[0, 1, 2, 3, 4],
        [0, 1, 2, 3, 4],
        [0, 1, 2, 3, 4],
        [0, 1, 2, 3, 4],
        [0, 1, 2, 3, 4]]])

In [3]: ogrid[0:5,0:5]
Out[3]:
[array([[0],
       [1],
       [2],
       [3],
       [4]]),
 array([[0, 1, 2, 3, 4]])]


mgrid gives you the fully broadcasted arrays. ogrid gives you partial arrays
which will broadcast against each other. Depending on how you are using them,
ogrid's results may be more efficient to use because you can often do the bulk
of your computations on the smaller arrays and only do a handful of operations
that broadcast against each other to form the final, full result.

-- 
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 spmcinerney at hotmail.com  Tue Jul 24 15:02:40 2007
From: spmcinerney at hotmail.com (Stephen McInerney)
Date: Tue, 24 Jul 2007 12:02:40 -0700
Subject: [SciPy-user] download and install python with all modules
In-Reply-To: <mailman.38432.1185296290.30507.scipy-user@scipy.org>
Message-ID: <BAY111-F3075131648666685A49E07A0F00@phx.gbl>


a) A very good zero-hassle batteries-included Python distribution for 
Windows
with scientific and visualization libraries is Enthought Python:
http://code.enthought.com/enthon/
Current release is Python 2.4.3 ; (Python 2.5.x is behind but supposedly due 
"this August").
Enthought Python 2.4.3 has NumPy 0.9.9.2706 + SciPy 0.5.0.2033 + IPython 
0.7.2 + Matplotlib 0.87.3.2478 + wx +much more ...
Typically Enthought bundles lag Python releases by ~1 yr.
You might contact enthought-dev at mail.enthought.com yourselves to ask for 
2.5.x :)

b) Compiling 15+ individual packages on Windows and reconciling multiple 
version conflicts
is a huge pain and a barrier to adoption, in particular if we need the retro 
version
Visual Express C++ for 2003. It's just not going to fly.

To add a question I have always wanted to know, are there any firm plans to 
move
NumPy + SciPy into Python? It really needs an arbitrary-precision numercial 
type (no, 'Decimal' is not good enough, too slow and basic methods like pow 
and math.* don't work on it )

Thanks,
Stephen

_________________________________________________________________
http://newlivehotmail.com



From robert.kern at gmail.com  Tue Jul 24 15:15:51 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 24 Jul 2007 14:15:51 -0500
Subject: [SciPy-user] download and install python with all modules
In-Reply-To: <BAY111-F3075131648666685A49E07A0F00@phx.gbl>
References: <BAY111-F3075131648666685A49E07A0F00@phx.gbl>
Message-ID: <46A64FE7.90709@gmail.com>

Stephen McInerney wrote:
> a) A very good zero-hassle batteries-included Python distribution for 
> Windows
> with scientific and visualization libraries is Enthought Python:
> http://code.enthought.com/enthon/
> Current release is Python 2.4.3 ; (Python 2.5.x is behind but supposedly due 
> "this August").
> Enthought Python 2.4.3 has NumPy 0.9.9.2706 + SciPy 0.5.0.2033 + IPython 
> 0.7.2 + Matplotlib 0.87.3.2478 + wx +much more ...
> Typically Enthought bundles lag Python releases by ~1 yr.
> You might contact enthought-dev at mail.enthought.com yourselves to ask for 
> 2.5.x :)

We're not updating the monolithic distribution anymore. Instead we are building
binaries for each package separately and using Python eggs as a distribution
mechanism.

  http://code.enthought.com/enstaller/

> b) Compiling 15+ individual packages on Windows and reconciling multiple 
> version conflicts
> is a huge pain and a barrier to adoption, in particular if we need the retro 
> version
> Visual Express C++ for 2003. It's just not going to fly.
> 
> To add a question I have always wanted to know, are there any firm plans to 
> move
> NumPy + SciPy into Python?

None. There is a PEP for putting the C array interface into Python for efficient
communication of multidimensional array data between C extensions. In a few
years, we might be able to contribute a slimmed down array type and ufuncs for
use at the Python level, but that seems less likely. scipy will never be a part
of Python.

> It really needs an arbitrary-precision numercial 
> type (no, 'Decimal' is not good enough, too slow and basic methods like pow 
> and math.* don't work on it )

We don't have an arbitrary-precision numerical type 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 fredmfp at gmail.com  Tue Jul 24 15:19:49 2007
From: fredmfp at gmail.com (fred)
Date: Tue, 24 Jul 2007 21:19:49 +0200
Subject: [SciPy-user] mgrid/ogrid...
In-Reply-To: <46A64694.9080307@gmail.com>
References: <46A6163E.7080501@gmail.com> <46A64694.9080307@gmail.com>
Message-ID: <46A650D5.9030909@gmail.com>

Robert Kern a ?crit :
> fred wrote:
>   
>> Hi all,
>>
>> What's difference between ogrid() and mgrid ?
>>     
>
> In [1]: from numpy import *
>
> In [2]: mgrid[0:5,0:5]
> Out[2]:
> array([[[0, 0, 0, 0, 0],
>         [1, 1, 1, 1, 1],
>         [2, 2, 2, 2, 2],
>         [3, 3, 3, 3, 3],
>         [4, 4, 4, 4, 4]],
>
>        [[0, 1, 2, 3, 4],
>         [0, 1, 2, 3, 4],
>         [0, 1, 2, 3, 4],
>         [0, 1, 2, 3, 4],
>         [0, 1, 2, 3, 4]]])
>
> In [3]: ogrid[0:5,0:5]
> Out[3]:
> [array([[0],
>        [1],
>        [2],
>        [3],
>        [4]]),
>  array([[0, 1, 2, 3, 4]])]
>
>
>   
Well, I should have try before, and not only read help, sorry.
> mgrid gives you the fully broadcasted arrays. ogrid gives you partial arrays
> which will broadcast against each other. Depending on how you are using them,
> ogrid's results may be more efficient to use because you can often do the bulk
> of your computations on the smaller arrays and only do a handful of operations
> that broadcast against each other to form the final, full result.
>   
Thanks for explanation as clear as possible, as usual ;-)

Cheers,

-- 
http://scipy.org/FredericPetit



From shao at msg.ucsf.edu  Tue Jul 24 15:42:07 2007
From: shao at msg.ucsf.edu (Lin Shao)
Date: Tue, 24 Jul 2007 12:42:07 -0700
Subject: [SciPy-user] mgrid/ogrid...
In-Reply-To: <46A64694.9080307@gmail.com>
References: <46A6163E.7080501@gmail.com> <46A64694.9080307@gmail.com>
Message-ID: <baa383440707241242v329270c8pa26b1b6a91616a0d@mail.gmail.com>

So, is there any difference between N.indices() calls and N.mgrid[]?


From robert.kern at gmail.com  Tue Jul 24 15:47:57 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 24 Jul 2007 14:47:57 -0500
Subject: [SciPy-user] mgrid/ogrid...
In-Reply-To: <baa383440707241242v329270c8pa26b1b6a91616a0d@mail.gmail.com>
References: <46A6163E.7080501@gmail.com> <46A64694.9080307@gmail.com>
	<baa383440707241242v329270c8pa26b1b6a91616a0d@mail.gmail.com>
Message-ID: <46A6576D.1090604@gmail.com>

Lin Shao wrote:
> So, is there any difference between N.indices() calls and N.mgrid[]?

Yes!

In [8]: mgrid[0:5:11j]
Out[8]: array([ 0. ,  0.5,  1. ,  1.5,  2. ,  2.5,  3. ,  3.5,  4. ,  4.5,  5. ])

In [9]: mgrid[0:5:0.5]
Out[9]: array([ 0. ,  0.5,  1. ,  1.5,  2. ,  2.5,  3. ,  3.5,  4. ,  4.5])

In [11]: mgrid[0:5:0.3]
Out[11]:
array([ 0. ,  0.3,  0.6,  0.9,  1.2,  1.5,  1.8,  2.1,  2.4,  2.7,  3. ,
        3.3,  3.6,  3.9,  4.2,  4.5,  4.8])

-- 
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 hetland at tamu.edu  Tue Jul 24 16:55:48 2007
From: hetland at tamu.edu (Rob Hetland)
Date: Tue, 24 Jul 2007 16:55:48 -0400
Subject: [SciPy-user] pynetcdf crashing on Python 2.5
In-Reply-To: <Pine.SOC.4.61.0707181655350.5424@windom.ucar.edu>
References: <Pine.SOC.4.61.0707181655350.5424@windom.ucar.edu>
Message-ID: <775208E0-E439-4D63-BD4B-7A06D8185E9A@tamu.edu>


On Jul 18, 2007, at 7:02 PM, Mary Haley wrote:

> Somebody told me about this thread, and I thought I'd jump in and
> offer another solution.

..And yet another:

http://code.google.com/p/netcdf4-python/

Although based on netcdf4, this package *does* support both reading  
and writing 'classic' netcdf3 files.  I use this package for all of  
my work, and have been very happy with it.  Particularly useful is  
the multifile support, as I often have 10+ gigs of data files spread  
out over 5+ files.

-Rob

----
Rob Hetland, Associate Professor
Dept. of Oceanography, Texas A&M University
http://pong.tamu.edu/~rob
phone: 979-458-0096, fax: 979-845-6331




From josh8912 at yahoo.com  Tue Jul 24 17:28:28 2007
From: josh8912 at yahoo.com (JJ)
Date: Tue, 24 Jul 2007 21:28:28 +0000 (UTC)
Subject: [SciPy-user] missing libacml.so when using cgi script
References: <186675.64366.qm@web54011.mail.re2.yahoo.com>
	<46A3F747.9030301@gmail.com>
Message-ID: <loom.20070724T232217-352@post.gmane.org>

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


> This won't work. The LD_LIBRARY_PATH environment variable needs to be set
 before
> the process starts, not just before the library is loaded. Read the Apache
> documentation for how to set up the environment for your CGI scripts.
> 

Thanks Robert:
I have now tried using the SetEnv directive, as Apache instructs.  I added the
following to the end of the httpd.config file
SetEnv LD_LIBRARY_PATH /home/john/usr/acml/gnu64/lib
SetEnv LAPACK /home/john/usr/acml/gnu64/lib/libacml.so
SetEnv LD_RUN_PATH /home/john/usr/acml/gnu64/lib

When I run my cgi script, I can verify that indeed these environment vaiables
have been set.  But the cgi script still fails when I try to import scipy,
giving the message that libacml.so cannot be found.  This is odd, as I can
import scipy just fine in a terminal window.  Any ideas?  I have written to the
Apache list also.
John




From david at ar.media.kyoto-u.ac.jp  Tue Jul 24 21:04:27 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Wed, 25 Jul 2007 10:04:27 +0900
Subject: [SciPy-user] download and install python with all modules
In-Reply-To: <59416.41.194.14.11.1185291936.squirrel@webmail.aims.ac.za>
References: <46A6163E.7080501@gmail.com>
	<59416.41.194.14.11.1185291936.squirrel@webmail.aims.ac.za>
Message-ID: <46A6A19B.1040805@ar.media.kyoto-u.ac.jp>

Kalala Mutombo Franck wrote:
> Hi
>
> I want to download python and install it on my computer, how can I
> download python and all the needed modules? for examples
> scipy,numeric,etc.
>   
Which platform are you on ? On Linux, depending on your distribution, it 
could be a simple matter of installing packaged binaries (for windows, 
other already pointed some possible solution).

cheers,

David


From oliver.tomic at matforsk.no  Wed Jul 25 07:06:18 2007
From: oliver.tomic at matforsk.no (oliver.tomic at matforsk.no)
Date: Wed, 25 Jul 2007 13:06:18 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <c5b438120707050536v7f863ee4jcc1b914cf4e7ad73@mail.gmail.com>
Message-ID: <OF0DA3F626.DFF14E7C-ONC1257323.003B6E11-C1257323.003D2F2A@matforsk.no>

Hi list,

I have searched the archive, but I really didn't come up with any
conclusion.

Is there now support in scipy.io (scipy 0.52) for reading .mat-files from
Matlab 7? I undestand that it supports older versions (v4 and v5). Are
there any plans to support for Matlab 7 .mat-files?

I found this mail from Nick Fotopoulos, but this might be outdated:
http://article.gmane.org/gmane.comp.python.scientific.user/4835/match=matlab+7+io+loadmat

Any comments appreciated.

cheers
Oliver




From fredmfp at gmail.com  Wed Jul 25 09:50:34 2007
From: fredmfp at gmail.com (fred)
Date: Wed, 25 Jul 2007 15:50:34 +0200
Subject: [SciPy-user] array dtype...
Message-ID: <46A7552A.1090402@gmail.com>

Hi all,

I want to set arrays for a given datatype.

I know I can use dtype='i', dtype='f', dtype='d' etc, which stand for
dtype=int32, dtype=float32, dtype=float64.

But I would like to know all the available types in the two declaration 
type.

I googled, browsed scipy.org, but did not find anything about this issue.

For instance, I would like to know if there is an equivalent "in one 
character" to the uint8 data type.

Any pointer, url are welcome.

TIA.

Cheers,

-- 
http://scipy.org/FredericPetit



From stefan at sun.ac.za  Wed Jul 25 09:57:31 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Wed, 25 Jul 2007 15:57:31 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <OF0DA3F626.DFF14E7C-ONC1257323.003B6E11-C1257323.003D2F2A@matforsk.no>
References: <c5b438120707050536v7f863ee4jcc1b914cf4e7ad73@mail.gmail.com>
	<OF0DA3F626.DFF14E7C-ONC1257323.003B6E11-C1257323.003D2F2A@matforsk.no>
Message-ID: <20070725135731.GJ8728@mentat.za.net>

Hi Oliver

On Wed, Jul 25, 2007 at 01:06:18PM +0200, oliver.tomic at matforsk.no wrote:
> Is there now support in scipy.io (scipy 0.52) for reading .mat-files from
> Matlab 7? I undestand that it supports older versions (v4 and v5). Are
> there any plans to support for Matlab 7 .mat-files?

>From the docstring:

"v4 (Level 1.0), v6 and v7.1 matfiles are supported."

Cheers
St?fan


From stefan at sun.ac.za  Wed Jul 25 10:49:37 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Wed, 25 Jul 2007 16:49:37 +0200
Subject: [SciPy-user] array dtype...
In-Reply-To: <46A7552A.1090402@gmail.com>
References: <46A7552A.1090402@gmail.com>
Message-ID: <20070725144937.GA26793@mentat.za.net>

On Wed, Jul 25, 2007 at 03:50:34PM +0200, fred wrote:
> For instance, I would like to know if there is an equivalent "in one 
> character" to the uint8 data type.

N.dtype(uint8).char

Cheers
St?fan


From rhc28 at cornell.edu  Wed Jul 25 11:05:28 2007
From: rhc28 at cornell.edu (Rob Clewley)
Date: Wed, 25 Jul 2007 11:05:28 -0400
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <20070725135731.GJ8728@mentat.za.net>
References: <c5b438120707050536v7f863ee4jcc1b914cf4e7ad73@mail.gmail.com>
	<OF0DA3F626.DFF14E7C-ONC1257323.003B6E11-C1257323.003D2F2A@matforsk.no>
	<20070725135731.GJ8728@mentat.za.net>
Message-ID: <a749952d0707250805x7a2d50e2ybef82d005b6eab5e@mail.gmail.com>

Stefan,

scipy.io.loadmat from scipy 0.5.2 doesn't work on anything but v4
files for me. I generated extremely simple mat files (one small matrix
in 'test.mat') with my matlab 7.3 using the -v6, -v5, and -v4 options,
and scipy complained with this in all but the -v4 case:

In [1]: import scipy.io as io

In [2]: io.loadmat('test')
---------------------------------------------------------------------------
exceptions.AttributeError                            Traceback (most
recent call last)

c:\temp\<ipython console>

C:\Python24\lib\site-packages\scipy\io\mio.py in loadmat(file_name, mdict, appen
dmat, basename, **kwargs)
     94     '''
     95     MR = mat_reader_factory(file_name, appendmat, **kwargs)
---> 96     matfile_dict = MR.get_variables()
     97     if mdict is not None:
     98         mdict.update(matfile_dict)

C:\Python24\lib\site-packages\scipy\io\miobase.py in
get_variables(self, variable_names)
    267             variable_names = [variable_names]
    268         self.mat_stream.seek(0)
--> 269         mdict = self.file_header()
    270         mdict['__globals__'] = []
    271         while not self.end_of_stream():

C:\Python24\lib\site-packages\scipy\io\mio5.py in file_header(self)
    508         hdict = {}
    509         hdr = self.read_dtype(self.dtypes['file_header'])
--> 510         hdict['__header__'] = hdr['description'].strip(' \t\n\000')
    511         v_major = hdr['version'] >> 8
    512         v_minor = hdr['version'] & 0xFF

AttributeError: 'numpy.ndarray' object has no attribute 'strip'


This is with numpy 1.0.2. I wonder if this is the problem that Oliver has?

-Rob

On 25/07/07, Stefan van der Walt <stefan at sun.ac.za> wrote:
> Hi Oliver
>
> On Wed, Jul 25, 2007 at 01:06:18PM +0200, oliver.tomic at matforsk.no wrote:
> > Is there now support in scipy.io (scipy 0.52) for reading .mat-files from
> > Matlab 7? I undestand that it supports older versions (v4 and v5). Are
> > there any plans to support for Matlab 7 .mat-files?
>
> >From the docstring:
>
> "v4 (Level 1.0), v6 and v7.1 matfiles are supported."
>
> Cheers
> St?fan
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


From fredmfp at gmail.com  Wed Jul 25 11:06:15 2007
From: fredmfp at gmail.com (fred)
Date: Wed, 25 Jul 2007 17:06:15 +0200
Subject: [SciPy-user] array dtype...
In-Reply-To: <20070725144937.GA26793@mentat.za.net>
References: <46A7552A.1090402@gmail.com> <20070725144937.GA26793@mentat.za.net>
Message-ID: <46A766E7.5080601@gmail.com>

Stefan van der Walt a ?crit :
> N.dtype(uint8).char
>   
Great, I like the answer ! :-))

Thanks a lot.

-- 
http://scipy.org/FredericPetit



From peridot.faceted at gmail.com  Wed Jul 25 11:36:16 2007
From: peridot.faceted at gmail.com (Anne Archibald)
Date: Wed, 25 Jul 2007 11:36:16 -0400
Subject: [SciPy-user] filtering without phase shift
In-Reply-To: <46A5B74C.2060304@ru.nl>
References: <c5b438120707232027w32c8a564v346f9292501616e8@mail.gmail.com>
	<1185261264.5678.9.camel@pc1.cole.uklinux.net>
	<46A5B74C.2060304@ru.nl>
Message-ID: <ce557a360707250836jf90fd93ya3fdd5d4a8d85659@mail.gmail.com>

On 24/07/07, Stef Mientki <s.mientki at ru.nl> wrote:
>
>
> Bryan Cole wrote:
> > The easiest way to do non-phase-shifted filtering is to use FFTs i.e.
> >
> > FFT -> Low Pass -> iFFT
> >
> > The only reason to move to time-domain filters is if you must do
> > filtering in real-time where the complete dataset is (yet) unknown and
> > you must obey the laws of causality.
> >
> > Using a frequency domain filter, you can have whatever cutoff you like.
> >
> I'm a bit rusty, but isn't there a real danger by filtering in the
> frequency domain,
> creating a rectangle cutoff in the frequency domain,
> leads to an infinite ringing signal in the time-domain.

There is a concern with ringing if you just use a brick-wall cutoff
(though often with an FFT you have so much frequency resolution the
ringing isn't actually a problem). The easiest way to do this is to
use a softer edge - there are standard filter shapes (Butterworth and
so on) which will minimize ringing, but any relatively smooth falloff
will help.

It sounds like this is not quite your issue, but if what you need is a
filter that preserves peak height, Kalman filtering is what you need.

Anne


From emanuelez at gmail.com  Wed Jul 25 11:46:40 2007
From: emanuelez at gmail.com (Emanuele Zattin)
Date: Wed, 25 Jul 2007 17:46:40 +0200
Subject: [SciPy-user] fitting mixed gaussians
Message-ID: <d7b942cd0707250846q752c842k1f30b3fafa1d34e@mail.gmail.com>

Hello,
i'm trying to perform mixed-gaussian fitting on some small greyscale
images (5x5 pixels, maximum 9x9). right now i'm using a leastsq
approach but it tends to get pretty slow and unreliable when it comes
to fit 4-5 2D gaussians even in such a small domain.
I've noticed PyEM and i have the feeling it might be useful for me,
but here i'm dealing with pixel intensities, not distributions. Any
hint is welcome!

Emanuele


From gael.varoquaux at normalesup.org  Wed Jul 25 11:51:32 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Wed, 25 Jul 2007 17:51:32 +0200
Subject: [SciPy-user] fitting mixed gaussians
In-Reply-To: <d7b942cd0707250846q752c842k1f30b3fafa1d34e@mail.gmail.com>
References: <d7b942cd0707250846q752c842k1f30b3fafa1d34e@mail.gmail.com>
Message-ID: <20070725155132.GD28080@clipper.ens.fr>

On Wed, Jul 25, 2007 at 05:46:40PM +0200, Emanuele Zattin wrote:
> i'm trying to perform mixed-gaussian fitting on some small greyscale
> images (5x5 pixels, maximum 9x9). right now i'm using a leastsq
> approach but it tends to get pretty slow and unreliable when it comes
> to fit 4-5 2D gaussians even in such a small domain.

Is the gaussian spot well centered on the image ?

If so then I reckon you are interested in finding its width ? Maybe I am
wrong, byt if I am right an approach that works very well is to determine
the second momentum of the intensity/position distribution. This is not
an optimisation procedure and is very quick. Reliability depends on the
kind of data you are processing.

Some code doing this can be found on
http://scipy.org/Cookbook/FittingData#head-11870c917b101bb0d4b34900a0da1b7deb613bf7

HTH,

Ga?l


From emanuelez at gmail.com  Wed Jul 25 11:58:50 2007
From: emanuelez at gmail.com (Emanuele Zattin)
Date: Wed, 25 Jul 2007 17:58:50 +0200
Subject: [SciPy-user] fitting mixed gaussians
In-Reply-To: <20070725155132.GD28080@clipper.ens.fr>
References: <d7b942cd0707250846q752c842k1f30b3fafa1d34e@mail.gmail.com>
	<20070725155132.GD28080@clipper.ens.fr>
Message-ID: <d7b942cd0707250858g39c47770x1e7fc9b16ecc8d2d@mail.gmail.com>

On 7/25/07, Gael Varoquaux <gael.varoquaux at normalesup.org> wrote:
> On Wed, Jul 25, 2007 at 05:46:40PM +0200, Emanuele Zattin wrote:
> > i'm trying to perform mixed-gaussian fitting on some small greyscale
> > images (5x5 pixels, maximum 9x9). right now i'm using a leastsq
> > approach but it tends to get pretty slow and unreliable when it comes
> > to fit 4-5 2D gaussians even in such a small domain.
>
> Is the gaussian spot well centered on the image ?
>
> If so then I reckon you are interested in finding its width ? Maybe I am
> wrong, byt if I am right an approach that works very well is to determine
> the second momentum of the intensity/position distribution. This is not
> an optimisation procedure and is very quick. Reliability depends on the
> kind of data you are processing.
>
> Some code doing this can be found on
> http://scipy.org/Cookbook/FittingData#head-11870c917b101bb0d4b34900a0da1b7deb613bf7
>
> HTH,
>
> Ga?l

The gaussian I am interested is indeed at the center of the images,
but very often more maxima are present at the boundaries so i try to
fit more gaussians. I am aware of the cookbook page... it's really
interesting and it helped me a lot in the first steps. On the other
side a momentum approach is not as easy when more gaussians are
present.


From stefan at sun.ac.za  Wed Jul 25 11:55:48 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Wed, 25 Jul 2007 17:55:48 +0200
Subject: [SciPy-user] fitting mixed gaussians
In-Reply-To: <d7b942cd0707250846q752c842k1f30b3fafa1d34e@mail.gmail.com>
References: <d7b942cd0707250846q752c842k1f30b3fafa1d34e@mail.gmail.com>
Message-ID: <20070725155548.GC26793@mentat.za.net>

Hi Emanuele

On Wed, Jul 25, 2007 at 05:46:40PM +0200, Emanuele Zattin wrote:
> i'm trying to perform mixed-gaussian fitting on some small greyscale
> images (5x5 pixels, maximum 9x9). right now i'm using a leastsq
> approach but it tends to get pretty slow and unreliable when it comes
> to fit 4-5 2D gaussians even in such a small domain.
> I've noticed PyEM and i have the feeling it might be useful for me,
> but here i'm dealing with pixel intensities, not distributions. Any
> hint is welcome!

Expectation maximisation should work well for your needs.  Take a look
at the lecture notes at

http://www.dsp.sun.ac.za/pr813/lectures/lecture06/pr813_lecture06.pdf

You can view your MxN image as an MN-dimensional vector and follow the
same process as you would for other types of features.  If the
dimensionality becomes a problem, you'll have to look at feature
extraction or dimension reduction (using methods like the
Karhunen-Loeve transform or linear discriminant analysis).

Regards
St?fan


From Joris.DeRidder at ster.kuleuven.be  Wed Jul 25 12:00:26 2007
From: Joris.DeRidder at ster.kuleuven.be (Joris De Ridder)
Date: Wed, 25 Jul 2007 18:00:26 +0200
Subject: [SciPy-user] array dtype...
In-Reply-To: <46A7552A.1090402@gmail.com>
References: <46A7552A.1090402@gmail.com>
Message-ID: <B6764343-482F-4199-A806-CFFCC738EC2D@ster.kuleuven.be>


> I googled, browsed scipy.org, but did not find anything about this  
> issue.
>
> For instance, I would like to know if there is an equivalent "in one
> character" to the uint8 data type.

Surf to the Numpy Example List: http://www.scipy.org/Numpy_Example_List
and look under the topic "dtype".

Cheers,
J.







From gael.varoquaux at normalesup.org  Wed Jul 25 12:08:34 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Wed, 25 Jul 2007 18:08:34 +0200
Subject: [SciPy-user] fitting mixed gaussians
In-Reply-To: <d7b942cd0707250858g39c47770x1e7fc9b16ecc8d2d@mail.gmail.com>
References: <d7b942cd0707250846q752c842k1f30b3fafa1d34e@mail.gmail.com>
	<20070725155132.GD28080@clipper.ens.fr>
	<d7b942cd0707250858g39c47770x1e7fc9b16ecc8d2d@mail.gmail.com>
Message-ID: <20070725160834.GE28080@clipper.ens.fr>

On Wed, Jul 25, 2007 at 05:58:50PM +0200, Emanuele Zattin wrote:
> The gaussian I am interested is indeed at the center of the images,
> but very often more maxima are present at the boundaries so i try to
> fit more gaussians. I am aware of the cookbook page... it's really
> interesting and it helped me a lot in the first steps. On the other
> side a momentum approach is not as easy when more gaussians are
> present.

Neither is the fitting.

Good luck with this problem, and keep me posted, I am interested in a
good solution.

Ga?l


From fredmfp at gmail.com  Wed Jul 25 12:09:51 2007
From: fredmfp at gmail.com (fred)
Date: Wed, 25 Jul 2007 18:09:51 +0200
Subject: [SciPy-user] array dtype...
In-Reply-To: <B6764343-482F-4199-A806-CFFCC738EC2D@ster.kuleuven.be>
References: <46A7552A.1090402@gmail.com>
	<B6764343-482F-4199-A806-CFFCC738EC2D@ster.kuleuven.be>
Message-ID: <46A775CF.6080304@gmail.com>

Joris De Ridder a ?crit :
>> I googled, browsed scipy.org, but did not find anything about this  
>> issue.
>>
>> For instance, I would like to know if there is an equivalent "in one
>> character" to the uint8 data type.
>>     
>
> Surf to the Numpy Example List: http://www.scipy.org/Numpy_Example_List
> and look under the topic "dtype".
>   

Great too ;-)

Bookmarked.
(just a pity that google did not find it, or I'm an idiot ;-)

Thanks.

-- 
http://scipy.org/FredericPetit



From stefan at sun.ac.za  Wed Jul 25 13:34:42 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Wed, 25 Jul 2007 19:34:42 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <a749952d0707250805x7a2d50e2ybef82d005b6eab5e@mail.gmail.com>
References: <c5b438120707050536v7f863ee4jcc1b914cf4e7ad73@mail.gmail.com>
	<OF0DA3F626.DFF14E7C-ONC1257323.003B6E11-C1257323.003D2F2A@matforsk.no>
	<20070725135731.GJ8728@mentat.za.net>
	<a749952d0707250805x7a2d50e2ybef82d005b6eab5e@mail.gmail.com>
Message-ID: <20070725173442.GB5361@mentat.za.net>

Hi Rob

On Wed, Jul 25, 2007 at 11:05:28AM -0400, Rob Clewley wrote:
> scipy.io.loadmat from scipy 0.5.2 doesn't work on anything but v4
> files for me. I generated extremely simple mat files (one small matrix
> in 'test.mat') with my matlab 7.3 using the -v6, -v5, and -v4 options,
> and scipy complained with this in all but the -v4 case:

Matthew made a number of changes since v0.5.2, so you may need to
upgrade.  He included test data files from Matlab 4 through 7.1, all
which are loaded as part of the automated unit test suite.

If you send me your file, I'd be glad to check that it loads under the
latest version.

Regards
St?fan


From stefan at sun.ac.za  Wed Jul 25 13:43:06 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Wed, 25 Jul 2007 19:43:06 +0200
Subject: [SciPy-user] fitting mixed gaussians
In-Reply-To: <20070725160834.GE28080@clipper.ens.fr>
References: <d7b942cd0707250846q752c842k1f30b3fafa1d34e@mail.gmail.com>
	<20070725155132.GD28080@clipper.ens.fr>
	<d7b942cd0707250858g39c47770x1e7fc9b16ecc8d2d@mail.gmail.com>
	<20070725160834.GE28080@clipper.ens.fr>
Message-ID: <20070725174306.GC5361@mentat.za.net>

On Wed, Jul 25, 2007 at 06:08:34PM +0200, Gael Varoquaux wrote:
> On Wed, Jul 25, 2007 at 05:58:50PM +0200, Emanuele Zattin wrote:
> > The gaussian I am interested is indeed at the center of the images,
> > but very often more maxima are present at the boundaries so i try to
> > fit more gaussians. I am aware of the cookbook page... it's really
> > interesting and it helped me a lot in the first steps. On the other
> > side a momentum approach is not as easy when more gaussians are
> > present.
> 
> Neither is the fitting.

Sorry, I misunderstood your application.  You don't want to use the
images as features, but want to identify features *in* the image.
Since your model is fitting so badly, you may want to consider another
model/method.

Since the images are only 5x5, you can use routines available in scipy
to do a bi-polynomial or higher order surface fit.  The maximum
turning point of the polynomial then indicates your feature centre.

Regards
St?fan


From oliver.tomic at matforsk.no  Wed Jul 25 15:25:36 2007
From: oliver.tomic at matforsk.no (oliver.tomic at matforsk.no)
Date: Wed, 25 Jul 2007 21:25:36 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <a749952d0707250805x7a2d50e2ybef82d005b6eab5e@mail.gmail.com>
Message-ID: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>


Rob and Stefan,

the problem below described by Rob is exactly what I have experienced. I
get the same error. I use scipy 0.5.2 and numpy 1.0.3.

Is there a newer version of scipy than 0.5.2 or do I have to use the svn?

Here is a file that failed to load with scipy.io:
(See attached file: brod.mat)
It was created with Matlab 7.3 with backward compatibility to (-v6).

Cheers
Oliver



scipy-user-bounces at scipy.org wrote on 25.07.2007 17:05:28:

> Stefan,
>
> scipy.io.loadmat from scipy 0.5.2 doesn't work on anything but v4
> files for me. I generated extremely simple mat files (one small matrix
> in 'test.mat') with my matlab 7.3 using the -v6, -v5, and -v4 options,
> and scipy complained with this in all but the -v4 case:
>
> In [1]: import scipy.io as io
>
> In [2]: io.loadmat('test')
>
---------------------------------------------------------------------------
> exceptions.AttributeError                            Traceback (most
> recent call last)
>
> c:\temp\<ipython console>
>
> C:\Python24\lib\site-packages\scipy\io\mio.py in loadmat(file_name,
> mdict, appen
> dmat, basename, **kwargs)
>      94     '''
>      95     MR = mat_reader_factory(file_name, appendmat, **kwargs)
> ---> 96     matfile_dict = MR.get_variables()
>      97     if mdict is not None:
>      98         mdict.update(matfile_dict)
>
> C:\Python24\lib\site-packages\scipy\io\miobase.py in
> get_variables(self, variable_names)
>     267             variable_names = [variable_names]
>     268         self.mat_stream.seek(0)
> --> 269         mdict = self.file_header()
>     270         mdict['__globals__'] = []
>     271         while not self.end_of_stream():
>
> C:\Python24\lib\site-packages\scipy\io\mio5.py in file_header(self)
>     508         hdict = {}
>     509         hdr = self.read_dtype(self.dtypes['file_header'])
> --> 510         hdict['__header__'] = hdr['description'].strip('
\t\n\000')
>     511         v_major = hdr['version'] >> 8
>     512         v_minor = hdr['version'] & 0xFF
>
> AttributeError: 'numpy.ndarray' object has no attribute 'strip'
>
>
> This is with numpy 1.0.2. I wonder if this is the problem that Oliver
has?
>
> -Rob
>
> On 25/07/07, Stefan van der Walt <stefan at sun.ac.za> wrote:
> > Hi Oliver
> >
> > On Wed, Jul 25, 2007 at 01:06:18PM +0200, oliver.tomic at matforsk.no
wrote:
> > > Is there now support in scipy.io (scipy 0.52) for reading .mat-files
from
> > > Matlab 7? I undestand that it supports older versions (v4 and v5).
Are
> > > there any plans to support for Matlab 7 .mat-files?
> >
> > >From the docstring:
> >
> > "v4 (Level 1.0), v6 and v7.1 matfiles are supported."
> >
> > Cheers
> > St?fan
> > _______________________________________________
> > 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 --------------
A non-text attachment was scrubbed...
Name: brod.mat
Type: application/octet-stream
Size: 8656 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070725/3a04b33d/attachment.obj>

From nwagner at iam.uni-stuttgart.de  Wed Jul 25 15:30:55 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Wed, 25 Jul 2007 21:30:55 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>
References: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>
Message-ID: <web-86735412@uni-stuttgart.de>

On Wed, 25 Jul 2007 21:25:36 +0200
  oliver.tomic at matforsk.no wrote:
> 
> Rob and Stefan,
> 
> the problem below described by Rob is exactly what I 
>have experienced. I
> get the same error. I use scipy 0.5.2 and numpy 1.0.3.
> 
> Is there a newer version of scipy than 0.5.2 or do I 
>have to use the svn?
> 
> Here is a file that failed to load with scipy.io:
> (See attached file: brod.mat)
> It was created with Matlab 7.3 with backward 
>compatibility to (-v6).
> 
> Cheers
> Oliver
> 
> 
> 
> scipy-user-bounces at scipy.org wrote on 25.07.2007 
>17:05:28:
> 
>> Stefan,
>>
>> scipy.io.loadmat from scipy 0.5.2 doesn't work on 
>>anything but v4
>> files for me. I generated extremely simple mat files 
>>(one small matrix
>> in 'test.mat') with my matlab 7.3 using the -v6, -v5, 
>>and -v4 options,
>> and scipy complained with this in all but the -v4 case:
>>
>> In [1]: import scipy.io as io
>>
>> In [2]: io.loadmat('test')
>>
> ---------------------------------------------------------------------------
>> exceptions.AttributeError 
>>                           Traceback (most
>> recent call last)
>>
>> c:\temp\<ipython console>
>>
>> C:\Python24\lib\site-packages\scipy\io\mio.py in 
>>loadmat(file_name,
>> mdict, appen
>> dmat, basename, **kwargs)
>>      94     '''
>>      95     MR = mat_reader_factory(file_name, 
>>appendmat, **kwargs)
>> ---> 96     matfile_dict = MR.get_variables()
>>      97     if mdict is not None:
>>      98         mdict.update(matfile_dict)
>>
>> C:\Python24\lib\site-packages\scipy\io\miobase.py in
>> get_variables(self, variable_names)
>>     267             variable_names = [variable_names]
>>     268         self.mat_stream.seek(0)
>> --> 269         mdict = self.file_header()
>>     270         mdict['__globals__'] = []
>>     271         while not self.end_of_stream():
>>
>> C:\Python24\lib\site-packages\scipy\io\mio5.py in 
>>file_header(self)
>>     508         hdict = {}
>>     509         hdr = 
>>self.read_dtype(self.dtypes['file_header'])
>> --> 510         hdict['__header__'] = 
>>hdr['description'].strip('
> \t\n\000')
>>     511         v_major = hdr['version'] >> 8
>>     512         v_minor = hdr['version'] & 0xFF
>>
>> AttributeError: 'numpy.ndarray' object has no attribute 
>>'strip'
>>
>>
>> This is with numpy 1.0.2. I wonder if this is the 
>>problem that Oliver
> has?
>>
>> -Rob
>>
>> On 25/07/07, Stefan van der Walt <stefan at sun.ac.za> 
>>wrote:
>> > Hi Oliver
>> >
>> > On Wed, Jul 25, 2007 at 01:06:18PM +0200, 
>>oliver.tomic at matforsk.no
> wrote:
>> > > Is there now support in scipy.io (scipy 0.52) for 
>>reading .mat-files
> from
>> > > Matlab 7? I undestand that it supports older 
>>versions (v4 and v5).
> Are
>> > > there any plans to support for Matlab 7 .mat-files?
>> >
>> > >From the docstring:
>> >
>> > "v4 (Level 1.0), v6 and v7.1 matfiles are supported."
>> >
>> > Cheers
>> > St?fan
>> > _______________________________________________
>> > 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

No problem with the svn version. It's time for a new
release ...

>>> from scipy import *
>>> io.loadmat('brod.mat')
{'Brod': array([[  1. ,   1. ,   1. , ...,   1. ,   5.6, 
  4.7],
        [  1. ,   1. ,   2. , ...,   1. ,   5.1,   4.1],
        [  1. ,   2. ,   1. , ...,   4.4,   4.8,   3.7],
        ...,
        [ 12. ,   6. ,   2. , ...,   1.8,   2.8,   4.2],
        [ 12. ,   7. ,   1. , ...,   2. ,   4.9,   4.4],
        [ 12. ,   7. ,   2. , ...,   1.3,   5.2,   3.9]]), 
'__version__': '1.0', '__header__': 'MATLAB 5.0 MAT-file, 
Platform: PCWIN, Created on: Wed Jul 25 21:21:07 2007', 
'__globals__': []}
  
Nils


From oliver.tomic at matforsk.no  Wed Jul 25 15:49:52 2007
From: oliver.tomic at matforsk.no (oliver.tomic at matforsk.no)
Date: Wed, 25 Jul 2007 21:49:52 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <web-86735412@uni-stuttgart.de>
Message-ID: <OF1AFB39ED.16C958F7-ONC1257323.006C76D0-C1257323.006D2503@matforsk.no>

Nils, Stefan and Rob,

thank you for your help.



scipy-user-bounces at scipy.org wrote on 25.07.2007 21:30:55:

>
> No problem with the svn version. It's time for a new
> release ...

Yes, it would be great to see a new release soon. The 0.5.2-release may be
a bit seasoned after 8 months.



>
> >>> from scipy import *
> >>> io.loadmat('brod.mat')
> {'Brod': array([[  1. ,   1. ,   1. , ...,   1. ,   5.6,
>   4.7],
>         [  1. ,   1. ,   2. , ...,   1. ,   5.1,   4.1],
>         [  1. ,   2. ,   1. , ...,   4.4,   4.8,   3.7],
>         ...,
>         [ 12. ,   6. ,   2. , ...,   1.8,   2.8,   4.2],
>         [ 12. ,   7. ,   1. , ...,   2. ,   4.9,   4.4],
>         [ 12. ,   7. ,   2. , ...,   1.3,   5.2,   3.9]]),
> '__version__': '1.0', '__header__': 'MATLAB 5.0 MAT-file,
> Platform: PCWIN, Created on: Wed Jul 25 21:21:07 2007',
> '__globals__': []}
>
> Nils
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user


Cheers
Oliver




From s.mientki at ru.nl  Wed Jul 25 17:05:15 2007
From: s.mientki at ru.nl (Stef Mientki)
Date: Wed, 25 Jul 2007 23:05:15 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <web-86735412@uni-stuttgart.de>
References: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>
	<web-86735412@uni-stuttgart.de>
Message-ID: <46A7BB0B.3060709@ru.nl>


> No problem with the svn version. It's time for a new
> release ...
>
>   
this would be very welcome, ...
... I don't understand svn
... I didn't even knew there was a "loadmat"
I translated my scripts by hand :-(

I think it would be very nice for those who consider switching from MatLab,
that the first line they read is "and there is a M-file conversion tool".
When I was looking for something different from MatLab,
I found 2 candidates: SciPy and SciLab.
Scipy had the advantage of better embedding (SciLab was just beginning),
which was important for me (but not for most people coming from MatLab).
SciLab had the advantage of a good M-file converter (I couldn't find it 
in SciPy),
which is important for most people coming from MatLab.

btw: I'm still very interesting in the "good" MatLab converter,
(where and how to install )
because it gives me more arguments to convice my colleagues.

cheers,
Stef Mientki



From david at ar.media.kyoto-u.ac.jp  Wed Jul 25 22:11:07 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Thu, 26 Jul 2007 11:11:07 +0900
Subject: [SciPy-user] fitting mixed gaussians
In-Reply-To: <d7b942cd0707250846q752c842k1f30b3fafa1d34e@mail.gmail.com>
References: <d7b942cd0707250846q752c842k1f30b3fafa1d34e@mail.gmail.com>
Message-ID: <46A802BB.8000707@ar.media.kyoto-u.ac.jp>

Emanuele Zattin wrote:
> Hello,
> i'm trying to perform mixed-gaussian fitting on some small greyscale
> images (5x5 pixels, maximum 9x9). right now i'm using a leastsq
> approach but it tends to get pretty slow and unreliable when it comes
> to fit 4-5 2D gaussians even in such a small domain.
> I've noticed PyEM and i have the feeling it might be useful for me,
> but here i'm dealing with pixel intensities, not distributions. Any
> hint is welcome!
>
>   
If I understand your problem well, you have one image and wants to fit 
one, possibly several Gaussian on it. pyem won't help you for that, or 
more exactly, would be both overkill and gives really poor performances 
(even for only one gaussian, if it is not centered, with a few tens 
samples, it would be pretty bad, I think).

David




From david at ar.media.kyoto-u.ac.jp  Wed Jul 25 22:19:27 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Thu, 26 Jul 2007 11:19:27 +0900
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <46A7BB0B.3060709@ru.nl>
References: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>	<web-86735412@uni-stuttgart.de>
	<46A7BB0B.3060709@ru.nl>
Message-ID: <46A804AF.3020504@ar.media.kyoto-u.ac.jp>

Stef Mientki wrote:
>> No problem with the svn version. It's time for a new
>> release ...
>>
>>   
> this would be very welcome, ...
> ... I don't understand svn
> ... I didn't even knew there was a "loadmat"
> I translated my scripts by hand :-(
>
> I think it would be very nice for those who consider switching from MatLab,
> that the first line they read is "and there is a M-file conversion tool".
I have the feeling that there is a confusion between mat files and m 
scripts. mat files are just a file format used by matlab to store data. 
m scripts are matlab code, which scipy has no way to understand.

There is no such thing as a scipy function which loads a matlab script 
and converts it to scipy. This would be impossible, or at least 
extremely difficult, since matlab and python are different languages 
with different semantics.

Scilab (and octave) were created as programming environments with an 
explicit support for matlab syntax and semantics. scipy is a framework 
for scientific computing in python. This may be seen as a weakness if 
you are coming from matlab, but for me, this is the whole point, since I 
consider matlab as a pretty bad environment for scientific computing.

If you want to be able to use matlab scripts with minimum fuss, I 
strongly advise you not to use scipy. If you are willing to spend time 
to convert them, to benefit from the advantages of scipy (a real 
programming language with advanced features, a full C Api), then use scipy.

David


From prabhu at aero.iitb.ac.in  Tue Jul 24 17:25:37 2007
From: prabhu at aero.iitb.ac.in (Prabhu Ramachandran)
Date: Wed, 25 Jul 2007 02:55:37 +0530
Subject: [SciPy-user] BoF on 3D visualization
Message-ID: <18086.28241.432931.478493@gargle.gargle.HOWL>

Hello,

Gael Varoquaux and myself are planning on holding a "Pythonic 3D
visualization" BoF at SciPy07 on Thursday evening.  We'd like to know
if any of you would be interested in this.  If you are, please do let
us know.  Please also let us know if you have anything in particular
you'd like to discuss.

Thanks.

Cheers,
-- 
Prabhu Ramachandran                 http://www.aero.iitb.ac.in/~prabhu


From fperez.net at gmail.com  Thu Jul 26 01:22:15 2007
From: fperez.net at gmail.com (Fernando Perez)
Date: Wed, 25 Jul 2007 23:22:15 -0600
Subject: [SciPy-user] BoF on 3D visualization
In-Reply-To: <18086.28241.432931.478493@gargle.gargle.HOWL>
References: <18086.28241.432931.478493@gargle.gargle.HOWL>
Message-ID: <db6b5ecc0707252222s1f1a2790sc76d09996f19d544@mail.gmail.com>

On 7/24/07, Prabhu Ramachandran <prabhu at aero.iitb.ac.in> wrote:
> Hello,
>
> Gael Varoquaux and myself are planning on holding a "Pythonic 3D
> visualization" BoF at SciPy07 on Thursday evening.  We'd like to know
> if any of you would be interested in this.  If you are, please do let
> us know.  Please also let us know if you have anything in particular
> you'd like to discuss.

Count me in.

Cheers,

f


From s.mientki at ru.nl  Thu Jul 26 04:33:52 2007
From: s.mientki at ru.nl (Stef Mientki)
Date: Thu, 26 Jul 2007 10:33:52 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <46A804AF.3020504@ar.media.kyoto-u.ac.jp>
References: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>
	<web-86735412@uni-stuttgart.de> <46A7BB0B.3060709@ru.nl>
	<46A804AF.3020504@ar.media.kyoto-u.ac.jp>
Message-ID: <46A85C70.9020506@ru.nl>




>>>   
>>>       
>> this would be very welcome, ...
>> ... I don't understand svn
>> ... I didn't even knew there was a "loadmat"
>> I translated my scripts by hand :-(
>>
>> I think it would be very nice for those who consider switching from MatLab,
>> that the first line they read is "and there is a M-file conversion tool".
>>     
> I have the feeling that there is a confusion between mat files and m 
> scripts. mat files are just a file format used by matlab to store data. 
> m scripts are matlab code, which scipy has no way to understand.
>   
thanks David,
you're completely right, that was the mis-understanding.
So happily the manual translation wasn't a waste of time after all ;-)

cheers,
Stef Mientki



From nwagner at iam.uni-stuttgart.de  Thu Jul 26 06:58:03 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Thu, 26 Jul 2007 12:58:03 +0200
Subject: [SciPy-user] argmin and optimization
Message-ID: <46A87E3B.6020206@iam.uni-stuttgart.de>

Hi all,

I would like to implement an algorithm described by Golub and Liao.
Get the source from here: http://citeseer.ist.psu.edu/638954.html

The central question is the following. How do I implement the operator
P_\Omega(\cdot)  in scipy ?

The definition is as follows (\LaTex notation)

P_\Omega(y) = \textrm{argmin}\limits_{x \in \Omega} \|x-y\|_2 \forall \,
y \in \mathds{R}^n

\Omega = \{ x \in \mathds{R}^n | <x,x> \le 1 \}.

I have attached a short script which implements the rest of the algorithm.
Any pointer how to implement P_\Omega(y) in scipy would be appreciated.

Thanks in advance
                                          Nils




From nwagner at iam.uni-stuttgart.de  Thu Jul 26 07:02:18 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Thu, 26 Jul 2007 13:02:18 +0200
Subject: [SciPy-user] argmin and optimization
Message-ID: <46A87F3A.9040900@iam.uni-stuttgart.de>

I missed out the attachment.

Nils

-------------- next part --------------
A non-text attachment was scrubbed...
Name: test_golub.py
Type: text/x-python
Size: 1036 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070726/37eef1d8/attachment.py>

From openopt at ukr.net  Thu Jul 26 07:21:21 2007
From: openopt at ukr.net (dmitrey)
Date: Thu, 26 Jul 2007 14:21:21 +0300
Subject: [SciPy-user] argmin and optimization
In-Reply-To: <46A87F3A.9040900@iam.uni-stuttgart.de>
References: <46A87F3A.9040900@iam.uni-stuttgart.de>
Message-ID: <46A883B1.3090302@ukr.net>

So your Omega is just all x: (x,x)<1?
(Can you attach non-Latex notation elseware?)
Why can't you just set P_Omega(y) = {
y, for ||y|| <=1
y/||y||, for ||y|| > 1
}
?

Regards, D.


From nwagner at iam.uni-stuttgart.de  Thu Jul 26 07:29:44 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Thu, 26 Jul 2007 13:29:44 +0200
Subject: [SciPy-user] argmin and optimization
In-Reply-To: <46A883B1.3090302@ukr.net>
References: <46A87F3A.9040900@iam.uni-stuttgart.de> <46A883B1.3090302@ukr.net>
Message-ID: <46A885A8.5010400@iam.uni-stuttgart.de>

dmitrey wrote:
> So your Omega is just all x: (x,x)<1?
>   
(x,x) <= 1 ("closed ball constraint")

> (Can you attach non-Latex notation elseware?)
>   
I can't see what you mean here.
> Why can't you just set P_Omega(y) = {
> y, for ||y|| <=1
> y/||y||, for ||y|| > 1
> }
> ?
>   
I don't see your point.

Nils
 


From openopt at ukr.net  Thu Jul 26 07:49:02 2007
From: openopt at ukr.net (dmitrey)
Date: Thu, 26 Jul 2007 14:49:02 +0300
Subject: [SciPy-user] argmin and optimization
In-Reply-To: <46A885A8.5010400@iam.uni-stuttgart.de>
References: <46A87F3A.9040900@iam.uni-stuttgart.de> <46A883B1.3090302@ukr.net>
	<46A885A8.5010400@iam.uni-stuttgart.de>
Message-ID: <46A88A2E.8070406@ukr.net>

Nils Wagner wrote:
> dmitrey wrote:
>   
>> So your Omega is just all x: (x,x)<1?
>>   
>>     
> (x,x) <= 1 ("closed ball constraint")
>
>   
>> (Can you attach non-Latex notation elseware?)
>>   
>>     
> I can't see what you mean here.
>   
I meant couldn't you describe what does Omega and P_Omega mean w/o Latex?
And/or describe howto see Latex strings in Linux?
>> Why can't you just set P_Omega(y) = {
>> y, for ||y|| <=1
>> y/||y||, for ||y|| > 1
>> }
>> ?
>>   
>>     
> I don't see your point.
>
> Nils
>   
I meant is this true that the func P_Omega will yield the result you intend?
from numpy.linalg import norm
def P_Omega(y):
  n_y = norm(y)
  if n_y <= 1: return y
  else: return y/ny
HTH, D
>  
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>
>   



From nwagner at iam.uni-stuttgart.de  Thu Jul 26 08:05:51 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Thu, 26 Jul 2007 14:05:51 +0200
Subject: [SciPy-user] argmin and optimization
In-Reply-To: <46A88A2E.8070406@ukr.net>
References: <46A87F3A.9040900@iam.uni-stuttgart.de>
	<46A883B1.3090302@ukr.net>	<46A885A8.5010400@iam.uni-stuttgart.de>
	<46A88A2E.8070406@ukr.net>
Message-ID: <46A88E1F.5070103@iam.uni-stuttgart.de>

dmitrey wrote:
> Nils Wagner wrote:
>   
>> dmitrey wrote:
>>   
>>     
>>> So your Omega is just all x: (x,x)<1?
>>>   
>>>     
>>>       
>> (x,x) <= 1 ("closed ball constraint")
>>
>>   
>>     
>>> (Can you attach non-Latex notation elseware?)
>>>   
>>>     
>>>       
>> I can't see what you mean here.
>>   
>>     
> I meant couldn't you describe what does Omega and P_Omega mean w/o Latex?
> And/or describe howto see Latex strings in Linux?
>   
>>> Why can't you just set P_Omega(y) = {
>>> y, for ||y|| <=1
>>> y/||y||, for ||y|| > 1
>>> }
>>> ?
>>>   
>>>     
>>>       
>> I don't see your point.
>>
>> Nils
>>   
>>     
> I meant is this true that the func P_Omega will yield the result you intend?
> from numpy.linalg import norm
> def P_Omega(y):
>   n_y = norm(y)
>   if n_y <= 1: return y
>   else: return y/ny
> HTH, D
>   
>>  
>> _______________________________________________
>> 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
>   
Dmitrey,

Thank you very much for your help. It works fine for me.
I have attached the script for the sake of completeness.

Cheers,
                    Nils

-------------- next part --------------
A non-text attachment was scrubbed...
Name: test_golub.py
Type: text/x-python
Size: 1151 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070726/341c1f05/attachment.py>

From openopt at ukr.net  Thu Jul 26 08:12:00 2007
From: openopt at ukr.net (dmitrey)
Date: Thu, 26 Jul 2007 15:12:00 +0300
Subject: [SciPy-user] argmin and optimization
In-Reply-To: <46A88E1F.5070103@iam.uni-stuttgart.de>
References: <46A87F3A.9040900@iam.uni-stuttgart.de>	<46A883B1.3090302@ukr.net>	<46A885A8.5010400@iam.uni-stuttgart.de>	<46A88A2E.8070406@ukr.net>
	<46A88E1F.5070103@iam.uni-stuttgart.de>
Message-ID: <46A88F90.2050803@ukr.net>

if you use python 2.5 you can use more short:
def P_Omega(y):
    from numpy.linalg import norm
    n_y = norm(y)
    return y if n_y <= 1 else y/n_y
HTH, D.

Nils Wagner wrote:
> dmitrey wrote:
>   
>> Nils Wagner wrote:
>>   
>>     
>>> dmitrey wrote:
>>>   
>>>     
>>>       
>>>> So your Omega is just all x: (x,x)<1?
>>>>   
>>>>     
>>>>       
>>>>         
>>> (x,x) <= 1 ("closed ball constraint")
>>>
>>>   
>>>     
>>>       
>>>> (Can you attach non-Latex notation elseware?)
>>>>   
>>>>     
>>>>       
>>>>         
>>> I can't see what you mean here.
>>>   
>>>     
>>>       
>> I meant couldn't you describe what does Omega and P_Omega mean w/o Latex?
>> And/or describe howto see Latex strings in Linux?
>>   
>>     
>>>> Why can't you just set P_Omega(y) = {
>>>> y, for ||y|| <=1
>>>> y/||y||, for ||y|| > 1
>>>> }
>>>> ?
>>>>   
>>>>     
>>>>       
>>>>         
>>> I don't see your point.
>>>
>>> Nils
>>>   
>>>     
>>>       
>> I meant is this true that the func P_Omega will yield the result you intend?
>> from numpy.linalg import norm
>> def P_Omega(y):
>>   n_y = norm(y)
>>   if n_y <= 1: return y
>>   else: return y/ny
>> HTH, D
>>   
>>     
>>>  
>>> _______________________________________________
>>> 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
>>   
>>     
> Dmitrey,
>
> Thank you very much for your help. It works fine for me.
> I have attached the script for the sake of completeness.
>
> Cheers,
>                     Nils
>
>   
> ------------------------------------------------------------------------
>
> from scipy import *
> from pylab import plot, show
> #
> # Gene H. Golub and Li-Zhi Liao
> # Continous methods for extreme and interior eigenvalue problems
> # Linear Algebra and its Applications Vol. 415 (2006) pp. 31-51 
> #
> def merit(x):
>     """ Merit function equation (3.1) """
>     return dot(x,dot(A,x))-c*dot(x,x)
>
> def meritp(x):
>     """ Gradient of the merit function """
>     return 2*dot(A,x)-2*c*x
>
> def P(y):
>     """ Projection onto \Omega defined as 
>     P_\Omega(y) = \argmin\limits_{x \in \Omega} \|x-y\|_2 """
>     n_y = linalg.norm(y)
>     if n_y <=1: return y
>     else: return y/n_y
>
> def dyn(x,t):
>     """ Dynamical system equation (3.2) """
>     tmp = zeros(n,float)
>     tmp = -(x-P(x-meritp(x)))
>     return tmp
>
> n = 6
> L = diag(r_[array(([-2.,-0.2])),zeros(2),ones(n-4)]) # Example 1 p. 46
> B = rand(n,n)
> Q,R = linalg.qr(B)
> A = dot(Q.T,dot(L,Q))
> A = 0.5*(A+A.T)
> c = linalg.norm(A,1)+1. # default value in our numerical computation
> print c
>
> x_0 =-ones(n) # Starting point
> T = linspace(0.,30.,100)
> x = integrate.odeint(dyn,x_0,T)
> plot(T,x[:,0],T,x[:,1],T,x[:,2],T,x[:,3],T,x[:,4],T,x[:,5])
> print dot(x[-1,:],dot(A,x[-1,:])),dot(x[-1,:],x[-1,:])
> show()
>   
> ------------------------------------------------------------------------
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   



From aisaac at american.edu  Thu Jul 26 08:18:53 2007
From: aisaac at american.edu (Alan G Isaac)
Date: Thu, 26 Jul 2007 08:18:53 -0400
Subject: [SciPy-user] argmin and optimization
In-Reply-To: <46A87E3B.6020206@iam.uni-stuttgart.de>
References: <46A87E3B.6020206@iam.uni-stuttgart.de>
Message-ID: <Mahogany-0.67.0-968-20070726-081853.00@american.edu>

On Thu, 26 Jul 2007, Nils Wagner apparently wrote:
> I would like to implement an algorithm described by Golub and Liao. 
> Get the source from here: http://citeseer.ist.psu.edu/638954.html 

> The central question is the following. How do I implement the operator 
> P_\Omega(\cdot)  in scipy ? 

> The definition is as follows (\LaTex notation)

> P_\Omega(y) = \textrm{argmin}\limits_{x \in \Omega} \|x-y\|_2 \forall \, 
> y \in \mathds{R}^n 

> \Omega = \{ x \in \mathds{R}^n | <x,x> \le 1 \}. 


It looks like you are saying:
get as close to y as possible without leaving the unit 
sphere.  Is that right?  This is just y if y is in the
unit sphere and it is y normalized to unit length if
y is outside the unit sphere.

Please post the code when you finish the algorithm.

Cheers,
Alan Isaac




From massimo.sandal at unibo.it  Thu Jul 26 08:40:20 2007
From: massimo.sandal at unibo.it (massimo sandal)
Date: Thu, 26 Jul 2007 14:40:20 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <46A7BB0B.3060709@ru.nl>
References: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>	<web-86735412@uni-stuttgart.de>
	<46A7BB0B.3060709@ru.nl>
Message-ID: <46A89634.20206@unibo.it>

Stef Mientki ha scritto:

> this would be very welcome, ...
> ... I don't understand svn

I've never used it for scipy, but SVN (SubVersion) for an end user that 
wants the latest-and-greatest is pretty easy to use. Just:
- install svn on your computer
- at a console, type "svn checkout http://url-of-svn-server-you-want"
and you'll obtain a directory with the (usually installable) source code 
you need.

Then (if everything's been done right) just install like you would 
install a tarball.

m.

-- 
Massimo Sandal
University of Bologna
Department of Biochemistry "G.Moruzzi"

snail mail:
Via Irnerio 48, 40126 Bologna, Italy

email:
massimo.sandal at unibo.it

tel: +39-051-2094388
fax: +39-051-2094387
-------------- next part --------------
A non-text attachment was scrubbed...
Name: massimo.sandal.vcf
Type: text/x-vcard
Size: 274 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070726/f38d85b9/attachment.vcf>

From nwagner at iam.uni-stuttgart.de  Thu Jul 26 08:37:25 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Thu, 26 Jul 2007 14:37:25 +0200
Subject: [SciPy-user] argmin and optimization
In-Reply-To: <Mahogany-0.67.0-968-20070726-081853.00@american.edu>
References: <46A87E3B.6020206@iam.uni-stuttgart.de>
	<Mahogany-0.67.0-968-20070726-081853.00@american.edu>
Message-ID: <46A89585.80506@iam.uni-stuttgart.de>

Alan G Isaac wrote:
> On Thu, 26 Jul 2007, Nils Wagner apparently wrote:
>   
>> I would like to implement an algorithm described by Golub and Liao. 
>> Get the source from here: http://citeseer.ist.psu.edu/638954.html 
>>     
>
>   
>> The central question is the following. How do I implement the operator 
>> P_\Omega(\cdot)  in scipy ? 
>>     
>
>   
>> The definition is as follows (\LaTex notation)
>>     
>
>   
>> P_\Omega(y) = \textrm{argmin}\limits_{x \in \Omega} \|x-y\|_2 \forall \, 
>> y \in \mathds{R}^n 
>>     
>
>   
>> \Omega = \{ x \in \mathds{R}^n | <x,x> \le 1 \}. 
>>     
>
>
> It looks like you are saying:
> get as close to y as possible without leaving the unit 
> sphere.  Is that right?  This is just y if y is in the
> unit sphere and it is y normalized to unit length if
> y is outside the unit sphere.
>
> Please post the code when you finish the algorithm.
>
> Cheers,
> Alan Isaac
>
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   
 Alan,

Your explanation is right. I have submitted the script in my previous email.

Cheers,
                    Nils



From s.mientki at mailbox.kun.nl  Thu Jul 26 09:39:59 2007
From: s.mientki at mailbox.kun.nl (stef mientki)
Date: Thu, 26 Jul 2007 15:39:59 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <46A89634.20206@unibo.it>
References: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>	<web-86735412@uni-stuttgart.de>	<46A7BB0B.3060709@ru.nl>
	<46A89634.20206@unibo.it>
Message-ID: <46A8A42F.5050002@gmail.com>

massimo sandal wrote:
> Stef Mientki ha scritto:
>
>> this would be very welcome, ...
>> ... I don't understand svn
>
> I've never used it for scipy, but SVN (SubVersion) for an end user 
> that wants the latest-and-greatest is pretty easy to use. Just:
> - install svn on your computer
> - at a console, type "svn checkout http://url-of-svn-server-you-want"
> and you'll obtain a directory with the (usually installable) source 
> code you need.
>
> Then (if everything's been done right) just install like you would 
> install a tarball.
thanks Massimo,
but I'm just a computer user,
so do you really think I know what a tarball is ;-)

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



From massimo.sandal at unibo.it  Thu Jul 26 09:45:54 2007
From: massimo.sandal at unibo.it (massimo sandal)
Date: Thu, 26 Jul 2007 15:45:54 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <46A8A42F.5050002@gmail.com>
References: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>	<web-86735412@uni-stuttgart.de>	<46A7BB0B.3060709@ru.nl>	<46A89634.20206@unibo.it>
	<46A8A42F.5050002@gmail.com>
Message-ID: <46A8A592.3010701@unibo.it>

stef mientki ha scritto:
>> Then (if everything's been done right) just install like you would 
>> install a tarball.
> thanks Massimo,
> but I'm just a computer user,
> so do you really think I know what a tarball is ;-)

Sorry, I'm always assuming everyone here's using Linux/Unix. :)

However you're also a developer, otherwise what would you do on a 
scientific programming library mailing list? Don't hide from your 
responsibilities! :P

m.

-- 
Massimo Sandal
University of Bologna
Department of Biochemistry "G.Moruzzi"

snail mail:
Via Irnerio 48, 40126 Bologna, Italy

email:
massimo.sandal at unibo.it

tel: +39-051-2094388
fax: +39-051-2094387
-------------- next part --------------
A non-text attachment was scrubbed...
Name: massimo.sandal.vcf
Type: text/x-vcard
Size: 274 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070726/14a11511/attachment.vcf>

From david at ar.media.kyoto-u.ac.jp  Thu Jul 26 09:37:48 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Thu, 26 Jul 2007 22:37:48 +0900
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <46A8A42F.5050002@gmail.com>
References: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>	<web-86735412@uni-stuttgart.de>	<46A7BB0B.3060709@ru.nl>	<46A89634.20206@unibo.it>
	<46A8A42F.5050002@gmail.com>
Message-ID: <46A8A3AC.5060405@ar.media.kyoto-u.ac.jp>

stef mientki wrote:
> massimo sandal wrote:
>   
>> Stef Mientki ha scritto:
>>
>>     
>>> this would be very welcome, ...
>>> ... I don't understand svn
>>>       
>> I've never used it for scipy, but SVN (SubVersion) for an end user 
>> that wants the latest-and-greatest is pretty easy to use. Just:
>> - install svn on your computer
>> - at a console, type "svn checkout http://url-of-svn-server-you-want"
>> and you'll obtain a directory with the (usually installable) source 
>> code you need.
>>
>> Then (if everything's been done right) just install like you would 
>> install a tarball.
>>     
> thanks Massimo,
> but I'm just a computer user,
> so do you really think I know what a tarball is ;-)
>
>   
On windows, you can use TortoiseSVN, which is a graphical interface to 
subversion. Once you fetch the sources with it, installing scipy is 
exactly like installing from the official sources.

David


From jeff.lyon at cox.net  Thu Jul 26 10:31:39 2007
From: jeff.lyon at cox.net (Jeff Lyon)
Date: Thu, 26 Jul 2007 07:31:39 -0700
Subject: [SciPy-user] filtering without phase shift
In-Reply-To: <ce557a360707250836jf90fd93ya3fdd5d4a8d85659@mail.gmail.com>
References: <c5b438120707232027w32c8a564v346f9292501616e8@mail.gmail.com>
	<1185261264.5678.9.camel@pc1.cole.uklinux.net>
	<46A5B74C.2060304@ru.nl>
	<ce557a360707250836jf90fd93ya3fdd5d4a8d85659@mail.gmail.com>
Message-ID: <00C8CDD1-F223-471D-87F2-DB1216611A6A@cox.net>

Another way to filter your data is with an finite impulse response  
(FIR) filter. One of the nice properties of FIR filters is that they  
can have linear phase, or constant group delay. This means that all  
the frequencies in the filtered signal are delayed uniformly, and you  
can simply subtract out the delay.

On Jul 25, 2007, at 8:36 AM, Anne Archibald wrote:

> On 24/07/07, Stef Mientki <s.mientki at ru.nl> wrote:
>>
>>
>> Bryan Cole wrote:
>>> The easiest way to do non-phase-shifted filtering is to use FFTs  
>>> i.e.
>>>
>>> FFT -> Low Pass -> iFFT
>>>
>>> The only reason to move to time-domain filters is if you must do
>>> filtering in real-time where the complete dataset is (yet)  
>>> unknown and
>>> you must obey the laws of causality.
>>>
>>> Using a frequency domain filter, you can have whatever cutoff you  
>>> like.
>>>
>> I'm a bit rusty, but isn't there a real danger by filtering in the
>> frequency domain,
>> creating a rectangle cutoff in the frequency domain,
>> leads to an infinite ringing signal in the time-domain.
>
> There is a concern with ringing if you just use a brick-wall cutoff
> (though often with an FFT you have so much frequency resolution the
> ringing isn't actually a problem). The easiest way to do this is to
> use a softer edge - there are standard filter shapes (Butterworth and
> so on) which will minimize ringing, but any relatively smooth falloff
> will help.
>
> It sounds like this is not quite your issue, but if what you need is a
> filter that preserves peak height, Kalman filtering is what you need.
>
> Anne
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user



From rmay at ou.edu  Thu Jul 26 11:05:44 2007
From: rmay at ou.edu (Ryan May)
Date: Thu, 26 Jul 2007 10:05:44 -0500
Subject: [SciPy-user] BoF on 3D visualization
In-Reply-To: <18086.28241.432931.478493@gargle.gargle.HOWL>
References: <18086.28241.432931.478493@gargle.gargle.HOWL>
Message-ID: <46A8B848.2040000@ou.edu>

Prabhu Ramachandran wrote:
> Hello,
> 
> Gael Varoquaux and myself are planning on holding a "Pythonic 3D
> visualization" BoF at SciPy07 on Thursday evening.  We'd like to know
> if any of you would be interested in this.  If you are, please do let
> us know.  Please also let us know if you have anything in particular
> you'd like to discuss.
> 

Count me in.

Ryan

-- 
Ryan May
Graduate Research Assistant
School of Meteorology
University of Oklahoma


From s.mientki at mailbox.kun.nl  Thu Jul 26 11:48:13 2007
From: s.mientki at mailbox.kun.nl (stef mientki)
Date: Thu, 26 Jul 2007 17:48:13 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <46A8A592.3010701@unibo.it>
References: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>	<web-86735412@uni-stuttgart.de>	<46A7BB0B.3060709@ru.nl>	<46A89634.20206@unibo.it>	<46A8A42F.5050002@gmail.com>
	<46A8A592.3010701@unibo.it>
Message-ID: <46A8C23D.6000101@gmail.com>


>
> Sorry, I'm always assuming everyone here's using Linux/Unix. :)
>
> However you're also a developer, otherwise what would you do on a 
> scientific programming library mailing list? Don't hide from your 
> responsibilities! :P
Well I already feel like a Don Quichotte,
trying to introduce Python (Scipy) into an organization,
as a replacement for MatLab and LabView.

for the record
compagny  (university, 20,000 people, excluding students) :
OS 100%  Windows
scientific computing 80% MatLab, 15% LabView, 5% other

country
OS 95% Windows, 4%Linux

personal:
I don't know anyone who is running Linux ;-)

I think that's one of the differences between the "rich world"  and the 
"poor world"

cheers,
Stef



From Alexander.Dietz at astro.cf.ac.uk  Thu Jul 26 12:15:25 2007
From: Alexander.Dietz at astro.cf.ac.uk (Alexander Dietz)
Date: Thu, 26 Jul 2007 17:15:25 +0100
Subject: [SciPy-user] Question to functions
Message-ID: <9cf809a00707260915i29520350gf8384f78f0ce6786@mail.gmail.com>

Hi,

I wanted to use a function to calculate the normal density (aka Gauss or
Normal) by doing something like:

y=scipy.???( x, m, s)

with x representing my variable, m the mean of this Normal distribution and
s the sigma value. I was looking now for 15 minutes in the numpy/scipy
documentation, but all I found was a vast list of functions with
scipy.special, but I did not found any function to just compute the normal
probability.
My question: in what package can I found this function with what name? If
there are dozen of functions implemented in scipy I never heard of, then
there must be the normal distribution as well.


Cheers
  Alex
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070726/9cb5caad/attachment.html>

From robert.kern at gmail.com  Thu Jul 26 14:47:11 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Thu, 26 Jul 2007 13:47:11 -0500
Subject: [SciPy-user] Question to functions
In-Reply-To: <9cf809a00707260915i29520350gf8384f78f0ce6786@mail.gmail.com>
References: <9cf809a00707260915i29520350gf8384f78f0ce6786@mail.gmail.com>
Message-ID: <46A8EC2F.6020405@gmail.com>

Alexander Dietz wrote:
> Hi,
> 
> I wanted to use a function to calculate the normal density (aka Gauss or
> Normal) by doing something like:
> 
> y=scipy.???( x, m, s)
> 
> with x representing my variable, m the mean of this Normal distribution
> and s the sigma value. I was looking now for 15 minutes in the
> numpy/scipy documentation, but all I found was a vast list of functions
> with scipy.special, but I did not found any function to just compute the
> normal probability.
> My question: in what package can I found this function with what name?
> If there are dozen of functions implemented in scipy I never heard of,
> then there must be the normal distribution as well.

scipy.stats.norm.pdf(x, m, s)

-- 
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 Karl.Young at ucsf.edu  Thu Jul 26 13:50:21 2007
From: Karl.Young at ucsf.edu (Karl Young)
Date: Thu, 26 Jul 2007 10:50:21 -0700
Subject: [SciPy-user] BoF on 3D visualization
In-Reply-To: <18086.28241.432931.478493@gargle.gargle.HOWL>
References: <18086.28241.432931.478493@gargle.gargle.HOWL>
Message-ID: <46A8DEDD.2060006@ucsf.edu>

Prabhu Ramachandran wrote:

>Hello,
>
>Gael Varoquaux and myself are planning on holding a "Pythonic 3D
>visualization" BoF at SciPy07 on Thursday evening.  We'd like to know
>if any of you would be interested in this.  If you are, please do let
>us know.  Please also let us know if you have anything in particular
>you'd like to discuss.
>
>Thanks.
>
>Cheers,
>  
>

I'll definitely show; issues that come up in 3D visualization of 
neuroimages are tops on my list but I'm sure any discussion will be of 
interest (I'm a real novice re. 3D visualization so I'm not sure if I 
contribute a whole lot of interest) 

-- KY

Karl Young
Center for Imaging of Neurodegenerative Diseases, UCSF          
VA Medical Center (114M)              Phone:  (415) 221-4810 x3114  lab        
4150 Clement Street                   FAX:    (415) 668-2864
San Francisco, CA 94121               Email:  karl young at ucsf edu


From fredmfp at gmail.com  Thu Jul 26 16:05:42 2007
From: fredmfp at gmail.com (fred)
Date: Thu, 26 Jul 2007 22:05:42 +0200
Subject: [SciPy-user] using imread...
Message-ID: <46A8FE96.40404@gmail.com>

Hi all,

I have to read some data got from a scan, ie data are printed on paper, 
then scanned.

Before to begin, I want to test some stuff, of course, ie make some test
on an image I know, and compare the results with the original. 

1) I have a data (scipy array) that I know.
2) I create a png file from this array using matplotlib, mayavi, etc.
3) I read this png file with imread().
4) This creates a scipy array I read and dump to a png file
with matplotlib/mayavi, following the cookbook.

Obviously,  the png created in 2) and the png created in 4) are not the 
same.
Scalar values of the png created in 4) ranges from 0 to 1, that's ok,
but even rescaled, I don't get the same values as the original.  

So, what am I doing wrong ?

Any pointer, url, are welcome.

TIA.


Cheers,

-- 
http://scipy.org/FredericPetit



From vaftrudner at gmail.com  Thu Jul 26 16:41:29 2007
From: vaftrudner at gmail.com (Martin Blom)
Date: Thu, 26 Jul 2007 15:41:29 -0500
Subject: [SciPy-user] Memory problems with stats.mannwhitneyu
Message-ID: <dc90fba40707261341u42fbb603x9171c869768185f2@mail.gmail.com>

I need to do lots and lots of mannwhitneyu-tests and my programs keep
running out of memory so something is wrong somewhere. When I run, for
example,

from scipy.stats import mannwhitneyu

x = [2,4,8,3]
y = [1,3,5,6]

while True:
    u,p = mannwhitneyu(x,y)

and run top to look at the memory used by python it just keeps increasing
until I run out. Does anyone who understands the inner workings of the stats
module know what's happening? Or is it that my little loop doesn't release
memory in some way? I've been using mannwhitneyu for some time now, and I
haven't noticed the problem until now, which might mean that I've had enough
memory to get by until now or that there is a recent error, which seems
unlikely when I look at the changelog.

yours confusedly,
Martin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070726/7a42daa9/attachment.html>

From ctrachte at gmail.com  Thu Jul 26 16:54:19 2007
From: ctrachte at gmail.com (Carl Trachte)
Date: Thu, 26 Jul 2007 13:54:19 -0700
Subject: [SciPy-user] BoF on 3D visualization
In-Reply-To: <46A8DEDD.2060006@ucsf.edu>
References: <18086.28241.432931.478493@gargle.gargle.HOWL>
	<46A8DEDD.2060006@ucsf.edu>
Message-ID: <426ada670707261354w216b195ctfa80ae3a334d11ad@mail.gmail.com>

I am mostly ignorant with regard to the BoF subject, but I need to get
smarter.
I will attend.

On 7/26/07, Karl Young <Karl.Young at ucsf.edu> wrote:
>
> Prabhu Ramachandran wrote:
>
> >Hello,
> >
> >Gael Varoquaux and myself are planning on holding a "Pythonic 3D
> >visualization" BoF at SciPy07 on Thursday evening.  We'd like to know
> >if any of you would be interested in this.  If you are, please do let
> >us know.  Please also let us know if you have anything in particular
> >you'd like to discuss.
> >
> >Thanks.
> >
> >Cheers,
> >
> >
>
> I'll definitely show; issues that come up in 3D visualization of
> neuroimages are tops on my list but I'm sure any discussion will be of
> interest (I'm a real novice re. 3D visualization so I'm not sure if I
> contribute a whole lot of interest)
>
> -- KY
>
> Karl Young
> Center for Imaging of Neurodegenerative Diseases, UCSF
> VA Medical Center (114M)              Phone:  (415) 221-4810 x3114  lab
> 4150 Clement Street                   FAX:    (415) 668-2864
> San Francisco, CA 94121               Email:  karl young at ucsf edu
> _______________________________________________
> 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/20070726/68629935/attachment.html>

From pgmdevlist at gmail.com  Thu Jul 26 16:54:16 2007
From: pgmdevlist at gmail.com (Pierre GM)
Date: Thu, 26 Jul 2007 16:54:16 -0400
Subject: [SciPy-user] Memory problems with stats.mannwhitneyu
In-Reply-To: <dc90fba40707261341u42fbb603x9171c869768185f2@mail.gmail.com>
References: <dc90fba40707261341u42fbb603x9171c869768185f2@mail.gmail.com>
Message-ID: <200707261654.17467.pgmdevlist@gmail.com>


A silly question, but how do you exit this loop ? 

> while True:
>     u,p = mannwhitneyu(x,y)

Unless you omitted some important part of your script, this loop will keep on 
running.


From vaftrudner at gmail.com  Thu Jul 26 17:15:25 2007
From: vaftrudner at gmail.com (Martin Blom)
Date: Thu, 26 Jul 2007 16:15:25 -0500
Subject: [SciPy-user] Memory problems with stats.mannwhitneyu
In-Reply-To: <200707261654.17467.pgmdevlist@gmail.com>
References: <dc90fba40707261341u42fbb603x9171c869768185f2@mail.gmail.com>
	<200707261654.17467.pgmdevlist@gmail.com>
Message-ID: <dc90fba40707261415v68ba0bfaof52d5b7594d3d790@mail.gmail.com>

Well, yes, it does run forever as it stands. I don't see how that would make
it eat all my memory though. The problem is not that it uses memory per se,
it is that the amount of memory used continues to increase until I run out,
and I don't see why that is.

The below script is just the last stage of my attempt at debugging. My real
program is much longer and trickier and does 18 million slightly different
u-tests, but the memory issue is the same for both, and I didn't want to
complicate the mail with a lot of unnecessary stuff.

thank you
Martin

2007/7/26, Pierre GM <pgmdevlist at gmail.com>:
>
>
> A silly question, but how do you exit this loop ?
>
> > while True:
> >     u,p = mannwhitneyu(x,y)
>
> Unless you omitted some important part of your script, this loop will keep
> on
> running.
> _______________________________________________
> 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/20070726/05c34f2f/attachment.html>

From robert.kern at gmail.com  Thu Jul 26 17:53:14 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Thu, 26 Jul 2007 16:53:14 -0500
Subject: [SciPy-user] Memory problems with stats.mannwhitneyu
In-Reply-To: <dc90fba40707261341u42fbb603x9171c869768185f2@mail.gmail.com>
References: <dc90fba40707261341u42fbb603x9171c869768185f2@mail.gmail.com>
Message-ID: <46A917CA.2040402@gmail.com>

Martin Blom wrote:
> I need to do lots and lots of mannwhitneyu-tests and my programs keep
> running out of memory so something is wrong somewhere. When I run, for
> example,
> 
> from scipy.stats import mannwhitneyu
> 
> x = [2,4,8,3]
> y = [1,3,5,6]
> 
> while True:
>     u,p = mannwhitneyu(x,y)
> 
> and run top to look at the memory used by python it just keeps
> increasing until I run out. Does anyone who understands the inner
> workings of the stats module know what's happening? Or is it that my
> little loop doesn't release memory in some way? I've been using
> mannwhitneyu for some time now, and I haven't noticed the problem until
> now, which might mean that I've had enough memory to get by until now or
> that there is a recent error, which seems unlikely when I look at the
> changelog.

That is odd. I don't see this on numpy 1.0.3, recentish scipy SVN (maybe up to a
week ago), Python 2.5.1, Intel Mac OS X. Both your code and the code for
mannwhitneyu() look fine to me.

Can you give us similarly-detailed information about the platform you are
running on? If you used binary builds, also please let us know where you got
them from, too. We have had problems in the past with numpy not freeing memory
for array scalars, but I think that was before the 1.0 release.

-- 
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 Jul 26 18:13:16 2007
From: cscheid at sci.utah.edu (Carlos Scheidegger)
Date: Thu, 26 Jul 2007 16:13:16 -0600
Subject: [SciPy-user] Memory problems with stats.mannwhitneyu
In-Reply-To: <46A917CA.2040402@gmail.com>
References: <dc90fba40707261341u42fbb603x9171c869768185f2@mail.gmail.com>
	<46A917CA.2040402@gmail.com>
Message-ID: <46A91C7C.8010604@sci.utah.edu>

> Can you give us similarly-detailed information about the platform you are
> running on? If you used binary builds, also please let us know where you got
> them from, too. We have had problems in the past with numpy not freeing memory
> for array scalars, but I think that was before the 1.0 release.

For whatever it's worth, I see the same issue here, running a plain Ubuntu
Feisty install on AMD64.

$ python
Python 2.5.1 (r251:54863, May  2 2007, 16:27:44)
[GCC 4.1.2 (Ubuntu 4.1.2-0ubuntu4)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import scipy
>>> scipy.__version__
'0.5.2'
>>> import numpy
>>> numpy.__version__
'1.0.1'

$ cat /proc/cpuinfo
processor       : 0
vendor_id       : AuthenticAMD
cpu family      : 15
model           : 33
model name      : Dual Core AMD Opteron(tm) Processor 275
stepping        : 2
cpu MHz         : 2210.186
cache size      : 1024 KB
physical id     : 0
siblings        : 2
core id         : 0
cpu cores       : 2
fpu             : yes
fpu_exception   : yes
cpuid level     : 1
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca
cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm
3dnowext 3dnow pni lahf_lm cmp_legacy
bogomips        : 4423.08
TLB size        : 1024 4K pages
clflush size    : 64
cache_alignment : 64
address sizes   : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp

-carlos


From gael.varoquaux at normalesup.org  Fri Jul 27 02:02:21 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Fri, 27 Jul 2007 08:02:21 +0200
Subject: [SciPy-user] using imread...
In-Reply-To: <46A8FE96.40404@gmail.com>
References: <46A8FE96.40404@gmail.com>
Message-ID: <20070727060221.GA15345@clipper.ens.fr>

On Thu, Jul 26, 2007 at 10:05:42PM +0200, fred wrote:
> So, what am I doing wrong ?

Can you post a minimal example of code ?

Cheers,

Ga?l


From mwojc at p.lodz.pl  Fri Jul 27 06:40:09 2007
From: mwojc at p.lodz.pl (Marek Wojciechowski)
Date: Fri, 27 Jul 2007 12:40:09 +0200 (CEST)
Subject: [SciPy-user] Numpy/Scipy on AIX
In-Reply-To: <mailman.39729.1185450521.30507.scipy-user@scipy.org>
References: <mailman.39729.1185450521.30507.scipy-user@scipy.org>
Message-ID: <34436.147.162.29.65.1185532809.squirrel@poczta.p.lodz.pl>

Hi!
I trying to instal numpy and scipy on AIX 5.3. Unfortunately, running
numpy's setup.py finished with error:

building 'numpy.core._dotblas' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

creating build/temp.aix-5.3-2.5/numpy/core/blasdot
compile options: '-DNO_ATLAS_INFO=1 -Inumpy/core/blasdot
-Inumpy/core/include -Ibuild/src.aix-5.3-2.5/numpy/core -Inumpy/core/src
-Inumpy/core/include -I/home/Marek/apython/include/python2.5 -c'
cc_r: numpy/core/blasdot/_dotblas.c
xlf95 -bshared -F/tmp/tmpwQurYM_xlf.cfg
build/temp.aix-5.3-2.5/numpy/core/blasdot/_dotblas.o -L/usr/lib -lblas -o
build/lib.aix-5.3-2.5/numpy/core/_dotblas.so
ld: 0711-317 ERROR: Undefined symbol: .main
ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more
information.
ld: 0711-317 ERROR: Undefined symbol: .main
ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more
information.
error: Command "xlf95 -bshared -F/tmp/tmpwQurYM_xlf.cfg
build/temp.aix-5.3-2.5/numpy/core/blasdot/_dotblas.o -L/usr/lib -lblas -o
build/lib.aix-5.3-2.5/numpy/core/_dotblas.so" failed with exit status 8

Is there any easy solution for this?
Thanks in advance.

Marek





From david at ar.media.kyoto-u.ac.jp  Fri Jul 27 07:06:59 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Fri, 27 Jul 2007 20:06:59 +0900
Subject: [SciPy-user] Numpy/Scipy on AIX
In-Reply-To: <34436.147.162.29.65.1185532809.squirrel@poczta.p.lodz.pl>
References: <mailman.39729.1185450521.30507.scipy-user@scipy.org>
	<34436.147.162.29.65.1185532809.squirrel@poczta.p.lodz.pl>
Message-ID: <46A9D1D3.7030703@ar.media.kyoto-u.ac.jp>

Marek Wojciechowski wrote:
> Hi!
> I trying to instal numpy and scipy on AIX 5.3. Unfortunately, running
> numpy's setup.py finished with error:
>
> building 'numpy.core._dotblas' extension
> compiling C sources
> C compiler: cc_r -DNDEBUG -O
>
> creating build/temp.aix-5.3-2.5/numpy/core/blasdot
> compile options: '-DNO_ATLAS_INFO=1 -Inumpy/core/blasdot
> -Inumpy/core/include -Ibuild/src.aix-5.3-2.5/numpy/core -Inumpy/core/src
> -Inumpy/core/include -I/home/Marek/apython/include/python2.5 -c'
> cc_r: numpy/core/blasdot/_dotblas.c
> xlf95 -bshared -F/tmp/tmpwQurYM_xlf.cfg
> build/temp.aix-5.3-2.5/numpy/core/blasdot/_dotblas.o -L/usr/lib -lblas -o
> build/lib.aix-5.3-2.5/numpy/core/_dotblas.so
> ld: 0711-317 ERROR: Undefined symbol: .main
> ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more
> information.
> ld: 0711-317 ERROR: Undefined symbol: .main
> ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more
> information.
> error: Command "xlf95 -bshared -F/tmp/tmpwQurYM_xlf.cfg
> build/temp.aix-5.3-2.5/numpy/core/blasdot/_dotblas.o -L/usr/lib -lblas -o
> build/lib.aix-5.3-2.5/numpy/core/_dotblas.so" failed with exit status 8
>
> Is there any easy solution for this?
Do you know which option is needed normally for compiling shared 
libraries on AIX ? The problem seems to be a missing symbol, .main, 
which I supposed is the main function, which is not supposed to be there 
for extensions.

David


From jelleferinga at gmail.com  Fri Jul 27 07:49:31 2007
From: jelleferinga at gmail.com (jelle)
Date: Fri, 27 Jul 2007 11:49:31 +0000 (UTC)
Subject: [SciPy-user] using imread...
References: <46A8FE96.40404@gmail.com> <20070727060221.GA15345@clipper.ens.fr>
Message-ID: <loom.20070727T134841-817@post.gmane.org>


pil 1.1.6 supports numpy arrays, perhaps its even more efficent to use this?




From elcorto at gmx.net  Fri Jul 27 08:39:43 2007
From: elcorto at gmx.net (Steve Schmerler)
Date: Fri, 27 Jul 2007 14:39:43 +0200
Subject: [SciPy-user] Memory problems with stats.mannwhitneyu
In-Reply-To: <46A91C7C.8010604@sci.utah.edu>
References: <dc90fba40707261341u42fbb603x9171c869768185f2@mail.gmail.com>	<46A917CA.2040402@gmail.com>
	<46A91C7C.8010604@sci.utah.edu>
Message-ID: <46A9E78F.6060107@gmx.net>

Carlos Scheidegger wrote:
>> Can you give us similarly-detailed information about the platform you are
>> running on? If you used binary builds, also please let us know where you got
>> them from, too. We have had problems in the past with numpy not freeing memory
>> for array scalars, but I think that was before the 1.0 release.
> 
> For whatever it's worth, I see the same issue here, running a plain Ubuntu
> Feisty install on AMD64.
> 
> $ python
> Python 2.5.1 (r251:54863, May  2 2007, 16:27:44)
> [GCC 4.1.2 (Ubuntu 4.1.2-0ubuntu4)] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
>>>> import scipy
>>>> scipy.__version__
> '0.5.2'
>>>> import numpy
>>>> numpy.__version__
> '1.0.1'
> 
> $ cat /proc/cpuinfo
> processor       : 0
> vendor_id       : AuthenticAMD
> cpu family      : 15
> model           : 33
> model name      : Dual Core AMD Opteron(tm) Processor 275
> stepping        : 2
> cpu MHz         : 2210.186
> cache size      : 1024 KB
> physical id     : 0
> siblings        : 2
> core id         : 0
> cpu cores       : 2
> fpu             : yes
> fpu_exception   : yes
> cpuid level     : 1
> wp              : yes
> flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca
> cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm
> 3dnowext 3dnow pni lahf_lm cmp_legacy
> bogomips        : 4423.08
> TLB size        : 1024 4K pages
> clflush size    : 64
> cache_alignment : 64
> address sizes   : 40 bits physical, 48 bits virtual
> power management: ts fid vid ttp
> 

same here:

Debian etch, python 2.4.4, numpy 1.0.1-8, scipy 0.5.2-0.1 from Debian's 
repos

processor       : 0
vendor_id       : AuthenticAMD
cpu family      : 15
model           : 31
model name      : AMD Athlon(tm) 64 Processor 3000+
stepping        : 0
cpu MHz         : 1808.928
cache size      : 512 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 1
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge 
mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt 
lm 3dnowext 3dnow up lahf_lm ts fid vid ttp
bogomips        : 3621.70

-- 
cheers,
steve

Random number generation is the art of producing pure gibberish as 
quickly as possible.


From stefan at sun.ac.za  Fri Jul 27 09:20:59 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Fri, 27 Jul 2007 15:20:59 +0200
Subject: [SciPy-user] reading .mat-files (Matlab 7)
In-Reply-To: <46A8C23D.6000101@gmail.com>
References: <OF5459A43A.44A36F78-ONC1257323.006904E4-C1257323.006AEC38@matforsk.no>
	<web-86735412@uni-stuttgart.de> <46A7BB0B.3060709@ru.nl>
	<46A89634.20206@unibo.it> <46A8A42F.5050002@gmail.com>
	<46A8A592.3010701@unibo.it> <46A8C23D.6000101@gmail.com>
Message-ID: <20070727132059.GB7447@mentat.za.net>

On Thu, Jul 26, 2007 at 05:48:13PM +0200, stef mientki wrote:
> I think that's one of the differences between the "rich world"  and the 
> "poor world"

You may be surprised at the amount of exploitation that takes place in
the "poor world" by big software companies.  Fortunately, the South
African government is pro free software, and invests big money in
developing alternatives to commercial packages.

At our university, the vast majority of all engineering is still done
using MATLAB.  It's usage is so deeply ingrained, that it is very
difficult to convince people to consider free alternatives.

We are fortunate to have talented teachers and ambassadors in our
midst.  Two years ago, for example, Fernando Perez came to South
Africa, championing the cause and, through his workshop, broadening
the vision of many engineers, mathematicians and engineers.

Sometimes, a bit of effort is needed to set the ball rolling and to
overcome the initial resistance offered by comfort zones.  Once that
has been breached, the facts speak for themselves.

Regards
St?fan


From vaftrudner at gmail.com  Fri Jul 27 12:33:12 2007
From: vaftrudner at gmail.com (Martin Blom)
Date: Fri, 27 Jul 2007 11:33:12 -0500
Subject: [SciPy-user] Memory problems with stats.mannwhitneyu
In-Reply-To: <46A9E78F.6060107@gmx.net>
References: <dc90fba40707261341u42fbb603x9171c869768185f2@mail.gmail.com>
	<46A917CA.2040402@gmail.com> <46A91C7C.8010604@sci.utah.edu>
	<46A9E78F.6060107@gmx.net>
Message-ID: <dc90fba40707270933t76ce2082k864e8eaf941b95ff@mail.gmail.com>

I'm on an Ubuntu Feisty machine, and the python installation is the standard
one for feisty. Machine and python details:


~$ python
Python 2.5.1 (r251:54863, May  2 2007, 16:56:35)
[GCC 4.1.2 (Ubuntu 4.1.2-0ubuntu4)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import scipy
>>> scipy.__version__
'0.5.2'
>>> import numpy
>>> numpy.__version__
'1.0.1'

~$ cat /proc/cpuinfo
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 15
model           : 4
model name      : Intel(R) Pentium(R) D CPU 3.20GHz
stepping        : 4
cpu MHz         : 3192.271
cache size      : 1024 KB
physical id     : 0
siblings        : 2
core id         : 0
cpu cores       : 2
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 5
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca
cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx lm
constant_tsc pni monitor ds_cpl est cid cx16 xtpr
bogomips        : 6389.70
clflush size    : 64

processor       : 1
vendor_id       : GenuineIntel
cpu family      : 15
model           : 4
model name      : Intel(R) Pentium(R) D CPU 3.20GHz
stepping        : 4
cpu MHz         : 3192.271
cache size      : 1024 KB
physical id     : 0
siblings        : 2
core id         : 1
cpu cores       : 2
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 5
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca
cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx lm
constant_tsc pni monitor ds_cpl est cid cx16 xtpr
bogomips        : 6384.39
clflush size    : 64


2007/7/27, Steve Schmerler <elcorto at gmx.net>:
>
> Carlos Scheidegger wrote:
> >> Can you give us similarly-detailed information about the platform you
> are
> >> running on? If you used binary builds, also please let us know where
> you got
> >> them from, too. We have had problems in the past with numpy not freeing
> memory
> >> for array scalars, but I think that was before the 1.0 release.
> >
> > For whatever it's worth, I see the same issue here, running a plain
> Ubuntu
> > Feisty install on AMD64.
> >
> > $ python
> > Python 2.5.1 (r251:54863, May  2 2007, 16:27:44)
> > [GCC 4.1.2 (Ubuntu 4.1.2-0ubuntu4)] on linux2
> > Type "help", "copyright", "credits" or "license" for more information.
> >>>> import scipy
> >>>> scipy.__version__
> > '0.5.2'
> >>>> import numpy
> >>>> numpy.__version__
> > '1.0.1'
> >
> > $ cat /proc/cpuinfo
> > processor       : 0
> > vendor_id       : AuthenticAMD
> > cpu family      : 15
> > model           : 33
> > model name      : Dual Core AMD Opteron(tm) Processor 275
> > stepping        : 2
> > cpu MHz         : 2210.186
> > cache size      : 1024 KB
> > physical id     : 0
> > siblings        : 2
> > core id         : 0
> > cpu cores       : 2
> > fpu             : yes
> > fpu_exception   : yes
> > cpuid level     : 1
> > wp              : yes
> > flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
> mca
> > cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt
> lm
> > 3dnowext 3dnow pni lahf_lm cmp_legacy
> > bogomips        : 4423.08
> > TLB size        : 1024 4K pages
> > clflush size    : 64
> > cache_alignment : 64
> > address sizes   : 40 bits physical, 48 bits virtual
> > power management: ts fid vid ttp
> >
>
> same here:
>
> Debian etch, python 2.4.4, numpy 1.0.1-8, scipy 0.5.2-0.1 from Debian's
> repos
>
> processor       : 0
> vendor_id       : AuthenticAMD
> cpu family      : 15
> model           : 31
> model name      : AMD Athlon(tm) 64 Processor 3000+
> stepping        : 0
> cpu MHz         : 1808.928
> cache size      : 512 KB
> fdiv_bug        : no
> hlt_bug         : no
> f00f_bug        : no
> coma_bug        : no
> fpu             : yes
> fpu_exception   : yes
> cpuid level     : 1
> wp              : yes
> flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
> mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt
> lm 3dnowext 3dnow up lahf_lm ts fid vid ttp
> bogomips        : 3621.70
>
> --
> 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
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070727/00ee6eef/attachment.html>

From zpincus at stanford.edu  Fri Jul 27 12:46:59 2007
From: zpincus at stanford.edu (Zachary Pincus)
Date: Fri, 27 Jul 2007 12:46:59 -0400
Subject: [SciPy-user] BoF on 3D visualization
In-Reply-To: <426ada670707261354w216b195ctfa80ae3a334d11ad@mail.gmail.com>
References: <18086.28241.432931.478493@gargle.gargle.HOWL>
	<46A8DEDD.2060006@ucsf.edu>
	<426ada670707261354w216b195ctfa80ae3a334d11ad@mail.gmail.com>
Message-ID: <E7CC7774-1F51-43BF-9FCB-D3AD6E3BBD73@stanford.edu>

Is there any chance that anyone took notes at this BoF and would be  
so gracious as to post anything for those of us who weren't at the  
meeting but nevertheless interested in this topic?

Thanks,
Zach Pincus


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



On Jul 26, 2007, at 4:54 PM, Carl Trachte wrote:

> I am mostly ignorant with regard to the BoF subject, but I need to  
> get smarter.
> I will attend.
>
> On 7/26/07, Karl Young < Karl.Young at ucsf.edu> wrote:Prabhu  
> Ramachandran wrote:
>
> >Hello,
> >
> >Gael Varoquaux and myself are planning on holding a "Pythonic 3D
> >visualization" BoF at SciPy07 on Thursday evening.  We'd like to know
> >if any of you would be interested in this.  If you are, please do let
> >us know.  Please also let us know if you have anything in particular
> >you'd like to discuss.
> >
> >Thanks.
> >
> >Cheers,
> >
> >
>
> I'll definitely show; issues that come up in 3D visualization of
> neuroimages are tops on my list but I'm sure any discussion will be of
> interest (I'm a real novice re. 3D visualization so I'm not sure if I
> contribute a whole lot of interest)
>
> -- KY
>
> Karl Young
> Center for Imaging of Neurodegenerative Diseases, UCSF
> VA Medical Center (114M)              Phone:  (415) 221-4810 x3114   
> lab
> 4150 Clement Street                   FAX:    (415) 668-2864
> San Francisco, CA 94121               Email:  karl young at ucsf edu
> _______________________________________________
> 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 gael.varoquaux at normalesup.org  Fri Jul 27 12:49:35 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Fri, 27 Jul 2007 18:49:35 +0200
Subject: [SciPy-user] BoF on 3D visualization
In-Reply-To: <E7CC7774-1F51-43BF-9FCB-D3AD6E3BBD73@stanford.edu>
References: <18086.28241.432931.478493@gargle.gargle.HOWL>
	<46A8DEDD.2060006@ucsf.edu>
	<426ada670707261354w216b195ctfa80ae3a334d11ad@mail.gmail.com>
	<E7CC7774-1F51-43BF-9FCB-D3AD6E3BBD73@stanford.edu>
Message-ID: <20070727164934.GB15290@clipper.ens.fr>

On Fri, Jul 27, 2007 at 12:46:59PM -0400, Zachary Pincus wrote:
> Is there any chance that anyone took notes at this BoF and would be  
> so gracious as to post anything for those of us who weren't at the  
> meeting but nevertheless interested in this topic?

It hasn't happened yet (it's on Tues august 14th )

Will try. No promises though.

Ga?l


From rmay at ou.edu  Fri Jul 27 12:57:21 2007
From: rmay at ou.edu (Ryan May)
Date: Fri, 27 Jul 2007 11:57:21 -0500
Subject: [SciPy-user] BoF on 3D visualization
In-Reply-To: <20070727164934.GB15290@clipper.ens.fr>
References: <18086.28241.432931.478493@gargle.gargle.HOWL>
	<46A8DEDD.2060006@ucsf.edu>
	<426ada670707261354w216b195ctfa80ae3a334d11ad@mail.gmail.com>
	<E7CC7774-1F51-43BF-9FCB-D3AD6E3BBD73@stanford.edu>
	<20070727164934.GB15290@clipper.ens.fr>
Message-ID: <46AA23F1.4010806@ou.edu>

Gael Varoquaux wrote:
> On Fri, Jul 27, 2007 at 12:46:59PM -0400, Zachary Pincus wrote:
>> Is there any chance that anyone took notes at this BoF and would be  
>> so gracious as to post anything for those of us who weren't at the  
>> meeting but nevertheless interested in this topic?
> 
> It hasn't happened yet (it's on Tues august 14th )
> 
> Will try. No promises though.
> 
> Ga?l

Don't you mean Thursday August 16th?

Ryan

-- 
Ryan May
Graduate Research Assistant
School of Meteorology
University of Oklahoma


From zpincus at stanford.edu  Fri Jul 27 13:01:03 2007
From: zpincus at stanford.edu (Zachary Pincus)
Date: Fri, 27 Jul 2007 13:01:03 -0400
Subject: [SciPy-user] BoF on 3D visualization
In-Reply-To: <20070727164934.GB15290@clipper.ens.fr>
References: <18086.28241.432931.478493@gargle.gargle.HOWL>
	<46A8DEDD.2060006@ucsf.edu>
	<426ada670707261354w216b195ctfa80ae3a334d11ad@mail.gmail.com>
	<E7CC7774-1F51-43BF-9FCB-D3AD6E3BBD73@stanford.edu>
	<20070727164934.GB15290@clipper.ens.fr>
Message-ID: <01A2E9AB-E898-44B6-A576-9A3B0CFAD4A8@stanford.edu>

> On Fri, Jul 27, 2007 at 12:46:59PM -0400, Zachary Pincus wrote:
>> Is there any chance that anyone took notes at this BoF and would be
>> so gracious as to post anything for those of us who weren't at the
>> meeting but nevertheless interested in this topic?
>
> It hasn't happened yet (it's on Tues august 14th )

Aah, I misread and thought it was yesterday!

> Will try. No promises though.

Thanks in advance! I'm currently wading through all of the different  
3DV options for python and and any insight on which tools are suited  
for which applications would be of course very valuable.

Zach


From prabhu at aero.iitb.ac.in  Fri Jul 27 13:01:20 2007
From: prabhu at aero.iitb.ac.in (Prabhu Ramachandran)
Date: Fri, 27 Jul 2007 22:31:20 +0530
Subject: [SciPy-user] BoF on 3D visualization
In-Reply-To: <20070727164934.GB15290@clipper.ens.fr>
References: <18086.28241.432931.478493@gargle.gargle.HOWL>
	<46A8DEDD.2060006@ucsf.edu>
	<426ada670707261354w216b195ctfa80ae3a334d11ad@mail.gmail.com>
	<E7CC7774-1F51-43BF-9FCB-D3AD6E3BBD73@stanford.edu>
	<20070727164934.GB15290@clipper.ens.fr>
Message-ID: <18090.9440.908892.566658@gargle.gargle.HOWL>

>>>>> "Gael" == Gael Varoquaux <gael.varoquaux at normalesup.org> writes:

    Gael> On Fri, Jul 27, 2007 at 12:46:59PM -0400, Zachary Pincus
    Gael> wrote:
    >> Is there any chance that anyone took notes at this BoF and
    >> would be so gracious as to post anything for those of us who
    >> weren't at the meeting but nevertheless interested in this
    >> topic?

    Gael> It hasn't happened yet (it's on Tues august 14th )

I this is supposed to be on Thursday, Aug. 16th and the exact time
will be announced at the conference so that we can coordinate it
correctly.

    Gael> Will try. No promises though.

:-) BoFs are supposed to be informal, so notes would be handy.  Lets
hope this can be arranged.

regards,
prabhu


From gael.varoquaux at normalesup.org  Fri Jul 27 13:17:11 2007
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Fri, 27 Jul 2007 19:17:11 +0200
Subject: [SciPy-user] BoF on 3D visualization
In-Reply-To: <46AA23F1.4010806@ou.edu>
References: <18086.28241.432931.478493@gargle.gargle.HOWL>
	<46A8DEDD.2060006@ucsf.edu>
	<426ada670707261354w216b195ctfa80ae3a334d11ad@mail.gmail.com>
	<E7CC7774-1F51-43BF-9FCB-D3AD6E3BBD73@stanford.edu>
	<20070727164934.GB15290@clipper.ens.fr> <46AA23F1.4010806@ou.edu>
Message-ID: <20070727171711.GF15290@clipper.ens.fr>

On Fri, Jul 27, 2007 at 11:57:21AM -0500, Ryan May wrote:
> Don't you mean Thursday August 16th?

$ ping brain.gael
no answer from brain.gael

But the fingers are here to correct the error. Yes, you are right. I
guess to a frenchman Thursday and Tuesday sound to close.

Ga?l


From robert.kern at gmail.com  Fri Jul 27 14:43:35 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Fri, 27 Jul 2007 13:43:35 -0500
Subject: [SciPy-user] Memory problems with stats.mannwhitneyu
In-Reply-To: <dc90fba40707270933t76ce2082k864e8eaf941b95ff@mail.gmail.com>
References: <dc90fba40707261341u42fbb603x9171c869768185f2@mail.gmail.com>	<46A917CA.2040402@gmail.com>
	<46A91C7C.8010604@sci.utah.edu>	<46A9E78F.6060107@gmx.net>
	<dc90fba40707270933t76ce2082k864e8eaf941b95ff@mail.gmail.com>
Message-ID: <46AA3CD7.4060504@gmail.com>

Martin Blom wrote:
> I'm on an Ubuntu Feisty machine, and the python installation is the
> standard one for feisty. Machine and python details:
> 
> 
> ~$ python
> Python 2.5.1 (r251:54863, May  2 2007, 16:56:35)
> [GCC 4.1.2 (Ubuntu 4.1.2-0ubuntu4 )] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
>>>> import scipy
>>>> scipy.__version__
> '0.5.2'
>>>> import numpy
>>>> numpy.__version__
> '1.0.1'

Well, you all seem to have numpy 1.0.1 in common. I do not see this on my AMD 64
Ubuntu box with numpy SVN. I suspect a bug in numpy 1.0.1 and suggest upgrading.

-- 
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 mwojc at p.lodz.pl  Fri Jul 27 17:17:12 2007
From: mwojc at p.lodz.pl (Marek)
Date: Fri, 27 Jul 2007 21:17:12 +0000 (UTC)
Subject: [SciPy-user] Numpy/Scipy on AIX
References: <mailman.39729.1185450521.30507.scipy-user@scipy.org>
	<34436.147.162.29.65.1185532809.squirrel@poczta.p.lodz.pl>
	<46A9D1D3.7030703@ar.media.kyoto-u.ac.jp>
Message-ID: <loom.20070727T151618-994@post.gmane.org>

David Cournapeau <david <at> ar.media.kyoto-u.ac.jp> writes:

> 
> Marek Wojciechowski wrote:
> > Hi!
> > I trying to instal numpy and scipy on AIX 5.3. Unfortunately, running
> > numpy's setup.py finished with error:
> >
> > building 'numpy.core._dotblas' extension
> > compiling C sources
> > C compiler: cc_r -DNDEBUG -O
> >
> > creating build/temp.aix-5.3-2.5/numpy/core/blasdot
> > compile options: '-DNO_ATLAS_INFO=1 -Inumpy/core/blasdot
> > -Inumpy/core/include -Ibuild/src.aix-5.3-2.5/numpy/core -Inumpy/core/src
> > -Inumpy/core/include -I/home/Marek/apython/include/python2.5 -c'
> > cc_r: numpy/core/blasdot/_dotblas.c
> > xlf95 -bshared -F/tmp/tmpwQurYM_xlf.cfg
> > build/temp.aix-5.3-2.5/numpy/core/blasdot/_dotblas.o -L/usr/lib -lblas -o
> > build/lib.aix-5.3-2.5/numpy/core/_dotblas.so
> > ld: 0711-317 ERROR: Undefined symbol: .main
> > ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more
> > information.
> > ld: 0711-317 ERROR: Undefined symbol: .main
> > ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more
> > information.
> > error: Command "xlf95 -bshared -F/tmp/tmpwQurYM_xlf.cfg
> > build/temp.aix-5.3-2.5/numpy/core/blasdot/_dotblas.o -L/usr/lib -lblas -o
> > build/lib.aix-5.3-2.5/numpy/core/_dotblas.so" failed with exit status 8
> >
> > Is there any easy solution for this?
> Do you know which option is needed normally for compiling shared 
> libraries on AIX ? The problem seems to be a missing symbol, .main, 
> which I supposed is the main function, which is not supposed to be there 
> for extensions.
> 
> David
> 
Thanks for your response, but I skipped the problem downloading svn version of 
numpy (previously i tried to install numpy-1.0.2), and this compiles nicely. 
But now i run into troubles with scipy. Namely, I get tons of unresolved 
symbols from lapack, for example:

ld: 0711-317 ERROR: Undefined symbol: .sgemm_ 

Any workaround for this? I know there is linker option -brename:.sgemm_,.sgemm, 
which can solve it but this si not a nice solution.

Thanks in advance
Marek




From david at ar.media.kyoto-u.ac.jp  Sat Jul 28 03:37:15 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Sat, 28 Jul 2007 16:37:15 +0900
Subject: [SciPy-user] Numpy/Scipy on AIX
In-Reply-To: <loom.20070727T151618-994@post.gmane.org>
References: <mailman.39729.1185450521.30507.scipy-user@scipy.org>	<34436.147.162.29.65.1185532809.squirrel@poczta.p.lodz.pl>	<46A9D1D3.7030703@ar.media.kyoto-u.ac.jp>
	<loom.20070727T151618-994@post.gmane.org>
Message-ID: <46AAF22B.4040207@ar.media.kyoto-u.ac.jp>

Marek wrote:
>
>
> Thanks for your response, but I skipped the problem downloading svn version of 
> numpy (previously i tried to install numpy-1.0.2), and this compiles nicely. 
> But now i run into troubles with scipy. Namely, I get tons of unresolved 
> symbols from lapack, for example:
>
> ld: 0711-317 ERROR: Undefined symbol: .sgemm_ 
>
> Any workaround for this? I know there is linker option -brename:.sgemm_,.sgemm, 
> which can solve it but this si not a nice solution.
>
>   
We'll need more informations to help you. The most useful thing would be 
to know the output of python setup.py config for numpy and scipy.

Also, did you used the release of scipy (0.5.2) or the svn version as well ?

David


From mwojc at p.lodz.pl  Sat Jul 28 07:47:02 2007
From: mwojc at p.lodz.pl (Marek Wojciechowski)
Date: Sat, 28 Jul 2007 13:47:02 +0200 (CEST)
Subject: [SciPy-user] Numpy/Scipy on AIX
Message-ID: <33682.147.162.29.65.1185623222.squirrel@poczta.p.lodz.pl>

David Cournapeau <david <at> ar.media.kyoto-u.ac.jp> writes:

>
> Marek wrote:
> >
> >
> > Thanks for your response, but I skipped the problem downloading svn
version
of
> > numpy (previously i tried to install numpy-1.0.2), and this compiles
nicely.
> > But now i run into troubles with scipy. Namely, I get tons of
unresolved
> > symbols from lapack, for example:
> >
> > ld: 0711-317 ERROR: Undefined symbol: .sgemm_
> >
> > Any workaround for this? I know there is linker
option -brename:.sgemm_,.sgemm,
> > which can solve it but this si not a nice solution.
> >
> >
> We'll need more informations to help you. The most useful thing would be

> to know the output of python setup.py config for numpy and scipy.
>
> Also, did you used the release of scipy (0.5.2) or the svn version as
well ?
>
> David
>

I'm using svn version of scipy. As far as i observed scipy 0.5.2 is not 
compatibile with numpy-svn. Below i'm pasting what setup.py produces. I
seems things are compiled and there is only linking problem with lapack.
Thanks for help.

========================================================================
mkl_info:
  libraries mkl,vml,guide not found in /home/Marek/apython/lib
  libraries mkl,vml,guide not found in /usr/lib
  NOT AVAILABLE

fftw3_info:
  libraries fftw3 not found in /home/Marek/apython/lib
  libraries fftw3 not found in /usr/lib
  fftw3 not found
  NOT AVAILABLE

fftw2_info:
  libraries rfftw,fftw not found in /home/Marek/apython/lib
  libraries rfftw,fftw not found in /usr/lib
  fftw2 not found
  NOT AVAILABLE

dfftw_info:
  libraries drfftw,dfftw not found in /home/Marek/apython/lib
  libraries drfftw,dfftw not found in /usr/lib
  dfftw not found
  NOT AVAILABLE

djbfft_info:
  NOT AVAILABLE

blas_opt_info:
blas_mkl_info:
  libraries mkl,vml,guide not found in /home/Marek/apython/lib
  libraries mkl,vml,guide not found in /usr/lib
  NOT AVAILABLE

atlas_blas_threads_info:
Setting PTATLAS=ATLAS
  libraries ptf77blas,ptcblas,atlas not found in /home/Marek/apython/lib
libraries ptf77blas,ptcblas,atlas not found in /usr/lib
  NOT AVAILABLE

atlas_blas_info:
  libraries f77blas,cblas,atlas not found in /home/Marek/apython/lib
libraries f77blas,cblas,atlas not found in /usr/lib
  NOT AVAILABLE

blas_info:
  libraries blas not found in /home/Marek/apython/lib
  FOUND:
    libraries = ['blas']
    library_dirs = ['/usr/lib']
    language = f77

  FOUND:
    libraries = ['blas']
    library_dirs = ['/usr/lib']
    define_macros = [('NO_ATLAS_INFO', 1)]
    language = f77

lapack_opt_info:
lapack_mkl_info:
  NOT AVAILABLE

atlas_threads_info:
Setting PTATLAS=ATLAS
  libraries ptf77blas,ptcblas,atlas not found in /home/Marek/apython/lib
libraries lapack_atlas not found in /home/Marek/apython/lib
  libraries ptf77blas,ptcblas,atlas not found in /usr/lib
  libraries lapack_atlas not found in /usr/lib
numpy.distutils.system_info.atlas_threads_info
  NOT AVAILABLE

atlas_info:
  libraries f77blas,cblas,atlas not found in /home/Marek/apython/lib
libraries lapack_atlas not found in /home/Marek/apython/lib
  libraries f77blas,cblas,atlas not found in /usr/lib
  libraries lapack_atlas not found in /usr/lib
numpy.distutils.system_info.atlas_info
  NOT AVAILABLE

lapack_info:
  FOUND:
    libraries = ['lapack']
    library_dirs = ['/home/Marek/apython/lib']
    language = f77

  FOUND:
    libraries = ['lapack', 'blas']
    library_dirs = ['/home/Marek/apython/lib', '/usr/lib']
    define_macros = [('NO_ATLAS_INFO', 1)]
    language = f77

non-existing path in 'Lib/linsolve': 'tests'
umfpack_info:
  libraries umfpack not found in /home/Marek/apython/lib
  libraries umfpack not found in /usr/lib
  NOT AVAILABLE

non-existing path in 'Lib/maxentropy': 'doc'
running build
running config_cc
unifing config_cc, config, build_clib, build_ext, build commands
--compiler options
running config_fc
unifing config_fc, config, build_clib, build_ext, build commands
--fcompiler options
running build_src
building py_modules sources
building library "dfftpack" sources
building library "linpack_lite" sources
building library "mach" sources
building library "quadpack" sources
building library "odepack" sources
building library "fitpack" sources
building library "superlu_src" sources
building library "odrpack" sources
building library "minpack" sources
building library "rootfind" sources
building library "c_misc" sources
building library "cephes" sources
building library "mach" sources
building library "toms" sources
building library "amos" sources
building library "cdf" sources
building library "specfun" sources
building library "statlib" sources
building extension "scipy.cluster._vq" sources
building extension "scipy.fftpack._fftpack" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.fftpack.convolve" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.integrate._quadpack" sources
building extension "scipy.integrate._odepack" sources
building extension "scipy.integrate.vode" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.interpolate._fitpack" sources
building extension "scipy.interpolate.dfitpack" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
  adding 'build/src.aix-5.3-2.5/Lib/interpolate/dfitpack-f2pywrappers.f'
to sources.
building extension "scipy.io.numpyio" sources
building extension "scipy.lib.blas.fblas" sources
f2py options: ['skip:', ':']
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
  adding
'build/src.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/lib/blas/fblas-f2pywrappers.f'
to sources.
building extension "scipy.lib.blas.cblas" sources
  adding 'build/src.aix-5.3-2.5/scipy/lib/blas/cblas.pyf' to sources.
f2py options: ['skip:', ':']
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.lib.lapack.flapack" sources
f2py options: ['skip:', ':']
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.lib.lapack.clapack" sources
  adding 'build/src.aix-5.3-2.5/scipy/lib/lapack/clapack.pyf' to sources.
f2py options: ['skip:', ':']
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.lib.lapack.calc_lwork" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.lib.lapack.atlas_version" sources
building extension "scipy.linalg.fblas" sources
  adding 'build/src.aix-5.3-2.5/scipy/linalg/fblas.pyf' to sources.
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
  adding
'build/src.aix-5.3-2.5/build/src.aix-5.3-2.5/scipy/linalg/fblas-f2pywrappers.f'
to sources.
building extension "scipy.linalg.cblas" sources
  adding 'build/src.aix-5.3-2.5/scipy/linalg/cblas.pyf' to sources.
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.linalg.flapack" sources
  adding 'build/src.aix-5.3-2.5/scipy/linalg/flapack.pyf' to sources.
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
  adding
'build/src.aix-5.3-2.5/build/src.aix-5.3-2.5/scipy/linalg/flapack-f2pywrappers.f'
to sources.
building extension "scipy.linalg.clapack" sources
  adding 'build/src.aix-5.3-2.5/scipy/linalg/clapack.pyf' to sources.
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.linalg._flinalg" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.linalg.calc_lwork" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.linalg.atlas_version" sources
building extension "scipy.linalg._iterative" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.linsolve._zsuperlu" sources
building extension "scipy.linsolve._dsuperlu" sources
building extension "scipy.linsolve._csuperlu" sources
building extension "scipy.linsolve._ssuperlu" sources
building extension "scipy.linsolve.umfpack.__umfpack" sources
building extension "scipy.odr.__odrpack" sources
building extension "scipy.optimize._minpack" sources
building extension "scipy.optimize._zeros" sources
building extension "scipy.optimize._lbfgsb" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.optimize.moduleTNC" sources
building extension "scipy.optimize._cobyla" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.optimize.minpack2" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.signal.sigtools" sources
building extension "scipy.signal.spline" sources
building extension "scipy.special._cephes" sources
building extension "scipy.special.specfun" sources
f2py options: ['--no-wrap-functions']
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.stats.statlib" sources
f2py options: ['--no-wrap-functions']
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.stats.futil" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
building extension "scipy.stats.mvn" sources
f2py options: []
  adding 'build/src.aix-5.3-2.5/fortranobject.c' to sources.
  adding 'build/src.aix-5.3-2.5' to include_dirs.
  adding 'build/src.aix-5.3-2.5/Lib/stats/mvn-f2pywrappers.f' to sources.
building extension "scipy.stsci.convolve._correlate" sources
building extension "scipy.stsci.convolve._lineshape" sources
building extension "scipy.stsci.image._combine" sources
building data_files sources
running build_py
copying Lib/__svn_version__.py -> build/lib.aix-5.3-2.5/scipy
copying build/src.aix-5.3-2.5/scipy/__config__.py ->
build/lib.aix-5.3-2.5/scipy
running build_clib
customize UnixCCompiler
customize UnixCCompiler using build_clib
customize IBMFCompiler
Found executable /usr/bin/xlf90
Found executable /usr/bin/xlf
Found executable /usr/bin/xlf95
Found executable /usr/bin/lslpp
#Nome Pacchetto:Sottopacchetto:Livello:Stato:ID PTF:St.
Correz.:Tipo:Descrizione:Indir. destinaz.:Uninstaller:Catalogo
messaggi:Insieme di messaggi:Numero messaggio:Parent:Automatico:EFIX bloccata
xlfcmp:xlfcmp:9.1.0.0: : :C: :XL Fortran Compiler: : : : : : :0:0:
Creating /tmp/tmpDeHt5Z/8iHliD_xlf.cfg
#Nome Pacchetto:Sottopacchetto:Livello:Stato:ID PTF:St.
Correz.:Tipo:Descrizione:Indir. destinaz.:Uninstaller:Catalogo
messaggi:Insieme di messaggi:Numero messaggio:Parent:Automatico:EFIX bloccata
xlfcmp:xlfcmp:9.1.0.0: : :C: :XL Fortran Compiler: : : : : : :0:0:
customize IBMFCompiler
#Nome Pacchetto:Sottopacchetto:Livello:Stato:ID PTF:St.
Correz.:Tipo:Descrizione:Indir. destinaz.:Uninstaller:Catalogo
messaggi:Insieme di messaggi:Numero messaggio:Parent:Automatico:EFIX bloccata
xlfcmp:xlfcmp:9.1.0.0: : :C: :XL Fortran Compiler: : : : : : :0:0:
Creating /tmp/tmpDeHt5Z/qt0Obz_xlf.cfg
customize IBMFCompiler using build_clib
#Nome Pacchetto:Sottopacchetto:Livello:Stato:ID PTF:St.
Correz.:Tipo:Descrizione:Indir. destinaz.:Uninstaller:Catalogo
messaggi:Insieme di messaggi:Numero messaggio:Parent:Automatico:EFIX bloccata
xlfcmp:xlfcmp:9.1.0.0: : :C: :XL Fortran Compiler: : : : : : :0:0: running
build_ext
customize UnixCCompiler
customize UnixCCompiler using build_ext
library 'mach' defined more than once, overwriting build_info
{'sources': ['Lib/integrate/mach/r1mach.f', 'Lib/integrate/mach/d1mach.f',
'Lib/integrate/mach/xerror.f', 'Lib/integrate/mach/i1mach.f'],
'config_fc': {'noopt': ('Lib/integrate/setup.pyc', 1)},
'source_languages': ['f77']}...
with
{'sources': ['Lib/special/mach/r1mach.f', 'Lib/special/mach/d1mach.f',
'Lib/special/mach/xerror.f', 'Lib/special/mach/i1mach.f'], 'config_fc':
{'noopt': ('Lib/special/setup.pyc', 1)}, 'source_languages': ['f77']}...
extending extension 'scipy.linsolve._zsuperlu' defined_macros with
[('USE_VENDOR_BLAS', 1)]
extending extension 'scipy.linsolve._dsuperlu' defined_macros with
[('USE_VENDOR_BLAS', 1)]
extending extension 'scipy.linsolve._csuperlu' defined_macros with
[('USE_VENDOR_BLAS', 1)]
extending extension 'scipy.linsolve._ssuperlu' defined_macros with
[('USE_VENDOR_BLAS', 1)]
customize IBMFCompiler
#Nome Pacchetto:Sottopacchetto:Livello:Stato:ID PTF:St.
Correz.:Tipo:Descrizione:Indir. destinaz.:Uninstaller:Catalogo
messaggi:Insieme di messaggi:Numero messaggio:Parent:Automatico:EFIX bloccata
xlfcmp:xlfcmp:9.1.0.0: : :C: :XL Fortran Compiler: : : : : : :0:0:
Creating /tmp/tmpDeHt5Z/KqMbHj_xlf.cfg
#Nome Pacchetto:Sottopacchetto:Livello:Stato:ID PTF:St.
Correz.:Tipo:Descrizione:Indir. destinaz.:Uninstaller:Catalogo
messaggi:Insieme di messaggi:Numero messaggio:Parent:Automatico:EFIX bloccata
xlfcmp:xlfcmp:9.1.0.0: : :C: :XL Fortran Compiler: : : : : : :0:0:
customize IBMFCompiler
#Nome Pacchetto:Sottopacchetto:Livello:Stato:ID PTF:St.
Correz.:Tipo:Descrizione:Indir. destinaz.:Uninstaller:Catalogo
messaggi:Insieme di messaggi:Numero messaggio:Parent:Automatico:EFIX bloccata
xlfcmp:xlfcmp:9.1.0.0: : :C: :XL Fortran Compiler: : : : : : :0:0:
Creating /tmp/tmpDeHt5Z/XAGDI2_xlf.cfg
#Nome Pacchetto:Sottopacchetto:Livello:Stato:ID PTF:St.
Correz.:Tipo:Descrizione:Indir. destinaz.:Uninstaller:Catalogo
messaggi:Insieme di messaggi:Numero messaggio:Parent:Automatico:EFIX bloccata
xlfcmp:xlfcmp:9.1.0.0: : :C: :XL Fortran Compiler: : : : : : :0:0:
customize IBMFCompiler using build_ext
#Nome Pacchetto:Sottopacchetto:Livello:Stato:ID PTF:St.
Correz.:Tipo:Descrizione:Indir. destinaz.:Uninstaller:Catalogo
messaggi:Insieme di messaggi:Numero messaggio:Parent:Automatico:EFIX bloccata
xlfcmp:xlfcmp:9.1.0.0: : :C: :XL Fortran Compiler: : : : : : :0:0:
building 'scipy.integrate._odepack' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/Lib/integrate/_odepackmodule.o -L/usr/lib
-Lbuild/temp.aix-5.3-2.5 -lodepack -llinpack_lite -lmach -lblas -o
build/lib.aix-5.3-2.5/scipy/integrate/_odepack.so
ld: 0711-317 ERRORE: Simbolo non definito : .idamax_
ld: 0711-317 ERRORE: Simbolo non definito : .dscal_
ld: 0711-317 ERRORE: Simbolo non definito : .daxpy_
ld: 0711-317 ERRORE: Simbolo non definito : .ddot_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.integrate.vode' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/integrate/vodemodule.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/fortranobject.o -L/usr/lib
-Lbuild/temp.aix-5.3-2.5 -lodepack -llinpack_lite -lmach -lblas -o
build/lib.aix-5.3-2.5/scipy/integrate/vode.so
ld: 0711-317 ERRORE: Simbolo non definito : .dscal_
ld: 0711-317 ERRORE: Simbolo non definito : .dcopy_
ld: 0711-317 ERRORE: Simbolo non definito : .idamax_
ld: 0711-317 ERRORE: Simbolo non definito : .daxpy_
ld: 0711-317 ERRORE: Simbolo non definito : .ddot_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.lib.blas.fblas' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
compiling Fortran sources
Fortran f77 compiler: /usr/bin/xlf -qextname -O5
Fortran f90 compiler: /usr/bin/xlf90 -qextname -O5
Fortran fix compiler: /usr/bin/xlf90 -qfixed -qextname -O5
compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/lib/blas/fblasmodule.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/fortranobject.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/lib/blas/fblaswrap.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/lib/blas/fblas-f2pywrappers.o
-L/usr/lib -Lbuild/temp.aix-5.3-2.5 -lblas -o
build/lib.aix-5.3-2.5/scipy/lib/blas/fblas.so
/usr/bin/xlf95: 1501-262 One or more input object files contain IPA
information: specify -qipa for additional optimization.
ld: 0711-317 ERRORE: Simbolo non definito : srotg_
ld: 0711-317 ERRORE: Simbolo non definito : drotg_
ld: 0711-317 ERRORE: Simbolo non definito : crotg_
ld: 0711-317 ERRORE: Simbolo non definito : zrotg_
ld: 0711-317 ERRORE: Simbolo non definito : srotmg_
ld: 0711-317 ERRORE: Simbolo non definito : drotmg_
ld: 0711-317 ERRORE: Simbolo non definito : srot_
ld: 0711-317 ERRORE: Simbolo non definito : drot_
ld: 0711-317 ERRORE: Simbolo non definito : csrot_
ld: 0711-317 ERRORE: Simbolo non definito : zdrot_
ld: 0711-317 ERRORE: Simbolo non definito : srotm_
ld: 0711-317 ERRORE: Simbolo non definito : drotm_
ld: 0711-317 ERRORE: Simbolo non definito : sswap_
ld: 0711-317 ERRORE: Simbolo non definito : dswap_
ld: 0711-317 ERRORE: Simbolo non definito : cswap_
ld: 0711-317 ERRORE: Simbolo non definito : zswap_
ld: 0711-317 ERRORE: Simbolo non definito : sscal_
ld: 0711-317 ERRORE: Simbolo non definito : dscal_
ld: 0711-317 ERRORE: Simbolo non definito : cscal_
ld: 0711-317 ERRORE: Simbolo non definito : zscal_
ld: 0711-317 ERRORE: Simbolo non definito : csscal_
ld: 0711-317 ERRORE: Simbolo non definito : zdscal_
ld: 0711-317 ERRORE: Simbolo non definito : scopy_
ld: 0711-317 ERRORE: Simbolo non definito : dcopy_
ld: 0711-317 ERRORE: Simbolo non definito : ccopy_
ld: 0711-317 ERRORE: Simbolo non definito : zcopy_
ld: 0711-317 ERRORE: Simbolo non definito : saxpy_
ld: 0711-317 ERRORE: Simbolo non definito : daxpy_
ld: 0711-317 ERRORE: Simbolo non definito : caxpy_
ld: 0711-317 ERRORE: Simbolo non definito : zaxpy_
ld: 0711-317 ERRORE: Simbolo non definito : .sdot_
ld: 0711-317 ERRORE: Simbolo non definito : .ddot_
ld: 0711-317 ERRORE: Simbolo non definito : .cdotu_
ld: 0711-317 ERRORE: Simbolo non definito : .zdotu_
ld: 0711-317 ERRORE: Simbolo non definito : .cdotc_
ld: 0711-317 ERRORE: Simbolo non definito : .zdotc_
ld: 0711-317 ERRORE: Simbolo non definito : .snrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .dnrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .scnrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .dznrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .sasum_
ld: 0711-317 ERRORE: Simbolo non definito : .dasum_
ld: 0711-317 ERRORE: Simbolo non definito : .scasum_
ld: 0711-317 ERRORE: Simbolo non definito : .dzasum_
ld: 0711-317 ERRORE: Simbolo non definito : isamax_
ld: 0711-317 ERRORE: Simbolo non definito : idamax_
ld: 0711-317 ERRORE: Simbolo non definito : icamax_
ld: 0711-317 ERRORE: Simbolo non definito : izamax_
ld: 0711-317 ERRORE: Simbolo non definito : sgemv_
ld: 0711-317 ERRORE: Simbolo non definito : dgemv_
ld: 0711-317 ERRORE: Simbolo non definito : cgemv_
ld: 0711-317 ERRORE: Simbolo non definito : zgemv_
ld: 0711-317 ERRORE: Simbolo non definito : ssymv_
ld: 0711-317 ERRORE: Simbolo non definito : dsymv_
ld: 0711-317 ERRORE: Simbolo non definito : chemv_
ld: 0711-317 ERRORE: Simbolo non definito : zhemv_
ld: 0711-317 ERRORE: Simbolo non definito : strmv_
ld: 0711-317 ERRORE: Simbolo non definito : dtrmv_
ld: 0711-317 ERRORE: Simbolo non definito : ctrmv_
ld: 0711-317 ERRORE: Simbolo non definito : ztrmv_
ld: 0711-317 ERRORE: Simbolo non definito : sger_
ld: 0711-317 ERRORE: Simbolo non definito : dger_
ld: 0711-317 ERRORE: Simbolo non definito : cgeru_
ld: 0711-317 ERRORE: Simbolo non definito : zgeru_
ld: 0711-317 ERRORE: Simbolo non definito : cgerc_
ld: 0711-317 ERRORE: Simbolo non definito : zgerc_
ld: 0711-317 ERRORE: Simbolo non definito : sgemm_
ld: 0711-317 ERRORE: Simbolo non definito : dgemm_
ld: 0711-317 ERRORE: Simbolo non definito : cgemm_
ld: 0711-317 ERRORE: Simbolo non definito : zgemm_
ld: 0711-317 ERRORE: Simbolo non definito : sdot_
ld: 0711-317 ERRORE: Simbolo non definito : ddot_
ld: 0711-317 ERRORE: Simbolo non definito : snrm2_
ld: 0711-317 ERRORE: Simbolo non definito : dnrm2_
ld: 0711-317 ERRORE: Simbolo non definito : scnrm2_
ld: 0711-317 ERRORE: Simbolo non definito : dznrm2_
ld: 0711-317 ERRORE: Simbolo non definito : sasum_
ld: 0711-317 ERRORE: Simbolo non definito : dasum_
ld: 0711-317 ERRORE: Simbolo non definito : scasum_
ld: 0711-317 ERRORE: Simbolo non definito : dzasum_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.lib.lapack.flapack' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/lib/lapack/flapackmodule.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/fortranobject.o
-L/home/Marek/apython/lib -L/usr/lib -Lbuild/temp.aix-5.3-2.5 -llapack
-lblas -o build/lib.aix-5.3-2.5/scipy/lib/lapack/flapack.so
ld: 0711-317 ERRORE: Simbolo non definito : sgesv_
ld: 0711-317 ERRORE: Simbolo non definito : dgesv_
ld: 0711-317 ERRORE: Simbolo non definito : cgesv_
ld: 0711-317 ERRORE: Simbolo non definito : zgesv_
ld: 0711-317 ERRORE: Simbolo non definito : sgbsv_
ld: 0711-317 ERRORE: Simbolo non definito : dgbsv_
ld: 0711-317 ERRORE: Simbolo non definito : cgbsv_
ld: 0711-317 ERRORE: Simbolo non definito : zgbsv_
ld: 0711-317 ERRORE: Simbolo non definito : sposv_
ld: 0711-317 ERRORE: Simbolo non definito : dposv_
ld: 0711-317 ERRORE: Simbolo non definito : cposv_
ld: 0711-317 ERRORE: Simbolo non definito : zposv_
ld: 0711-317 ERRORE: Simbolo non definito : sgelss_
ld: 0711-317 ERRORE: Simbolo non definito : dgelss_
ld: 0711-317 ERRORE: Simbolo non definito : cgelss_
ld: 0711-317 ERRORE: Simbolo non definito : zgelss_
ld: 0711-317 ERRORE: Simbolo non definito : ssyev_
ld: 0711-317 ERRORE: Simbolo non definito : dsyev_
ld: 0711-317 ERRORE: Simbolo non definito : cheev_
ld: 0711-317 ERRORE: Simbolo non definito : zheev_
ld: 0711-317 ERRORE: Simbolo non definito : ssyevd_
ld: 0711-317 ERRORE: Simbolo non definito : dsyevd_
ld: 0711-317 ERRORE: Simbolo non definito : cheevd_
ld: 0711-317 ERRORE: Simbolo non definito : zheevd_
ld: 0711-317 ERRORE: Simbolo non definito : ssyevr_
ld: 0711-317 ERRORE: Simbolo non definito : dsyevr_
ld: 0711-317 ERRORE: Simbolo non definito : cheevr_
ld: 0711-317 ERRORE: Simbolo non definito : zheevr_
ld: 0711-317 ERRORE: Simbolo non definito : sgees_
ld: 0711-317 ERRORE: Simbolo non definito : dgees_
ld: 0711-317 ERRORE: Simbolo non definito : cgees_
ld: 0711-317 ERRORE: Simbolo non definito : zgees_
ld: 0711-317 ERRORE: Simbolo non definito : sgeev_
ld: 0711-317 ERRORE: Simbolo non definito : dgeev_
ld: 0711-317 ERRORE: Simbolo non definito : cgeev_
ld: 0711-317 ERRORE: Simbolo non definito : zgeev_
ld: 0711-317 ERRORE: Simbolo non definito : sgesdd_
ld: 0711-317 ERRORE: Simbolo non definito : dgesdd_
ld: 0711-317 ERRORE: Simbolo non definito : cgesdd_
ld: 0711-317 ERRORE: Simbolo non definito : zgesdd_
ld: 0711-317 ERRORE: Simbolo non definito : ssygv_
ld: 0711-317 ERRORE: Simbolo non definito : dsygv_
ld: 0711-317 ERRORE: Simbolo non definito : chegv_
ld: 0711-317 ERRORE: Simbolo non definito : zhegv_
ld: 0711-317 ERRORE: Simbolo non definito : ssygvd_
ld: 0711-317 ERRORE: Simbolo non definito : dsygvd_
ld: 0711-317 ERRORE: Simbolo non definito : chegvd_
ld: 0711-317 ERRORE: Simbolo non definito : zhegvd_
ld: 0711-317 ERRORE: Simbolo non definito : sggev_
ld: 0711-317 ERRORE: Simbolo non definito : dggev_
ld: 0711-317 ERRORE: Simbolo non definito : cggev_
ld: 0711-317 ERRORE: Simbolo non definito : zggev_
ld: 0711-317 ERRORE: Simbolo non definito : sgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : dgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : cgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : zgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : spotrf_
ld: 0711-317 ERRORE: Simbolo non definito : dpotrf_
ld: 0711-317 ERRORE: Simbolo non definito : cpotrf_
ld: 0711-317 ERRORE: Simbolo non definito : zpotrf_
ld: 0711-317 ERRORE: Simbolo non definito : sgetrs_
ld: 0711-317 ERRORE: Simbolo non definito : dgetrs_
ld: 0711-317 ERRORE: Simbolo non definito : cgetrs_
ld: 0711-317 ERRORE: Simbolo non definito : zgetrs_
ld: 0711-317 ERRORE: Simbolo non definito : spotrs_
ld: 0711-317 ERRORE: Simbolo non definito : dpotrs_
ld: 0711-317 ERRORE: Simbolo non definito : cpotrs_
ld: 0711-317 ERRORE: Simbolo non definito : zpotrs_
ld: 0711-317 ERRORE: Simbolo non definito : sgetri_
ld: 0711-317 ERRORE: Simbolo non definito : dgetri_
ld: 0711-317 ERRORE: Simbolo non definito : cgetri_
ld: 0711-317 ERRORE: Simbolo non definito : zgetri_
ld: 0711-317 ERRORE: Simbolo non definito : spotri_
ld: 0711-317 ERRORE: Simbolo non definito : dpotri_
ld: 0711-317 ERRORE: Simbolo non definito : cpotri_
ld: 0711-317 ERRORE: Simbolo non definito : zpotri_
ld: 0711-317 ERRORE: Simbolo non definito : strtri_
ld: 0711-317 ERRORE: Simbolo non definito : dtrtri_
ld: 0711-317 ERRORE: Simbolo non definito : ctrtri_
ld: 0711-317 ERRORE: Simbolo non definito : ztrtri_
ld: 0711-317 ERRORE: Simbolo non definito : sgeqrf_
ld: 0711-317 ERRORE: Simbolo non definito : dgeqrf_
ld: 0711-317 ERRORE: Simbolo non definito : cgeqrf_
ld: 0711-317 ERRORE: Simbolo non definito : zgeqrf_
ld: 0711-317 ERRORE: Simbolo non definito : sorgqr_
ld: 0711-317 ERRORE: Simbolo non definito : dorgqr_
ld: 0711-317 ERRORE: Simbolo non definito : cungqr_
ld: 0711-317 ERRORE: Simbolo non definito : zungqr_
ld: 0711-317 ERRORE: Simbolo non definito : sgehrd_
ld: 0711-317 ERRORE: Simbolo non definito : dgehrd_
ld: 0711-317 ERRORE: Simbolo non definito : cgehrd_
ld: 0711-317 ERRORE: Simbolo non definito : zgehrd_
ld: 0711-317 ERRORE: Simbolo non definito : sgebal_
ld: 0711-317 ERRORE: Simbolo non definito : dgebal_
ld: 0711-317 ERRORE: Simbolo non definito : cgebal_
ld: 0711-317 ERRORE: Simbolo non definito : zgebal_
ld: 0711-317 ERRORE: Simbolo non definito : slauum_
ld: 0711-317 ERRORE: Simbolo non definito : dlauum_
ld: 0711-317 ERRORE: Simbolo non definito : clauum_
ld: 0711-317 ERRORE: Simbolo non definito : zlauum_
ld: 0711-317 ERRORE: Simbolo non definito : slaswp_
ld: 0711-317 ERRORE: Simbolo non definito : dlaswp_
ld: 0711-317 ERRORE: Simbolo non definito : claswp_
ld: 0711-317 ERRORE: Simbolo non definito : zlaswp_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.lib.lapack.calc_lwork' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
compiling Fortran sources
Fortran f77 compiler: /usr/bin/xlf -qextname -O5
Fortran f90 compiler: /usr/bin/xlf90 -qextname -O5
Fortran fix compiler: /usr/bin/xlf90 -qfixed -qextname -O5
compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/scipy/lib/lapack/calc_lworkmodule.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/fortranobject.o
build/temp.aix-5.3-2.5/Lib/lib/lapack/calc_lwork.o
-L/home/Marek/apython/lib -L/usr/lib -Lbuild/temp.aix-5.3-2.5 -llapack
-lblas -o build/lib.aix-5.3-2.5/scipy/lib/lapack/calc_lwork.so
/usr/bin/xlf95: 1501-262 One or more input object files contain IPA
information: specify -qipa for additional optimization.
ld: 0711-317 ERRORE: Simbolo non definito : .ilaenv_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.linalg.fblas' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
compiling Fortran sources
Fortran f77 compiler: /usr/bin/xlf -qextname -O5
Fortran f90 compiler: /usr/bin/xlf90 -qextname -O5
Fortran fix compiler: /usr/bin/xlf90 -qfixed -qextname -O5
compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/build/src.aix-5.3-2.5/scipy/linalg/fblasmodule.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/fortranobject.o
build/temp.aix-5.3-2.5/Lib/linalg/src/fblaswrap.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/build/src.aix-5.3-2.5/scipy/linalg/fblas-f2pywrappers.o
-L/home/Marek/apython/lib -L/usr/lib -Lbuild/temp.aix-5.3-2.5 -llapack
-lblas -o build/lib.aix-5.3-2.5/scipy/linalg/fblas.so
/usr/bin/xlf95: 1501-262 One or more input object files contain IPA
information: specify -qipa for additional optimization.
ld: 0711-317 ERRORE: Simbolo non definito : srotg_
ld: 0711-317 ERRORE: Simbolo non definito : drotg_
ld: 0711-317 ERRORE: Simbolo non definito : crotg_
ld: 0711-317 ERRORE: Simbolo non definito : zrotg_
ld: 0711-317 ERRORE: Simbolo non definito : srotmg_
ld: 0711-317 ERRORE: Simbolo non definito : drotmg_
ld: 0711-317 ERRORE: Simbolo non definito : srot_
ld: 0711-317 ERRORE: Simbolo non definito : drot_
ld: 0711-317 ERRORE: Simbolo non definito : csrot_
ld: 0711-317 ERRORE: Simbolo non definito : zdrot_
ld: 0711-317 ERRORE: Simbolo non definito : srotm_
ld: 0711-317 ERRORE: Simbolo non definito : drotm_
ld: 0711-317 ERRORE: Simbolo non definito : sswap_
ld: 0711-317 ERRORE: Simbolo non definito : dswap_
ld: 0711-317 ERRORE: Simbolo non definito : cswap_
ld: 0711-317 ERRORE: Simbolo non definito : zswap_
ld: 0711-317 ERRORE: Simbolo non definito : sscal_
ld: 0711-317 ERRORE: Simbolo non definito : dscal_
ld: 0711-317 ERRORE: Simbolo non definito : cscal_
ld: 0711-317 ERRORE: Simbolo non definito : zscal_
ld: 0711-317 ERRORE: Simbolo non definito : csscal_
ld: 0711-317 ERRORE: Simbolo non definito : zdscal_
ld: 0711-317 ERRORE: Simbolo non definito : scopy_
ld: 0711-317 ERRORE: Simbolo non definito : dcopy_
ld: 0711-317 ERRORE: Simbolo non definito : ccopy_
ld: 0711-317 ERRORE: Simbolo non definito : zcopy_
ld: 0711-317 ERRORE: Simbolo non definito : saxpy_
ld: 0711-317 ERRORE: Simbolo non definito : daxpy_
ld: 0711-317 ERRORE: Simbolo non definito : caxpy_
ld: 0711-317 ERRORE: Simbolo non definito : zaxpy_
ld: 0711-317 ERRORE: Simbolo non definito : .cdotu_
ld: 0711-317 ERRORE: Simbolo non definito : .zdotu_
ld: 0711-317 ERRORE: Simbolo non definito : .cdotc_
ld: 0711-317 ERRORE: Simbolo non definito : .zdotc_
ld: 0711-317 ERRORE: Simbolo non definito : sgemv_
ld: 0711-317 ERRORE: Simbolo non definito : dgemv_
ld: 0711-317 ERRORE: Simbolo non definito : cgemv_
ld: 0711-317 ERRORE: Simbolo non definito : zgemv_
ld: 0711-317 ERRORE: Simbolo non definito : chemv_
ld: 0711-317 ERRORE: Simbolo non definito : zhemv_
ld: 0711-317 ERRORE: Simbolo non definito : ssymv_
ld: 0711-317 ERRORE: Simbolo non definito : dsymv_
ld: 0711-317 ERRORE: Simbolo non definito : strmv_
ld: 0711-317 ERRORE: Simbolo non definito : dtrmv_
ld: 0711-317 ERRORE: Simbolo non definito : ctrmv_
ld: 0711-317 ERRORE: Simbolo non definito : ztrmv_
ld: 0711-317 ERRORE: Simbolo non definito : sger_
ld: 0711-317 ERRORE: Simbolo non definito : dger_
ld: 0711-317 ERRORE: Simbolo non definito : cgeru_
ld: 0711-317 ERRORE: Simbolo non definito : zgeru_
ld: 0711-317 ERRORE: Simbolo non definito : cgerc_
ld: 0711-317 ERRORE: Simbolo non definito : zgerc_
ld: 0711-317 ERRORE: Simbolo non definito : sgemm_
ld: 0711-317 ERRORE: Simbolo non definito : dgemm_
ld: 0711-317 ERRORE: Simbolo non definito : cgemm_
ld: 0711-317 ERRORE: Simbolo non definito : zgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .sdot_
ld: 0711-317 ERRORE: Simbolo non definito : .ddot_
ld: 0711-317 ERRORE: Simbolo non definito : .snrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .dnrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .scnrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .dznrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .sasum_
ld: 0711-317 ERRORE: Simbolo non definito : .dasum_
ld: 0711-317 ERRORE: Simbolo non definito : .scasum_
ld: 0711-317 ERRORE: Simbolo non definito : .dzasum_
ld: 0711-317 ERRORE: Simbolo non definito : isamax_
ld: 0711-317 ERRORE: Simbolo non definito : idamax_
ld: 0711-317 ERRORE: Simbolo non definito : icamax_
ld: 0711-317 ERRORE: Simbolo non definito : izamax_
ld: 0711-317 ERRORE: Simbolo non definito : sdot_
ld: 0711-317 ERRORE: Simbolo non definito : ddot_
ld: 0711-317 ERRORE: Simbolo non definito : snrm2_
ld: 0711-317 ERRORE: Simbolo non definito : dnrm2_
ld: 0711-317 ERRORE: Simbolo non definito : scnrm2_
ld: 0711-317 ERRORE: Simbolo non definito : dznrm2_
ld: 0711-317 ERRORE: Simbolo non definito : sasum_
ld: 0711-317 ERRORE: Simbolo non definito : dasum_
ld: 0711-317 ERRORE: Simbolo non definito : scasum_
ld: 0711-317 ERRORE: Simbolo non definito : dzasum_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.linalg.flapack' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
compiling Fortran sources
Fortran f77 compiler: /usr/bin/xlf -qextname -O5
Fortran f90 compiler: /usr/bin/xlf90 -qextname -O5
Fortran fix compiler: /usr/bin/xlf90 -qfixed -qextname -O5
compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/build/src.aix-5.3-2.5/scipy/linalg/flapackmodule.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/fortranobject.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/build/src.aix-5.3-2.5/scipy/linalg/flapack-f2pywrappers.o
-L/home/Marek/apython/lib -L/usr/lib -Lbuild/temp.aix-5.3-2.5 -llapack
-lblas -o build/lib.aix-5.3-2.5/scipy/linalg/flapack.so
/usr/bin/xlf95: 1501-262 One or more input object files contain IPA
information: specify -qipa for additional optimization.
ld: 0711-317 ERRORE: Simbolo non definito : spbtrf_
ld: 0711-317 ERRORE: Simbolo non definito : dpbtrf_
ld: 0711-317 ERRORE: Simbolo non definito : cpbtrf_
ld: 0711-317 ERRORE: Simbolo non definito : zpbtrf_
ld: 0711-317 ERRORE: Simbolo non definito : spbsv_
ld: 0711-317 ERRORE: Simbolo non definito : dpbsv_
ld: 0711-317 ERRORE: Simbolo non definito : cpbsv_
ld: 0711-317 ERRORE: Simbolo non definito : zpbsv_
ld: 0711-317 ERRORE: Simbolo non definito : sgebal_
ld: 0711-317 ERRORE: Simbolo non definito : dgebal_
ld: 0711-317 ERRORE: Simbolo non definito : cgebal_
ld: 0711-317 ERRORE: Simbolo non definito : zgebal_
ld: 0711-317 ERRORE: Simbolo non definito : sgehrd_
ld: 0711-317 ERRORE: Simbolo non definito : dgehrd_
ld: 0711-317 ERRORE: Simbolo non definito : cgehrd_
ld: 0711-317 ERRORE: Simbolo non definito : zgehrd_
ld: 0711-317 ERRORE: Simbolo non definito : sgbsv_
ld: 0711-317 ERRORE: Simbolo non definito : dgbsv_
ld: 0711-317 ERRORE: Simbolo non definito : cgbsv_
ld: 0711-317 ERRORE: Simbolo non definito : zgbsv_
ld: 0711-317 ERRORE: Simbolo non definito : sgesv_
ld: 0711-317 ERRORE: Simbolo non definito : dgesv_
ld: 0711-317 ERRORE: Simbolo non definito : cgesv_
ld: 0711-317 ERRORE: Simbolo non definito : zgesv_
ld: 0711-317 ERRORE: Simbolo non definito : sgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : dgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : cgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : zgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : sgetrs_
ld: 0711-317 ERRORE: Simbolo non definito : dgetrs_
ld: 0711-317 ERRORE: Simbolo non definito : cgetrs_
ld: 0711-317 ERRORE: Simbolo non definito : zgetrs_
ld: 0711-317 ERRORE: Simbolo non definito : sgetri_
ld: 0711-317 ERRORE: Simbolo non definito : dgetri_
ld: 0711-317 ERRORE: Simbolo non definito : cgetri_
ld: 0711-317 ERRORE: Simbolo non definito : zgetri_
ld: 0711-317 ERRORE: Simbolo non definito : sgesdd_
ld: 0711-317 ERRORE: Simbolo non definito : dgesdd_
ld: 0711-317 ERRORE: Simbolo non definito : cgesdd_
ld: 0711-317 ERRORE: Simbolo non definito : zgesdd_
ld: 0711-317 ERRORE: Simbolo non definito : sgelss_
ld: 0711-317 ERRORE: Simbolo non definito : dgelss_
ld: 0711-317 ERRORE: Simbolo non definito : cgelss_
ld: 0711-317 ERRORE: Simbolo non definito : zgelss_
ld: 0711-317 ERRORE: Simbolo non definito : sgeqrf_
ld: 0711-317 ERRORE: Simbolo non definito : dgeqrf_
ld: 0711-317 ERRORE: Simbolo non definito : cgeqrf_
ld: 0711-317 ERRORE: Simbolo non definito : zgeqrf_
ld: 0711-317 ERRORE: Simbolo non definito : sgerqf_
ld: 0711-317 ERRORE: Simbolo non definito : dgerqf_
ld: 0711-317 ERRORE: Simbolo non definito : cgerqf_
ld: 0711-317 ERRORE: Simbolo non definito : zgerqf_
ld: 0711-317 ERRORE: Simbolo non definito : sorgqr_
ld: 0711-317 ERRORE: Simbolo non definito : dorgqr_
ld: 0711-317 ERRORE: Simbolo non definito : cungqr_
ld: 0711-317 ERRORE: Simbolo non definito : zungqr_
ld: 0711-317 ERRORE: Simbolo non definito : sgeev_
ld: 0711-317 ERRORE: Simbolo non definito : dgeev_
ld: 0711-317 ERRORE: Simbolo non definito : cgeev_
ld: 0711-317 ERRORE: Simbolo non definito : zgeev_
ld: 0711-317 ERRORE: Simbolo non definito : sgegv_
ld: 0711-317 ERRORE: Simbolo non definito : dgegv_
ld: 0711-317 ERRORE: Simbolo non definito : cgegv_
ld: 0711-317 ERRORE: Simbolo non definito : zgegv_
ld: 0711-317 ERRORE: Simbolo non definito : ssyev_
ld: 0711-317 ERRORE: Simbolo non definito : dsyev_
ld: 0711-317 ERRORE: Simbolo non definito : cheev_
ld: 0711-317 ERRORE: Simbolo non definito : zheev_
ld: 0711-317 ERRORE: Simbolo non definito : sposv_
ld: 0711-317 ERRORE: Simbolo non definito : dposv_
ld: 0711-317 ERRORE: Simbolo non definito : cposv_
ld: 0711-317 ERRORE: Simbolo non definito : zposv_
ld: 0711-317 ERRORE: Simbolo non definito : spotrf_
ld: 0711-317 ERRORE: Simbolo non definito : dpotrf_
ld: 0711-317 ERRORE: Simbolo non definito : cpotrf_
ld: 0711-317 ERRORE: Simbolo non definito : zpotrf_
ld: 0711-317 ERRORE: Simbolo non definito : spotrs_
ld: 0711-317 ERRORE: Simbolo non definito : dpotrs_
ld: 0711-317 ERRORE: Simbolo non definito : cpotrs_
ld: 0711-317 ERRORE: Simbolo non definito : zpotrs_
ld: 0711-317 ERRORE: Simbolo non definito : spotri_
ld: 0711-317 ERRORE: Simbolo non definito : dpotri_
ld: 0711-317 ERRORE: Simbolo non definito : cpotri_
ld: 0711-317 ERRORE: Simbolo non definito : zpotri_
ld: 0711-317 ERRORE: Simbolo non definito : slauum_
ld: 0711-317 ERRORE: Simbolo non definito : dlauum_
ld: 0711-317 ERRORE: Simbolo non definito : clauum_
ld: 0711-317 ERRORE: Simbolo non definito : zlauum_
ld: 0711-317 ERRORE: Simbolo non definito : strtri_
ld: 0711-317 ERRORE: Simbolo non definito : dtrtri_
ld: 0711-317 ERRORE: Simbolo non definito : ctrtri_
ld: 0711-317 ERRORE: Simbolo non definito : ztrtri_
ld: 0711-317 ERRORE: Simbolo non definito : slaswp_
ld: 0711-317 ERRORE: Simbolo non definito : dlaswp_
ld: 0711-317 ERRORE: Simbolo non definito : claswp_
ld: 0711-317 ERRORE: Simbolo non definito : zlaswp_
ld: 0711-317 ERRORE: Simbolo non definito : cgees_
ld: 0711-317 ERRORE: Simbolo non definito : zgees_
ld: 0711-317 ERRORE: Simbolo non definito : dgees_
ld: 0711-317 ERRORE: Simbolo non definito : sgees_
ld: 0711-317 ERRORE: Simbolo non definito : sggev_
ld: 0711-317 ERRORE: Simbolo non definito : dggev_
ld: 0711-317 ERRORE: Simbolo non definito : cggev_
ld: 0711-317 ERRORE: Simbolo non definito : zggev_
ld: 0711-317 ERRORE: Simbolo non definito : ssbev_
ld: 0711-317 ERRORE: Simbolo non definito : dsbev_
ld: 0711-317 ERRORE: Simbolo non definito : ssbevd_
ld: 0711-317 ERRORE: Simbolo non definito : dsbevd_
ld: 0711-317 ERRORE: Simbolo non definito : ssbevx_
ld: 0711-317 ERRORE: Simbolo non definito : dsbevx_
ld: 0711-317 ERRORE: Simbolo non definito : chbevd_
ld: 0711-317 ERRORE: Simbolo non definito : zhbevd_
ld: 0711-317 ERRORE: Simbolo non definito : chbevx_
ld: 0711-317 ERRORE: Simbolo non definito : zhbevx_
ld: 0711-317 ERRORE: Simbolo non definito : sgbtrf_
ld: 0711-317 ERRORE: Simbolo non definito : dgbtrf_
ld: 0711-317 ERRORE: Simbolo non definito : cgbtrf_
ld: 0711-317 ERRORE: Simbolo non definito : zgbtrf_
ld: 0711-317 ERRORE: Simbolo non definito : sgbtrs_
ld: 0711-317 ERRORE: Simbolo non definito : dgbtrs_
ld: 0711-317 ERRORE: Simbolo non definito : cgbtrs_
ld: 0711-317 ERRORE: Simbolo non definito : zgbtrs_
ld: 0711-317 ERRORE: Simbolo non definito : .slamch_
ld: 0711-317 ERRORE: Simbolo non definito : .dlamch_
ld: 0711-317 ERRORE: Simbolo non definito : slamch_
ld: 0711-317 ERRORE: Simbolo non definito : dlamch_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.linalg._flinalg' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
compiling Fortran sources
Fortran f77 compiler: /usr/bin/xlf -qextname -O5
Fortran f90 compiler: /usr/bin/xlf90 -qextname -O5
Fortran fix compiler: /usr/bin/xlf90 -qfixed -qextname -O5
compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/scipy/linalg/_flinalgmodule.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/fortranobject.o
build/temp.aix-5.3-2.5/Lib/linalg/src/det.o
build/temp.aix-5.3-2.5/Lib/linalg/src/lu.o -L/home/Marek/apython/lib
-L/usr/lib -Lbuild/temp.aix-5.3-2.5 -llapack -lblas -o
build/lib.aix-5.3-2.5/scipy/linalg/_flinalg.so
/usr/bin/xlf95: 1501-262 One or more input object files contain IPA
information: specify -qipa for additional optimization.
ld: 0711-317 ERRORE: Simbolo non definito : .dgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : .sgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : .zgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : .cgetrf_
ld: 0711-317 ERRORE: Simbolo non definito : .dlaswp_
ld: 0711-317 ERRORE: Simbolo non definito : .zlaswp_
ld: 0711-317 ERRORE: Simbolo non definito : .slaswp_
ld: 0711-317 ERRORE: Simbolo non definito : .claswp_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.linalg.calc_lwork' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
compiling Fortran sources
Fortran f77 compiler: /usr/bin/xlf -qextname -O5
Fortran f90 compiler: /usr/bin/xlf90 -qextname -O5
Fortran fix compiler: /usr/bin/xlf90 -qfixed -qextname -O5
compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/scipy/linalg/calc_lworkmodule.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/fortranobject.o
build/temp.aix-5.3-2.5/Lib/linalg/src/calc_lwork.o
-L/home/Marek/apython/lib -L/usr/lib -Lbuild/temp.aix-5.3-2.5 -llapack
-lblas -o build/lib.aix-5.3-2.5/scipy/linalg/calc_lwork.so
/usr/bin/xlf95: 1501-262 One or more input object files contain IPA
information: specify -qipa for additional optimization.
ld: 0711-317 ERRORE: Simbolo non definito : .ilaenv_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.linalg._iterative' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
compiling Fortran sources
Fortran f77 compiler: /usr/bin/xlf -qextname -O5
Fortran f90 compiler: /usr/bin/xlf90 -qextname -O5
Fortran fix compiler: /usr/bin/xlf90 -qfixed -qextname -O5
compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/linalg/iterative/_iterativemodule.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/fortranobject.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/linalg/iterative/STOPTEST2.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/linalg/iterative/getbreak.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/linalg/iterative/BiCGREVCOM.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/linalg/iterative/BiCGSTABREVCOM.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/linalg/iterative/CGREVCOM.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/linalg/iterative/CGSREVCOM.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/linalg/iterative/GMRESREVCOM.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/linalg/iterative/QMRREVCOM.o
-L/home/Marek/apython/lib -L/usr/lib -Lbuild/temp.aix-5.3-2.5 -llapack
-lblas -o build/lib.aix-5.3-2.5/scipy/linalg/_iterative.so
/usr/bin/xlf95: 1501-262 One or more input object files contain IPA
information: specify -qipa for additional optimization.
ld: 0711-317 ERRORE: Simbolo non definito : .saxpy_
ld: 0711-317 ERRORE: Simbolo non definito : .sdot_
ld: 0711-317 ERRORE: Simbolo non definito : .scopy_
ld: 0711-317 ERRORE: Simbolo non definito : .snrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .slamch_
ld: 0711-317 ERRORE: Simbolo non definito : .daxpy_
ld: 0711-317 ERRORE: Simbolo non definito : .ddot_
ld: 0711-317 ERRORE: Simbolo non definito : .dcopy_
ld: 0711-317 ERRORE: Simbolo non definito : .dnrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .dlamch_
ld: 0711-317 ERRORE: Simbolo non definito : .caxpy_
ld: 0711-317 ERRORE: Simbolo non definito : .cdotc_
ld: 0711-317 ERRORE: Simbolo non definito : .ccopy_
ld: 0711-317 ERRORE: Simbolo non definito : .scnrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .zaxpy_
ld: 0711-317 ERRORE: Simbolo non definito : .zdotc_
ld: 0711-317 ERRORE: Simbolo non definito : .zcopy_
ld: 0711-317 ERRORE: Simbolo non definito : .dznrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .sscal_
ld: 0711-317 ERRORE: Simbolo non definito : .dscal_
ld: 0711-317 ERRORE: Simbolo non definito : .cscal_
ld: 0711-317 ERRORE: Simbolo non definito : .zscal_
ld: 0711-317 ERRORE: Simbolo non definito : .strsv_
ld: 0711-317 ERRORE: Simbolo non definito : .dtrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .ctrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .ztrsv_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.linsolve._zsuperlu' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -DUSE_VENDOR_BLAS=1
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/Lib/linsolve/_zsuperlumodule.o
build/temp.aix-5.3-2.5/Lib/linsolve/_superlu_utils.o
build/temp.aix-5.3-2.5/Lib/linsolve/_superluobject.o
-L/home/Marek/apython/lib -L/usr/lib -Lbuild/temp.aix-5.3-2.5
-lsuperlu_src -llapack -lblas -o
build/lib.aix-5.3-2.5/scipy/linsolve/_zsuperlu.so
ld: 0711-317 ERRORE: Simbolo non definito : .ztrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .zgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .ztrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .zgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .ctrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .cgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .ctrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .cgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .dtrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .dgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .dtrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .dgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .strsv_
ld: 0711-317 ERRORE: Simbolo non definito : .sgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .strsm_
ld: 0711-317 ERRORE: Simbolo non definito : .sgemm_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.linsolve._dsuperlu' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -DUSE_VENDOR_BLAS=1
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/Lib/linsolve/_dsuperlumodule.o
build/temp.aix-5.3-2.5/Lib/linsolve/_superlu_utils.o
build/temp.aix-5.3-2.5/Lib/linsolve/_superluobject.o
-L/home/Marek/apython/lib -L/usr/lib -Lbuild/temp.aix-5.3-2.5
-lsuperlu_src -llapack -lblas -o
build/lib.aix-5.3-2.5/scipy/linsolve/_dsuperlu.so
ld: 0711-317 ERRORE: Simbolo non definito : .ztrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .zgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .ztrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .zgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .ctrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .cgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .ctrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .cgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .dtrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .dgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .dtrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .dgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .strsv_
ld: 0711-317 ERRORE: Simbolo non definito : .sgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .strsm_
ld: 0711-317 ERRORE: Simbolo non definito : .sgemm_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.linsolve._csuperlu' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -DUSE_VENDOR_BLAS=1
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/Lib/linsolve/_csuperlumodule.o
build/temp.aix-5.3-2.5/Lib/linsolve/_superlu_utils.o
build/temp.aix-5.3-2.5/Lib/linsolve/_superluobject.o
-L/home/Marek/apython/lib -L/usr/lib -Lbuild/temp.aix-5.3-2.5
-lsuperlu_src -llapack -lblas -o
build/lib.aix-5.3-2.5/scipy/linsolve/_csuperlu.so
ld: 0711-317 ERRORE: Simbolo non definito : .ztrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .zgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .ztrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .zgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .ctrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .cgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .ctrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .cgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .dtrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .dgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .dtrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .dgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .strsv_
ld: 0711-317 ERRORE: Simbolo non definito : .sgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .strsm_
ld: 0711-317 ERRORE: Simbolo non definito : .sgemm_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.linsolve._ssuperlu' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -DUSE_VENDOR_BLAS=1
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/Lib/linsolve/_ssuperlumodule.o
build/temp.aix-5.3-2.5/Lib/linsolve/_superlu_utils.o
build/temp.aix-5.3-2.5/Lib/linsolve/_superluobject.o
-L/home/Marek/apython/lib -L/usr/lib -Lbuild/temp.aix-5.3-2.5
-lsuperlu_src -llapack -lblas -o
build/lib.aix-5.3-2.5/scipy/linsolve/_ssuperlu.so
ld: 0711-317 ERRORE: Simbolo non definito : .ztrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .zgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .ztrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .zgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .ctrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .cgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .ctrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .cgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .dtrsv_
ld: 0711-317 ERRORE: Simbolo non definito : .dgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .dtrsm_
ld: 0711-317 ERRORE: Simbolo non definito : .dgemm_
ld: 0711-317 ERRORE: Simbolo non definito : .strsv_
ld: 0711-317 ERRORE: Simbolo non definito : .sgemv_
ld: 0711-317 ERRORE: Simbolo non definito : .strsm_
ld: 0711-317 ERRORE: Simbolo non definito : .sgemm_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.odr.__odrpack' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -ILib/odr
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/Lib/odr/__odrpack.o -L/usr/lib
-Lbuild/temp.aix-5.3-2.5 -lodrpack -lblas -o
build/lib.aix-5.3-2.5/scipy/odr/__odrpack.so
ld: 0711-317 ERRORE: Simbolo non definito : .dcopy_
ld: 0711-317 ERRORE: Simbolo non definito : .ddot_
ld: 0711-317 ERRORE: Simbolo non definito : .dnrm2_
ld: 0711-317 ERRORE: Simbolo non definito : .daxpy_
ld: 0711-317 ERRORE: Simbolo non definito : .dswap_
ld: 0711-317 ERRORE: Simbolo non definito : .dscal_
ld: 0711-317 ERRORE: Simbolo non definito : .drotg_
ld: 0711-317 ERRORE: Simbolo non definito : .drot_
ld: 0711-317 ERRORE: Simbolo non definito : .dasum_
ld: 0711-317 ERRORE: Simbolo non definito : .idamax_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.
building 'scipy.optimize._lbfgsb' extension
compiling C sources
C compiler: cc_r -DNDEBUG -O

compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
compiling Fortran sources
Fortran f77 compiler: /usr/bin/xlf -qextname -O5
Fortran f90 compiler: /usr/bin/xlf90 -qextname -O5
Fortran fix compiler: /usr/bin/xlf90 -qfixed -qextname -O5
compile options: '-DNO_ATLAS_INFO=1 -Ibuild/src.aix-5.3-2.5
-I/home/Marek/apython/lib/python2.5/site-packages/numpy/core/include
-I/home/Marek/apython/include/python2.5 -c'
/home/Marek/apython/lib/python2.5/config/ld_so_aix /usr/bin/xlf95
-bI:/home/Marek/apython/lib/python2.5/config/python.exp
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/Lib/optimize/lbfgsb/_lbfgsbmodule.o
build/temp.aix-5.3-2.5/build/src.aix-5.3-2.5/fortranobject.o
build/temp.aix-5.3-2.5/Lib/optimize/lbfgsb/routines.o
-L/home/Marek/apython/lib -L/usr/lib -Lbuild/temp.aix-5.3-2.5 -llapack
-lblas -o build/lib.aix-5.3-2.5/scipy/optimize/_lbfgsb.so
/usr/bin/xlf95: 1501-262 One or more input object files contain IPA
information: specify -qipa for additional optimization.
ld: 0711-317 ERRORE: Simbolo non definito : .dcopy_
ld: 0711-317 ERRORE: Simbolo non definito : .dscal_
ld: 0711-317 ERRORE: Simbolo non definito : .dtrtrs_
ld: 0711-317 ERRORE: Simbolo non definito : .ddot_
ld: 0711-317 ERRORE: Simbolo non definito : .daxpy_
ld: 0711-317 ERRORE: Simbolo non definito : .dpotrf_
ld: 0711-345 Usare l'opzione -bloadmap o -bnoquiet per ulteriori
informazioni.






From david at ar.media.kyoto-u.ac.jp  Sat Jul 28 07:51:22 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Sat, 28 Jul 2007 20:51:22 +0900
Subject: [SciPy-user] Numpy/Scipy on AIX
In-Reply-To: <33682.147.162.29.65.1185623222.squirrel@poczta.p.lodz.pl>
References: <33682.147.162.29.65.1185623222.squirrel@poczta.p.lodz.pl>
Message-ID: <46AB2DBA.4020601@ar.media.kyoto-u.ac.jp>

Marek Wojciechowski wrote:
> David Cournapeau <david <at> ar.media.kyoto-u.ac.jp> writes:
>
>   
>> Marek wrote:
>>     
>>> Thanks for your response, but I skipped the problem downloading svn
>>>       
> version
> of
>   
>>> numpy (previously i tried to install numpy-1.0.2), and this compiles
>>>       
> nicely.
>   
>>> But now i run into troubles with scipy. Namely, I get tons of
>>>       
> unresolved
>   
>>> symbols from lapack, for example:
>>>
>>> ld: 0711-317 ERROR: Undefined symbol: .sgemm_
>>>
>>> Any workaround for this? I know there is linker
>>>       
> option -brename:.sgemm_,.sgemm,
>   
>>> which can solve it but this si not a nice solution.
>>>
>>>
>>>       
>> We'll need more informations to help you. The most useful thing would be
>>     
>
>   
>> to know the output of python setup.py config for numpy and scipy.
>>
>> Also, did you used the release of scipy (0.5.2) or the svn version as
>>     
> well ?
>   
>> David
>>
>>     
>
> I'm using svn version of scipy. As far as i observed scipy 0.5.2 is not 
> compatibile with numpy-svn. Below i'm pasting what setup.py produces. I
> seems things are compiled and there is only linking problem with lapack.
> Thanks for help.
>   
Where are your blas and lapack coming from ? Did you compile them by 
yourself ? How ? (compiler, version, etc...)

David


From mwojc at p.lodz.pl  Sat Jul 28 10:41:03 2007
From: mwojc at p.lodz.pl (Marek Wojciechowski)
Date: Sat, 28 Jul 2007 16:41:03 +0200 (CEST)
Subject: [SciPy-user] Numpy/Scipy on AIX
In-Reply-To: <mailman.41279.1185623245.30507.scipy-user@scipy.org>
References: <mailman.41279.1185623245.30507.scipy-user@scipy.org>
Message-ID: <34352.147.162.29.65.1185633663.squirrel@poczta.p.lodz.pl>

> > I'm using svn version of scipy. As far as i observed scipy 0.5.2 is not
> > compatibile with numpy-svn. Below i'm pasting what setup.py produces. I
> > seems things are compiled and there is only linking problem with lapack.
> > Thanks for help.
> >
> Where are your blas and lapack coming from ? Did you compile them by
> yourself ? How ? (compiler, version, etc...)
>

I'm using IBM(R) XL Fortran Enterprise Edition V9.1 and IBM(R) XL C
Enterprise Edition V7.0. Maybe there is some incompatibility between
compilers for Fortran and C? That's what is installed on the machine i'm
working on. I have to compile lapack by myself (actually scipy installer
can do that), this is lapack-lite-3.1.1 downloaded from the web, and blas
is the one shipped with XL Fortran. Are there some obvious
incompatibilities?

Thanks
Marek




From fredmfp at gmail.com  Sat Jul 28 19:03:07 2007
From: fredmfp at gmail.com (fred)
Date: Sun, 29 Jul 2007 01:03:07 +0200
Subject: [SciPy-user] using imread...
In-Reply-To: <loom.20070727T134841-817@post.gmane.org>
References: <46A8FE96.40404@gmail.com> <20070727060221.GA15345@clipper.ens.fr>
	<loom.20070727T134841-817@post.gmane.org>
Message-ID: <46ABCB2B.5070602@gmail.com>

jelle a ?crit :
> pil 1.1.6 supports numpy arrays, perhaps its even more efficent to use this?
>
>   
Well, I think I need to learn a little bit more about using gimp before...

Thanks anyway.

Cheers,

-- 
http://scipy.org/FredericPetit



From david at ar.media.kyoto-u.ac.jp  Sun Jul 29 05:49:32 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Sun, 29 Jul 2007 18:49:32 +0900
Subject: [SciPy-user] Numpy/Scipy on AIX
In-Reply-To: <34352.147.162.29.65.1185633663.squirrel@poczta.p.lodz.pl>
References: <mailman.41279.1185623245.30507.scipy-user@scipy.org>
	<34352.147.162.29.65.1185633663.squirrel@poczta.p.lodz.pl>
Message-ID: <46AC62AC.3000500@ar.media.kyoto-u.ac.jp>

Marek Wojciechowski wrote:
>>> I'm using svn version of scipy. As far as i observed scipy 0.5.2 is not
>>> compatibile with numpy-svn. Below i'm pasting what setup.py produces. I
>>> seems things are compiled and there is only linking problem with lapack.
>>> Thanks for help.
>>>
>> Where are your blas and lapack coming from ? Did you compile them by
>> yourself ? How ? (compiler, version, etc...)
>>
>
> I'm using IBM(R) XL Fortran Enterprise Edition V9.1 and IBM(R) XL C
> Enterprise Edition V7.0. Maybe there is some incompatibility between
> compilers for Fortran and C? That's what is installed on the machine i'm
> working on. I have to compile lapack by myself (actually scipy installer
> can do that), 
scipy cannot compile lapack by itself. I don't know the exact process in 
this case, but scipy has a "lite" lapack as a fall back; I am not 100 % 
sure, though.

There is a pretty good chance that the LAPACK you compiled has something 
wrong: LAPACK is difficult to compile correctly when you want to use it 
in a non Fortran environment; at least it is on Linux. Did you test LAPACK ?

LAPACK 3.1.1 causes problems with scipy at the moment, too, so I 
recommend you to compile the version 3.0 (beware though, you should NOT 
use the BLAS from LAPACK 3.0, because it is buggy). If I were you, I 
would do this first: compile LAPACK 3.0 with the AIX BLAS, and compile 
the testers and run them (doing a make lapack_testing should be enough).

If this works, then the problem is either in the calling convention 
(python calls fortran code "through C", so you need to be sure that C 
and Fortran convention are the same; I would guess that AIX being 
proprietary and used for HPC, IBM tries to have compatible convention by 
default), or in scipy. If this is the case, we can focus on this point.

Cheers,

David


From pearu at cens.ioc.ee  Sun Jul 29 11:33:55 2007
From: pearu at cens.ioc.ee (Pearu Peterson)
Date: Sun, 29 Jul 2007 17:33:55 +0200
Subject: [SciPy-user] Numpy/Scipy on AIX
In-Reply-To: <46AC62AC.3000500@ar.media.kyoto-u.ac.jp>
References: <mailman.41279.1185623245.30507.scipy-user@scipy.org>	<34352.147.162.29.65.1185633663.squirrel@poczta.p.lodz.pl>
	<46AC62AC.3000500@ar.media.kyoto-u.ac.jp>
Message-ID: <46ACB363.9050601@cens.ioc.ee>



David Cournapeau wrote:

> scipy cannot compile lapack by itself.

yes, it can. Just define env. variables BLAS_SRC, LAPACK_SRC
pointing to blas, lapack source directories, respectively.
If there are some blas or lapack libraries lying
around in the system  then one must also set BLAS=None, LAPACK=None, 
ATLAS=None so that numpy.distutils will not pick up the unwanted libraries.

 > I don't know the exact process in
> this case, but scipy has a "lite" lapack as a fall back; I am not 100 % 
> sure, though.

Did you mean numpy? numpy has lite lapack, not scipy.

> There is a pretty good chance that the LAPACK you compiled has something 
> wrong: LAPACK is difficult to compile correctly when you want to use it 
> in a non Fortran environment; at least it is on Linux. Did you test LAPACK ?
> 
> LAPACK 3.1.1 causes problems with scipy at the moment, too, so I 
> recommend you to compile the version 3.0 (beware though, you should NOT 
> use the BLAS from LAPACK 3.0, because it is buggy). If I were you, I 
> would do this first: compile LAPACK 3.0 with the AIX BLAS, and compile 
> the testers and run them (doing a make lapack_testing should be enough).

I have fixed the problems with LAPACK 3.1.1 recently. Just use numpy
from SVN and scipy should compile fine with LAPACK 3.1.1.

Pearu


From david at ar.media.kyoto-u.ac.jp  Mon Jul 30 01:27:34 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Mon, 30 Jul 2007 14:27:34 +0900
Subject: [SciPy-user] Numpy/Scipy on AIX
In-Reply-To: <46ACB363.9050601@cens.ioc.ee>
References: <mailman.41279.1185623245.30507.scipy-user@scipy.org>	<34352.147.162.29.65.1185633663.squirrel@poczta.p.lodz.pl>	<46AC62AC.3000500@ar.media.kyoto-u.ac.jp>
	<46ACB363.9050601@cens.ioc.ee>
Message-ID: <46AD76C6.3040206@ar.media.kyoto-u.ac.jp>

Pearu Peterson wrote:
>
> David Cournapeau wrote:
>
>> scipy cannot compile lapack by itself.
>
> yes, it can. Just define env. variables BLAS_SRC, LAPACK_SRC
> pointing to blas, lapack source directories, respectively.
> If there are some blas or lapack libraries lying
> around in the system  then one must also set BLAS=None, LAPACK=None, 
> ATLAS=None so that numpy.distutils will not pick up the unwanted libraries.
Wow, I didn't see this feature while I packaged scipy. Sorry for the 
misinformation.
>
>> There is a pretty good chance that the LAPACK you compiled has something 
>> wrong: LAPACK is difficult to compile correctly when you want to use it 
>> in a non Fortran environment; at least it is on Linux. Did you test LAPACK ?
>>
>> LAPACK 3.1.1 causes problems with scipy at the moment, too, so I 
>> recommend you to compile the version 3.0 (beware though, you should NOT 
>> use the BLAS from LAPACK 3.0, because it is buggy). If I were you, I 
>> would do this first: compile LAPACK 3.0 with the AIX BLAS, and compile 
>> the testers and run them (doing a make lapack_testing should be enough).
>
> I have fixed the problems with LAPACK 3.1.1 recently. Just use numpy
> from SVN and scipy should compile fine with LAPACK 3.1.1.
Do all the tests pass now ? As far as I remembered (but this was a few 
months ago), building with LAPACK 3.1.1 is possible, but some tests 
fail. I don't have the time right now to recompile my own lapack to 
reproduce the problem, but I can do it within the new few days if you 
need more precise informations.

David


From elcorto at gmx.net  Mon Jul 30 02:22:31 2007
From: elcorto at gmx.net (Steve Schmerler)
Date: Mon, 30 Jul 2007 08:22:31 +0200
Subject: [SciPy-user] Memory problems with stats.mannwhitneyu
In-Reply-To: <46AA3CD7.4060504@gmail.com>
References: <dc90fba40707261341u42fbb603x9171c869768185f2@mail.gmail.com>	<46A917CA.2040402@gmail.com>	<46A91C7C.8010604@sci.utah.edu>	<46A9E78F.6060107@gmx.net>	<dc90fba40707270933t76ce2082k864e8eaf941b95ff@mail.gmail.com>
	<46AA3CD7.4060504@gmail.com>
Message-ID: <46AD83A7.90001@gmx.net>

Robert Kern wrote:
> Martin Blom wrote:
>> I'm on an Ubuntu Feisty machine, and the python installation is the
>> standard one for feisty. Machine and python details:
>>
>>
>> ~$ python
>> Python 2.5.1 (r251:54863, May  2 2007, 16:56:35)
>> [GCC 4.1.2 (Ubuntu 4.1.2-0ubuntu4 )] on linux2
>> Type "help", "copyright", "credits" or "license" for more information.
>>>>> import scipy
>>>>> scipy.__version__
>> '0.5.2'
>>>>> import numpy
>>>>> numpy.__version__
>> '1.0.1'
> 
> Well, you all seem to have numpy 1.0.1 in common. I do not see this on my AMD 64
> Ubuntu box with numpy SVN. I suspect a bug in numpy 1.0.1 and suggest upgrading.

Yes, the problem isn't there with numpy 1.0.4.dev3869, scipy
0.5.3.dev3105.

-- 
cheers,
steve

I love deadlines. I like the whooshing sound they make as they fly by. --
Douglas Adams



From robert.vergnes at yahoo.fr  Mon Jul 30 10:37:46 2007
From: robert.vergnes at yahoo.fr (Robert VERGNES)
Date: Mon, 30 Jul 2007 16:37:46 +0200 (CEST)
Subject: [SciPy-user] odeint memoryError
Message-ID: <106788.73226.qm@web27402.mail.ukl.yahoo.com>

Hello,

when I have a time vector too thin or long (...), then odeint crash and returns a "MemoryError". 

Apart from splitting my calculation in parts is there some ways around ? ie like a wrapper dealing with this exception?


Best Regards

Robert

       
---------------------------------
 Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070730/5c2a0c15/attachment.html>

From fredmfp at gmail.com  Mon Jul 30 11:26:18 2007
From: fredmfp at gmail.com (fred)
Date: Mon, 30 Jul 2007 17:26:18 +0200
Subject: [SciPy-user] using imread... [fixed]
In-Reply-To: <loom.20070727T134841-817@post.gmane.org>
References: <46A8FE96.40404@gmail.com> <20070727060221.GA15345@clipper.ens.fr>
	<loom.20070727T134841-817@post.gmane.org>
Message-ID: <46AE031A.4050803@gmail.com>

jelle a ?crit :
> pil 1.1.6 supports numpy arrays, perhaps its even more efficent to use this?
>   
Well, the problem was only between the keyboard and the chair...

Cheers,

-- 
http://scipy.org/FredericPetit



From rhc28 at cornell.edu  Mon Jul 30 14:32:43 2007
From: rhc28 at cornell.edu (Rob Clewley)
Date: Mon, 30 Jul 2007 14:32:43 -0400
Subject: [SciPy-user] odeint memoryError
In-Reply-To: <106788.73226.qm@web27402.mail.ukl.yahoo.com>
References: <106788.73226.qm@web27402.mail.ukl.yahoo.com>
Message-ID: <a749952d0707301132pce1cc6an85cde3d8166fecac@mail.gmail.com>

Robert,

You need to post more information about your system of ODEs as there
are many reasons (some even mathematical) why you might be running out
of memory. You may need to adjust some algorithmic parameters such as
step size or accuracy tolerances.

-Rob

On 30/07/07, Robert VERGNES <robert.vergnes at yahoo.fr> wrote:
> Hello,
>
> when I have a time vector too thin or long (...), then odeint crash and
> returns a "MemoryError".
>
> Apart from splitting my calculation in parts is there some ways around ? ie
> like a wrapper dealing with this exception?
>
>
> Best Regards
>
> Robert
>
>
>  ________________________________
>  Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo!
> Mail
>
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>


From Alexander.Dietz at astro.cf.ac.uk  Mon Jul 30 16:01:35 2007
From: Alexander.Dietz at astro.cf.ac.uk (Alexander Dietz)
Date: Mon, 30 Jul 2007 21:01:35 +0100
Subject: [SciPy-user] Problems installing scipy
Message-ID: <9cf809a00707301301x53086d46q8f4242d3876545bb@mail.gmail.com>

Hi,

I am trying to install scipy on my laptop, running FC5. I have installed
numpy, so I downloaded the scipy-package and run the setup script which
creates a warning only:

> python setup.py install
usr/lib/python2.4/distutils/dist.py:236: UserWarning: Unknown distribution
option: 'configuration'
  warnings.warn(msg)
running install
running build
running config_fc

When testing the installation within (i)python I get:

In [1]: import scipy
---------------------------------------------------------------------------
exceptions.ImportError                               Traceback (most recent
call last)

/opt/users/src/scipy-0.5.2/<ipython console>

ImportError: No module named scipy


I tried also to scipy to install in the directory
/usr/lib/python2.4/site-packages where numpy is installed (at least that's
what numpy.__path__ is saying) with
 python setup.py install --prefix=/usr/lib/python2.4/site-packages

but this also did not install scipy so I can use it. Can someone help me?
What's going wrong? In which directory is scipy installed if no path is
defined?

Thanks
  Alex
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070730/ed9623d2/attachment.html>

From david at ar.media.kyoto-u.ac.jp  Mon Jul 30 22:58:05 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Tue, 31 Jul 2007 11:58:05 +0900
Subject: [SciPy-user] Problems installing scipy
In-Reply-To: <9cf809a00707301301x53086d46q8f4242d3876545bb@mail.gmail.com>
References: <9cf809a00707301301x53086d46q8f4242d3876545bb@mail.gmail.com>
Message-ID: <46AEA53D.5090306@ar.media.kyoto-u.ac.jp>

Alexander Dietz wrote:
> Hi,
>
> I am trying to install scipy on my laptop, running FC5. I have 
> installed numpy, so I downloaded the scipy-package and run the setup 
> script which creates a warning only:
>
Note that to avoid all this hassle, there are some rpm for FC5 for numpy 
and scipy:

http://download.opensuse.org/repositories/home:/ashigabou/Fedora_Extras_5/
> > python setup.py install
> usr/lib/python2.4/distutils/dist.py:236: UserWarning: Unknown 
> distribution option: 'configuration'
>   warnings.warn(msg)
> running install
> running build
> running config_fc
>
> When testing the installation within (i)python I get:
>
> In [1]: import scipy
> --------------------------------------------------------------------------- 
>
> exceptions.ImportError                               Traceback (most 
> recent call last)
>
> /opt/users/src/scipy-0.5.2/<ipython console>
>
> ImportError: No module named scipy
>
>
> I tried also to scipy to install in the directory 
> /usr/lib/python2.4/site-packages where numpy is installed (at least 
> that's what numpy.__path__ is saying) with
>  python setup.py install --prefix=/usr/lib/python2.4/site-packages    
>
You did not give enough details, but the line above suggests that you 
are making a mistake concerning where the packages are installed.

If you do python setup.py install -> install in the default 
site-packages of your python installation, in this case 
/usr/lib/python2.4/site-packages

if you do python setup.py install --prefix=SOMEPATH, this will install 
the packages in SOMEPATH/python2.4/site-packages (assuming your python 
is 2.4, of course).

So the path you give to the --prefix option and the one where python is 
looking at are different. For example, not having administrative rights 
on my machine, I install everything in my $HOME/local directory:

python setup.py install --prefix=$HOME/local

And the add $HOME/local/python2.4/site-packages to my PYTHONPATH.

cheers,

David


From v-nijs at kellogg.northwestern.edu  Tue Jul 31 00:24:53 2007
From: v-nijs at kellogg.northwestern.edu (Vincent Nijs)
Date: Mon, 30 Jul 2007 23:24:53 -0500
Subject: [SciPy-user] google group
In-Reply-To: <46AEA53D.5090306@ar.media.kyoto-u.ac.jp>
Message-ID: <C2D423C5.876B%v-nijs@kellogg.northwestern.edu>

Is there a reason why the numpy list is available as a Google group and the
scipy list is not? It would be great to have both available as Google
groups.

Vincent




From robert.kern at gmail.com  Tue Jul 31 00:31:27 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Mon, 30 Jul 2007 23:31:27 -0500
Subject: [SciPy-user] google group
In-Reply-To: <C2D423C5.876B%v-nijs@kellogg.northwestern.edu>
References: <C2D423C5.876B%v-nijs@kellogg.northwestern.edu>
Message-ID: <46AEBB1F.8030507@gmail.com>

Vincent Nijs wrote:
> Is there a reason why the numpy list is available as a Google group and the
> scipy list is not? It would be great to have both available as Google
> groups.

Someone decided to request a Google Groups gateway for numpy-discussion, but
they didn't care about the scipy lists. If you would like to make such a
gateway, please do! Just be sure to pay attention to it. The first I had heard
about the Google Groups gateway for numpy-discussion was just after we moved the
list from sourceforge.net, and the Google Groups info hadn't been updated!

-- 
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 topengineer at gmail.com  Tue Jul 31 02:15:40 2007
From: topengineer at gmail.com (HuiChang Moon)
Date: Tue, 31 Jul 2007 15:15:40 +0900
Subject: [SciPy-user] I want to improve the speed of eig().
Message-ID: <296323b50707302315l580fe2adgabd1be3b338846af@mail.gmail.com>

Hi, users,
I have a problem with calculating the eigenvalue problem using the eig().
The array I want to calculate is about 5000*5000 but, the array is very
sparse.
At each row it has only 3 non-zero items, and the matrix is diagonal.
I have to solve eigenvalue problem, but it takes very long time.
How can I do?
I try to use sparse matix in Scipy, but eig() function doent' work with
sparse_matrix.
Please help me~!
Thank you.

-- 
Nano SOI Process Lab.
Hanyang University.

Contacts;
smartmoon at hanyang.ac.kr
HeeChang.Moon at Gmail.com
+82-2-2220-0247
+82-10-6455-7444

________________________________________
Dream but be awake.
________________________________________
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070731/043b5ba9/attachment.html>

From robert.vergnes at yahoo.fr  Tue Jul 31 02:31:59 2007
From: robert.vergnes at yahoo.fr (Robert VERGNES)
Date: Tue, 31 Jul 2007 08:31:59 +0200 (CEST)
Subject: [SciPy-user] RE : Re:  odeint memoryError
In-Reply-To: <a749952d0707301132pce1cc6an85cde3d8166fecac@mail.gmail.com>
Message-ID: <793925.17944.qm@web27407.mail.ukl.yahoo.com>

Dear Rob,

It is related with step size and accuracy. Is there some way I could know if ODEINT will crash or not ? - hence I would adapt the length of my t_vector...


an example of crashing odeint is below:

from scipy import *
from numpy import *
from scipy.integrate import *

# Equation Parameters
t_step= 1E-12
t_start=1.0172
t_end = 1.01726095276
t_interim = t_range = arange(t_start, t_end,t_step)
print 'Length of t_range: ' + str(len(t_range))
print 'Length should be around: '+ str(((t_end-t_start)/t_step))
w0 =sqrt (10)
fo = 0.075
y0_interim = [-2.20182185029, 0.0120342071085] #[Theta_dot, Theta]

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

if __name__ == '__main__':
    # ODE calculation
    y, infodict = odeint(func_f3,y0_interim,t_interim, full_output =1)
    print y
    print infodict



Rob Clewley <rhc28 at cornell.edu> a ?crit : Robert,

You need to post more information about your system of ODEs as there
are many reasons (some even mathematical) why you might be running out
of memory. You may need to adjust some algorithmic parameters such as
step size or accuracy tolerances.

-Rob

On 30/07/07, Robert VERGNES  wrote:
> Hello,
>
> when I have a time vector too thin or long (...), then odeint crash and
> returns a "MemoryError".
>
> Apart from splitting my calculation in parts is there some ways around ? ie
> like a wrapper dealing with this exception?
>
>
> Best Regards
>
> Robert
>
>
>  ________________________________
>  Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo!
> Mail
>
>
> _______________________________________________
> 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


       
---------------------------------
 Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070731/be4ef717/attachment.html>

From nwagner at iam.uni-stuttgart.de  Tue Jul 31 03:26:12 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Tue, 31 Jul 2007 09:26:12 +0200
Subject: [SciPy-user] I want to improve the speed of eig().
In-Reply-To: <296323b50707302315l580fe2adgabd1be3b338846af@mail.gmail.com>
References: <296323b50707302315l580fe2adgabd1be3b338846af@mail.gmail.com>
Message-ID: <46AEE414.9000501@iam.uni-stuttgart.de>

HuiChang Moon wrote:
> Hi, users,
> I have a problem with calculating the eigenvalue problem using the eig().
> The array I want to calculate is about 5000*5000 but, the array is
> very sparse.
> At each row it has only 3 non-zero items, and the matrix is diagonal.
A diagonal matrix has only one entry per row. I guess you mean a
tridiagonal matrix ?
eig is restricted to dense arrays.
You may try arpack or lobpcg. It depends on the properties of your matrix
(symmetrie, definiteness, ...)
>>> from scipy.sandbox import lobpcg
>>> help (lobpcg)

>>> from scipy.sandbox import arpack
>>> help (arpack)

Can you send your matrix (in matrixmarket format) off-list ?

HTH
            Nils


> I have to solve eigenvalue problem, but it takes very long time.
> How can I do?
> I try to use sparse matix in Scipy, but eig() function doent' work
> with sparse_matrix.
> Please help me~!
> Thank you.
>
> -- 
> Nano SOI Process Lab.
> Hanyang University.
>
> Contacts;
> smartmoon at hanyang.ac.kr <mailto:smartmoon at hanyang.ac.kr>
> HeeChang.Moon at Gmail.com <mailto:HeeChang.Moon at Gmail.com>
> +82-2-2220-0247
> +82-10-6455-7444
>
> ________________________________________
> Dream but be awake.
> ________________________________________
> ------------------------------------------------------------------------
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>   


 



From elcorto at gmx.net  Tue Jul 31 05:09:55 2007
From: elcorto at gmx.net (Steve Schmerler)
Date: Tue, 31 Jul 2007 11:09:55 +0200
Subject: [SciPy-user] odeint memoryError
In-Reply-To: <793925.17944.qm@web27407.mail.ukl.yahoo.com>
References: <793925.17944.qm@web27407.mail.ukl.yahoo.com>
Message-ID: <46AEFC63.9090806@gmx.net>

Robert VERGNES wrote:
> Dear Rob,
> 
> It is related with step size and accuracy. Is there some way I could
> know if ODEINT will crash or not ? - hence I would adapt the length of
> my t_vector...
> 
> 
> an example of crashing odeint is below:
> 
> from scipy import *
> from numpy import *
> from scipy.integrate import *
> 
> # Equation Parameters
> t_step= 1E-12
> t_start=1.0172
> t_end = 1.01726095276
> t_interim = t_range = arange(t_start, t_end,t_step)
> print 'Length of t_range: ' + str(len(t_range))
> print 'Length should be around: '+ str(((t_end-t_start)/t_step))
> w0 =sqrt (10)
> fo = 0.075
> y0_interim = [-2.20182185029, 0.0120342071085] #[Theta_dot, Theta]
> 
> def func_f3(y,t):
>     """  Jacobian function """
>     return [ - fo*y[0]- w0**2*sin(y[1]), y[0]]
> 
> if __name__ == '__main__':
>     # ODE calculation
>     y, infodict = odeint(func_f3,y0_interim,t_interim, full_output =1)
>     print y
>     print infodict
> 

With t_step= 1E-12, you t_range array has a size of about 6e7 floats, which is
about 460 MB of memory for t_range alone. With 2 additional arrays for y[0] and
y[1], you will run out of memory or at least force extensive swap usage. On my
machine with 512 MB RAM I also get a MemoryError. Are you sure you need such a
high resolution? Usually I'm happy with, say 50000 points. Like this:

# Equation Parameters
##t_step= 1E-12
npts=50000
t_start=1.0172
t_end = 1.01726095276
##t_interim = t_range = arange(t_start, t_end,t_step)
t_interim = t_range = linspace(t_start, t_end, npts)
print 't_step: %e' %(t_range[-1]-t_range[-2])
print 'Length of t_range: ' + str(len(t_range))
print 'Length should be around: '+ str(((t_end-t_start)/t_step))
.....

HTH

-- 
cheers,
steve

I love deadlines. I like the whooshing sound they make as they fly by. --
Douglas Adams



From nwagner at iam.uni-stuttgart.de  Tue Jul 31 05:23:47 2007
From: nwagner at iam.uni-stuttgart.de (Nils Wagner)
Date: Tue, 31 Jul 2007 11:23:47 +0200
Subject: [SciPy-user] [Fwd: Re:  I want to improve the speed of eig().]
Message-ID: <46AEFFA3.1070600@iam.uni-stuttgart.de>

Ok. I am using the svn version of scipy on linux and I am not familiar
with Windows. Since arpack and lobpcg are still in the sandbox (bleeding edge http://en.wikipedia.org/wiki/Bleeding_edge) I can't give
you further advice.
Anyway it might be a good idea to move arpack and lobpcg into the main tree.
Comments would be appreciated.

However you can send me your matrix off-list. Which eigenvalues do you need ?
I mean you can't compute all of them in case of large sparse matrices.
How about the corresponding eigenvectors ?
Is your matrix real and symmetric ?

Cheers,
                      Nils



-------------- next part --------------
An embedded message was scrubbed...
From: "HuiChang Moon" <topengineer at gmail.com>
Subject: Re: [SciPy-user] I want to improve the speed of eig().
Date: Tue, 31 Jul 2007 18:02:04 +0900
Size: 9735
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070731/ef7e8435/attachment.mht>

From rademacherchristoph at googlemail.com  Tue Jul 31 09:25:59 2007
From: rademacherchristoph at googlemail.com (Christoph Rademacher)
Date: Tue, 31 Jul 2007 15:25:59 +0200
Subject: [SciPy-user] restricting optimize.leastsq to positive results
Message-ID: <C64F545A-CC84-4094-8D12-EA19E3C0FC76@gmail.com>

Hi all,

I am using optimize.leastsq to fit my experimental data. Is there a  
way to restrict the results to be a positive float?
e.g. my target function is a simple linear combination of three  
variables to fit
f(x) : return a*A(x) + b*B(x)

 From my experiment I know that a,b,A,B can only be positive floats.  
How do I put this extra information in optimize.leastsq?

Thanks for your help,

Christoph



From v-nijs at kellogg.northwestern.edu  Tue Jul 31 10:51:29 2007
From: v-nijs at kellogg.northwestern.edu (Vincent Nijs)
Date: Tue, 31 Jul 2007 09:51:29 -0500
Subject: [SciPy-user] google group
In-Reply-To: <46AEBB1F.8030507@gmail.com>
Message-ID: <C2D4B6A1.877A%v-nijs@kellogg.northwestern.edu>

Robert,

I can't find any information about setting up a 'gateway' to the scipy list
in Google groups. Making a new group looks easy but that is not the idea.

Do you know who setup the numpy group?

Vincent


On 7/30/07 11:31 PM, "Robert Kern" <robert.kern at gmail.com> wrote:

> Vincent Nijs wrote:
>> Is there a reason why the numpy list is available as a Google group and the
>> scipy list is not? It would be great to have both available as Google
>> groups.
> 
> Someone decided to request a Google Groups gateway for numpy-discussion, but
> they didn't care about the scipy lists. If you would like to make such a
> gateway, please do! Just be sure to pay attention to it. The first I had heard
> about the Google Groups gateway for numpy-discussion was just after we moved
> the
> list from sourceforge.net, and the Google Groups info hadn't been updated!




From scyang at nist.gov  Tue Jul 31 11:30:07 2007
From: scyang at nist.gov (Stephen Yang)
Date: Tue, 31 Jul 2007 11:30:07 -0400
Subject: [SciPy-user] optimize.leastsq ier flag
Message-ID: <46AF557F.7000707@nist.gov>

Hello everyone,

I'm getting an ier flag of '4' when I run this code to fit some data to 
a gaussian:

p0 = [1, 1, 1]

gauss = lambda p, x: p[0]*numpy.exp(-((x-p[1])**2/(2*p[2]**2)))
res = lambda p, x, y: gauss(p, x) - y
optimized, success = optimize.minpack.leastsq(res, p0, args = (omega, 
counts))
print 'success flag:', success

In the minpack documentation, there is supposedly a 'mesg' return value, 
which states:

    mesg -- a string message giving information about the cause of failure.

So.. first, does anyone know what a flag of '4' means? Secondly, how 
would you go about getting this flag? I tried

optimized, mesg, success = optimize.minpack.leastsq(res, p0, args = 
(omega, counts))

with no luck..
ValueError: need more than 2 values to unpack

Any help would be greatly appreciated.





From robert.kern at gmail.com  Tue Jul 31 12:02:40 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 31 Jul 2007 11:02:40 -0500
Subject: [SciPy-user] google group
In-Reply-To: <C2D4B6A1.877A%v-nijs@kellogg.northwestern.edu>
References: <C2D4B6A1.877A%v-nijs@kellogg.northwestern.edu>
Message-ID: <46AF5D20.9030809@gmail.com>

Vincent Nijs wrote:
> Robert,
> 
> I can't find any information about setting up a 'gateway' to the scipy list
> in Google groups. Making a new group looks easy but that is not the idea.

http://groups.google.com/support/bin/answer.py?answer=46387&topic=9249

> Do you know who setup the numpy group?

Mike McLay, if I remember correctly.

-- 
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 Alexander.Dietz at astro.cf.ac.uk  Tue Jul 31 14:29:57 2007
From: Alexander.Dietz at astro.cf.ac.uk (Alexander Dietz)
Date: Tue, 31 Jul 2007 19:29:57 +0100
Subject: [SciPy-user] Problems installing scipy
In-Reply-To: <46AEA53D.5090306@ar.media.kyoto-u.ac.jp>
References: <9cf809a00707301301x53086d46q8f4242d3876545bb@mail.gmail.com>
	<46AEA53D.5090306@ar.media.kyoto-u.ac.jp>
Message-ID: <9cf809a00707311129n64475320w1380d58832c80cc4@mail.gmail.com>

Hi,

thanks for the tipps bot none of them are working!

On 7/31/07, David Cournapeau <david at ar.media.kyoto-u.ac.jp > wrote:
>
> Alexander Dietz wrote:
> > Hi,
> >
> > I am trying to install scipy on my laptop, running FC5. I have
> > installed numpy, so I downloaded the scipy-package and run the setup
> > script which creates a warning only:
> >
> Note that to avoid all this hassle, there are some rpm for FC5 for numpy
> and scipy:
>
> http://download.opensuse.org/repositories/home:/ashigabou/Fedora_Extras_5/


I tried to install the rpm package, but, as I expected before, several other
packages are required to install scipy:

error: Failed dependencies:
        libblas.so.3 is needed by python-scipy-0.5.2-14.1.i386
        liblapack.so.3 is needed by python-scipy-0.5.2-14.1.i386
        python-numpy is needed by python-scipy-0.5.2-14.1.i386

And as I pointed out, I have numpy installed already...

Also, when running the setup script absolutely nothing gets installed. This
might be related to the warning printed out:

/usr/lib/python2.4/distutils/dist.py:236: UserWarning: Unknown distribution
option: 'configuration'
  warnings.warn(msg)

Do I need a new version of 'distutils'? Any other idea in order to install
scipy?

Thanks
  Alex
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070731/0bcb684d/attachment.html>

From v-nijs at kellogg.northwestern.edu  Tue Jul 31 15:02:40 2007
From: v-nijs at kellogg.northwestern.edu (Vincent Nijs)
Date: Tue, 31 Jul 2007 14:02:40 -0500
Subject: [SciPy-user] google group
In-Reply-To: <46AF5D20.9030809@gmail.com>
Message-ID: <C2D4F180.A083%v-nijs@kellogg.northwestern.edu>

New messages to the scipy-user mailing list will now also be archived in a
google group at:

http://groups.google.com/group/scipy-user

Some relevant information from Google:

1. "A remote archive works a bit differently than a regular Google Group;
you'll still be able to search the archive and post messages to your mailing
list via the online form in the group, but you won't be able to post to the
Google Group via email because all of the messages will be coming from your
mailing list. Also, your subscription choices for the group will be limited
to No Email, Digest, and Abridged."

2. "Only new messages to the list will be stored in the Google Group"


Best,

Vincent




From Alexander.Dietz at astro.cf.ac.uk  Tue Jul 31 16:35:11 2007
From: Alexander.Dietz at astro.cf.ac.uk (Alexander Dietz)
Date: Tue, 31 Jul 2007 21:35:11 +0100
Subject: [SciPy-user] General question on scipy
In-Reply-To: <9cf809a00706290548u42ce009w32e6b279085553c2@mail.gmail.com>
References: <9cf809a00706290354l7ff898bexcb4f8adab3041a5d@mail.gmail.com>
	<f62qnu$ag7$1@sea.gmane.org> <4684EEC3.4040000@ar.media.kyoto-u.ac.jp>
	<9cf809a00706290502m4207e366xb5d7401f6568bbdb@mail.gmail.com>
	<e76aa17f0706290512i6ed900f7m937008476bf023aa@mail.gmail.com>
	<9cf809a00706290524p5fbc2e40m754ef6dc2c1994cf@mail.gmail.com>
	<e76aa17f0706290527p1d175ed0w6cc9039e5b16bde7@mail.gmail.com>
	<9cf809a00706290548u42ce009w32e6b279085553c2@mail.gmail.com>
Message-ID: <9cf809a00707311335w726da9ebh836e7a0f3905a67d@mail.gmail.com>

Hi,

I tried to install scipy the same way on my laptop, but this time it did not
work. I got an error this time:

error: Command "/usr/bin/g77 -g -Wall -fno-second-underscore -fPIC -O2
-funroll-loops -march=i686 -mmmx -msse2 -msse -fomit-frame-pointer
-malign-double -c -c Lib/fftpack/dfftpack/zfftb1.f -o build/temp.linux-
i686-2.4/Lib/fftpack/dfftpack/zfftb1.o" failed with exit status 1

Any ideas?

Cheers
  Alex



On 6/29/07, Alexander Dietz <Alexander.Dietz at astro.cf.ac.uk> wrote:
>
> Hi,
>
> I got it to work, finally, thanks to the back-online installing page. What
> I did was:
>
> yum install lapack-devel blas-devel
> cd scipy-0.5.2
> python setup.py install >& install.log
>
> This finally worked. Thanks a lot for helping me installing scipy!!!
>
> Alex
>
>
> On 6/29/07, Matthieu Brucher <matthieu.brucher at gmail.com > wrote:
> >
> > try python-scipy (or even more better, yum search scipy to find the real
> > name of the package)
> >
> > Matthieu
> >
> > 2007/6/29, Alexander Dietz < Alexander.Dietz at astro.cf.ac.uk>:
> > >
> > > Hi,
> > >
> > > here is my try:
> > >
> > > # yum install scipy
> > > Loading "installonlyn" plugin
> > > Setting up Install Process
> > > ....
> > > .....
> > > No Match for argument: scipy
> > > Nothing to do
> > >
> > >
> > > So there is no scipy found or so.
> > >
> > >
> > > Cheers
> > >   Alex
> > >
> > >
> > >
> > > On 6/29/07, Matthieu Brucher < matthieu.brucher at gmail.com> wrote:
> > >
> > > >
> > > >
> > > > 2007/6/29, Alexander Dietz < Alexander.Dietz at astro.cf.ac.uk>:
> > > > >
> > > > > Hi,
> > > > >
> > > > > here is what I am using exactly: FC5 and Linux
> > > > > 2.6.20-1.2316.fc5smp. Hope this helps.
> > > > > Can I install scipy (and all the dependend libraries) using 'yum'
> > > > > or something similar? 'yum' itself does not work...
> > > > >
> > > > > Cheers
> > > > >   Alex
> > > >
> > > >
> > > > yum install numpy
> > > > yum install scipy
> > > >
> > > > But FC5's scipy is old IIRC, can you update to FC 6 at least ?
> > > >
> > > > Matthieu
> > > >
> > > > _______________________________________________
> > > > 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
> >
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070731/4328ce37/attachment.html>

From william.ratcliff at gmail.com  Tue Jul 31 17:25:46 2007
From: william.ratcliff at gmail.com (william ratcliff)
Date: Tue, 31 Jul 2007 17:25:46 -0400
Subject: [SciPy-user] Question about scipy.test()
Message-ID: <827183970707311425x15810362jacd3907a3a85bca4@mail.gmail.com>

Hi!  I'm trying to compile scipy on a windows box running windows xp
professional.  The processor is an Intel Core 2 Extreme QX67--(it's a
quad).  I compiled atlas3.7.11 from source and it was fine.  Next, I built
numpy from source and it passed all of its tests.   Next, I built scipy from
source.  However, when I ran scipy.tests(level=1, verbosity=2),
it crashed python (2.5.1--latest release from python.org).   I checked and
both are using the libraries I generated from atlas and lapack.

The tests reached:

...
generation of a binary structure 4 ... ok
generic filter 1 ... ERROR
generic 1d filter 1

Then, it crashed python.

Any ideas?  I am using mingw to compile from the command window.

Thanks,
William Ratcliff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070731/bf08da0b/attachment.html>

From haase at msg.ucsf.edu  Tue Jul 31 17:58:19 2007
From: haase at msg.ucsf.edu (Sebastian Haase)
Date: Tue, 31 Jul 2007 23:58:19 +0200
Subject: [SciPy-user] histogram equalization using scipy.interpolate.splrep
Message-ID: <bc657ead0707311458m232158c5se7ccb9b6092ad207@mail.gmail.com>

Hi,
I'm trying to implement an efficiant numpy / scipy based
implementation of an image analysis procedure called "histogram
equalization".
http://en.wikipedia.org/wiki/Histogram_equalization

(I don't want to use the Python Imaging Library !)

For that I calculate the intesity histogram of my 2D or 3D images
(Let's call this 1D array `h`). Then I calculate `hh=
h.astype(N.float).cumsum()`

Now I need to create a new 2D / 3D image replacing each pixel value
with the corresponding "look-up value" in `hh`.
The means essentially doing something like

`a2 = hh[a]`  -- if `a` was the original image.
Of course this syntax does not work, because
a) the array index lookup probably is not possible using 2D/3D indices
like this - right !?
and
b) my histogram is just sampling / approximating all the actually
occuring values in `a`. In other words:  the histogram is based on
creating bins, and nearby pixel values in `a` are then counted into
same bin.  Consequently there is no simple  "array index lookup" to
get the needed value out of the `h` array. Instead one needs to
interpolate.

This is were I thought the scipy.interpolate package might come in handy ;-)
In fact, if `a` was a 1D "image" I think this would work:
 >>> rep = scipy.interpolate.splrep(x,y)   # x,y being the
horizontal,count axis in my histogram
 >>> aa = scipy.interpolate.splev(a,rep)

But for a.ndim being 2 I get:
Traceback (most recent call last):
  File "<input>", line 1, in ?
  File "C:\PrWinN\scipy\interpolate\fitpack.py", line 443, in splev
    y,ier=_fitpack._spl_(x,der,t,c,k)
ValueError: object too deep for desired array

Using N.vectorize(lambda x: scipy.interpolate.splev(x,rep))
works but is slow.

I there a fast vectorized version of some interpolation already in
scipy.interpolate ?
Am I missing something ?

Thanks,
Sebastian Haase


From Alexander.Dietz at astro.cf.ac.uk  Tue Jul 31 17:58:51 2007
From: Alexander.Dietz at astro.cf.ac.uk (Alexander Dietz)
Date: Tue, 31 Jul 2007 22:58:51 +0100
Subject: [SciPy-user] Another error while installing scipy
Message-ID: <9cf809a00707311458u26b2e5cehd3fd0a22e4a74010@mail.gmail.com>

Hi,

while trying to install scipy on my laptop (FC5, numpy 1.0.3) I encountered
finally an error never occured before. I would appreciate to resolve this
problem in order to get scipy installed.

Cheers
  Alex

Traceback (most recent call last):
  File "setup.py", line 55, in ?
    setup_package()
  File "setup.py", line 47, in setup_package
    configuration=configuration )
  File "/usr/lib/python2.4/site-packages/numpy/distutils/core.py", line 142,
in setup
    config = configuration()
  File "setup.py", line 19, in configuration
    config.add_subpackage('Lib')
  File "/usr/lib/python2.4/site-packages/numpy/distutils/misc_util.py", line
772, in add_subpackage
    caller_level = 2)
  File "/usr/lib/python2.4/site-packages/numpy/distutils/misc_util.py", line
755, in get_subpackage
    caller_level = caller_level + 1)
  File "/usr/lib/python2.4/site-packages/numpy/distutils/misc_util.py", line
702, in _get_configuration_from_setup_py
    config = setup_module.configuration(*args)
  File "/opt/users/src/scipy-0.5.2/Lib/setup.py", line 15, in configuration
    config.add_subpackage('odr')
  File "/usr/lib/python2.4/site-packages/numpy/distutils/misc_util.py", line
772, in add_subpackage
    caller_level = 2)
  File "/usr/lib/python2.4/site-packages/numpy/distutils/misc_util.py", line
755, in get_subpackage
    caller_level = caller_level + 1)
  File "/usr/lib/python2.4/site-packages/numpy/distutils/misc_util.py", line
687, in _get_configuration_from_setup_py
    ('.py', 'U', 1))
  File "Lib/odr/setup.py", line 9, in ?
    from numpy.distutils.misc_util import get_path, Configuration, dot_join
ImportError: cannot import name get_path


As far as I know I have all these lapack, blas and atlas packages installed.
Any ideas?

Cheers
 Alex
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20070731/fb396400/attachment.html>

From amcmorl at gmail.com  Tue Jul 31 18:00:50 2007
From: amcmorl at gmail.com (Angus McMorland)
Date: Wed, 1 Aug 2007 10:00:50 +1200
Subject: [SciPy-user] optimize.leastsq ier flag
In-Reply-To: <46AF557F.7000707@nist.gov>
References: <46AF557F.7000707@nist.gov>
Message-ID: <e85f13c40707311500m9b79731i400d2c25f2bb74d@mail.gmail.com>

Hi Stephen,

On 01/08/07, Stephen Yang <scyang at nist.gov> wrote:
> Hello everyone,
>
> I'm getting an ier flag of '4' when I run this code to fit some data to
> a gaussian:
>
> p0 = [1, 1, 1]
>
> gauss = lambda p, x: p[0]*numpy.exp(-((x-p[1])**2/(2*p[2]**2)))
> res = lambda p, x, y: gauss(p, x) - y
> optimized, success = optimize.minpack.leastsq(res, p0, args = (omega,
> counts))
> print 'success flag:', success
>
> In the minpack documentation, there is supposedly a 'mesg' return value,
> which states:
>
>     mesg -- a string message giving information about the cause of failure.
>
> So.. first, does anyone know what a flag of '4' means?

>From the MINPACK ldif documentation (http://www.netlib.org/minpack/lmdif.f):

c info = 4  the cosine of the angle between fvec and any
c                   column of the jacobian is at most gtol in
c                   absolute value.

I'm afraid I can't offer any advice as to what that means exactly...
will leave that to an expert.

> Secondly, how would you go about getting this flag?

>From the docstring:

Inputs:
    <snip>
    full_output -- non-zero to return all optional outputs.
    <snip>
Outputs: (x, {cov_x, infodict, mesg}, ier)

So, you need to call as:

optimized, blah,moreblah, mesg, success =
optimize.minpack.leastsq(res, p0, args = (omega, counts), full_output
= 1)

>I tried
>
> optimized, mesg, success = optimize.minpack.leastsq(res, p0, args =
> (omega, counts))
>
> with no luck..
> ValueError: need more than 2 values to unpack
>
> Any help would be greatly appreciated.
>

I hope that helps,

Angus.
-- 
AJC McMorland, PhD Student
Physiology, University of Auckland


From robert.kern at gmail.com  Tue Jul 31 18:04:17 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 31 Jul 2007 17:04:17 -0500
Subject: [SciPy-user] Another error while installing scipy
In-Reply-To: <9cf809a00707311458u26b2e5cehd3fd0a22e4a74010@mail.gmail.com>
References: <9cf809a00707311458u26b2e5cehd3fd0a22e4a74010@mail.gmail.com>
Message-ID: <46AFB1E1.5080402@gmail.com>

Alexander Dietz wrote:
> Hi,
> 
> while trying to install scipy on my laptop (FC5, numpy 1.0.3) I
> encountered finally an error never occured before. I would appreciate to
> resolve this problem in order to get scipy installed.

>   File "Lib/odr/setup.py", line 9, in ?
>     from numpy.distutils.misc_util import get_path, Configuration, dot_join
> ImportError: cannot import name get_path
> 
> 
> As far as I know I have all these lapack, blas and atlas packages
> installed. Any ideas?

I goofed before the 1.0.3 release. I thought that get_path() was only used
internally; it *is* currently deprecated in favor of using the Configuration
object. Lib/odr/setup.py is fixed in scipy SVN, though.

-- 
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 stefan at sun.ac.za  Tue Jul 31 18:15:35 2007
From: stefan at sun.ac.za (Stefan van der Walt)
Date: Wed, 1 Aug 2007 00:15:35 +0200
Subject: [SciPy-user] Question about scipy.test()
In-Reply-To: <827183970707311425x15810362jacd3907a3a85bca4@mail.gmail.com>
References: <827183970707311425x15810362jacd3907a3a85bca4@mail.gmail.com>
Message-ID: <20070731221535.GY7447@mentat.za.net>

Hi William

Are you compiling the latest version of the scipy source from SVN?  If
so, which version of the mingw gcc compiler are you using?

You'll find a bug-report related to this problem on the SciPy trac;
any additional information you can provide will be valuable.

Regards
St?fan

On Tue, Jul 31, 2007 at 05:25:46PM -0400, william ratcliff wrote:
> Hi!  I'm trying to compile scipy on a windows box running windows xp
> professional.  The processor is an Intel Core 2 Extreme QX67--(it's a quad).  I
> compiled atlas3.7.11 from source and it was fine.  Next, I built numpy from
> source and it passed all of its tests.   Next, I built scipy from source. 
> However, when I ran scipy.tests(level=1, verbosity=2),
> it crashed python (2.5.1--latest release from python.org).   I checked and both
> are using the libraries I generated from atlas and lapack.
> 
> The tests reached:
> 
> ...
> generation of a binary structure 4 ... ok
> generic filter 1 ... ERROR
> generic 1d filter 1
> 
> Then, it crashed python.
> 
> Any ideas?  I am using mingw to compile from the command window. 
> 
> Thanks,
> William Ratcliff


From zunzun at zunzun.com  Tue Jul 31 18:23:01 2007
From: zunzun at zunzun.com (zunzun at zunzun.com)
Date: Tue, 31 Jul 2007 18:23:01 -0400
Subject: [SciPy-user] General question on scipy
In-Reply-To: <9cf809a00707311335w726da9ebh836e7a0f3905a67d@mail.gmail.com>
References: <9cf809a00706290354l7ff898bexcb4f8adab3041a5d@mail.gmail.com>
	<f62qnu$ag7$1@sea.gmane.org>
	<4684EEC3.4040000@ar.media.kyoto-u.ac.jp>
	<9cf809a00706290502m4207e366xb5d7401f6568bbdb@mail.gmail.com>
	<e76aa17f0706290512i6ed900f7m937008476bf023aa@mail.gmail.com>
	<9cf809a00706290524p5fbc2e40m754ef6dc2c1994cf@mail.gmail.com>
	<e76aa17f0706290527p1d175ed0w6cc9039e5b16bde7@mail.gmail.com>
	<9cf809a00706290548u42ce009w32e6b279085553c2@mail.gmail.com>
	<9cf809a00707311335w726da9ebh836e7a0f3905a67d@mail.gmail.com>
Message-ID: <20070731222301.GA23052@zunzun.com>

On Tue, Jul 31, 2007 at 09:35:11PM +0100, Alexander Dietz wrote:
> I got an error this time:
> 
> error: Command "/usr/bin/g77 -g -Wall -fno-second-underscore -fPIC -O2
> -funroll-loops -march=i686 -mmmx -msse2 -msse -fomit-frame-pointer
> -malign-double -c -c Lib/fftpack/dfftpack/zfftb1.f -o build/temp.linux-
> i686-2.4/Lib/fftpack/dfftpack/zfftb1.o" failed with exit status 1

What happens if you type '/usr/bin/g77' on a command line?

     James



From william.ratcliff at gmail.com  Tue Jul 31 18:47:18 2007
From: william.ratcliff at gmail.com (william ratcliff)
Date: Tue, 31 Jul 2007 18:47:18 -0400
Subject: [SciPy-user] Question about scipy.test()
In-Reply-To: <20070731221535.GY7447@mentat.za.net>
References: <827183970707311425x15810362jacd3907a3a85bca4@mail.gmail.com>
	<20070731221535.GY7447@mentat.za.net>
Message-ID: <827183970707311547w3a57c809i81a2fac4768cea92@mail.gmail.com>

I am using the latest version of the scipy source from SVN.  I am using the
mingw from the enthought sumo distribution of python (2.4.3), just copied
over to the python25 directory.  I'm not sure of the version--

It is version 4.0.3 of gcc according to the folder in
C:\Python25\MingW\bin\lib\gcc-lib\i686-pc-mingw32

The only addition I made to the library was a recent download of g95.  Has
this bug come up before?

Thanks,
William

On 7/31/07, Stefan van der Walt <stefan at sun.ac.za> wrote:
>
> Hi William
>
> Are you compiling the latest version of the scipy source from SVN?  If
> so, which version of the mingw gcc compiler are you using?
>
> You'll find a bug-report related to this problem on the SciPy trac;
> any additional information you can provide will be valuable.
>
> Regards
> St?fan
>
> On Tue, Jul 31, 2007 at 05:25:46PM -0400, william ratcliff wrote:
> > Hi!  I'm trying to compile scipy on a windows box running windows xp
> > professional.  The processor is an Intel Core 2 Extreme QX67--(it's a
> quad).  I
> > compiled atlas3.7.11 from source and it was fine.  Next, I built numpy
> from
> > source and it passed all of its tests.   Next, I built scipy from
> source.
> > However, when I ran scipy.tests(level=1, verbosity=2),
> > it crashed python (2.5.1--latest release from python.org).   I checked
> and both
> > are using the libraries I generated from atlas and lapack.
> >
> > The tests reached:
> >
> > ...
> > generation of a binary structure 4 ... ok
> > generic filter 1 ... ERROR
> > generic 1d filter 1
> >
> > Then, it crashed python.
> >
> > Any ideas?  I am using mingw to compile from the command window.
> >
> > Thanks,
> > William Ratcliff
> _______________________________________________
> 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/20070731/85418335/attachment.html>

From aisaac at american.edu  Tue Jul 31 22:01:12 2007
From: aisaac at american.edu (Alan G Isaac)
Date: Tue, 31 Jul 2007 22:01:12 -0400
Subject: [SciPy-user] Finding Neighboors
In-Reply-To: <e76aa17f0704170541o5e18f9a8i9269fe48184ce287@mail.gmail.com>
References: <e76aa17f0704170541o5e18f9a8i9269fe48184ce287@mail.gmail.com>
Message-ID: <Mahogany-0.67.0-1024-20070731-220112.00@american.edu>

On Tue, 17 Apr 2007, Matthieu Brucher apparently wrote:
> I wanted to know if there was a module in scipy that is able to find the 
> k-neighboors of a point ? 
> If so, is there an optimized one - tree-based search - ? 
> If not, I'm doing the optimized version. 

Hi Matthieu,

Where did you go with this?

Thanks!
Alan






From david at ar.media.kyoto-u.ac.jp  Tue Jul 31 21:57:01 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Wed, 01 Aug 2007 10:57:01 +0900
Subject: [SciPy-user] Problems installing scipy
In-Reply-To: <9cf809a00707311129n64475320w1380d58832c80cc4@mail.gmail.com>
References: <9cf809a00707301301x53086d46q8f4242d3876545bb@mail.gmail.com>	<46AEA53D.5090306@ar.media.kyoto-u.ac.jp>
	<9cf809a00707311129n64475320w1380d58832c80cc4@mail.gmail.com>
Message-ID: <46AFE86D.6090306@ar.media.kyoto-u.ac.jp>

Alexander Dietz wrote:
> Hi,
>
> thanks for the tipps bot none of them are working!
>
> On 7/31/07, *David Cournapeau* < david at ar.media.kyoto-u.ac.jp 
> <mailto:david at ar.media.kyoto-u.ac.jp>> wrote:
>
>     Alexander Dietz wrote:
>     > Hi,
>     >
>     > I am trying to install scipy on my laptop, running FC5. I have
>     > installed numpy, so I downloaded the scipy-package and run the setup
>     > script which creates a warning only:
>     >
>     Note that to avoid all this hassle, there are some rpm for FC5 for
>     numpy
>     and scipy:
>
>     http://download.opensuse.org/repositories/home:/ashigabou/Fedora_Extras_5/
>     <http://download.opensuse.org/repositories/home:/ashigabou/Fedora_Extras_5/>
>
>
> I tried to install the rpm package, but, as I expected before, several 
> other packages are required to install scipy:
>
> error: Failed dependencies:
>         libblas.so.3 is needed by python-scipy-0.5.2-14.1.i386
>         liblapack.so.3 is needed by python-scipy-0.5.2-14.1.i386
>         python-numpy is needed by python-scipy-0.5.2-14.1.i386
Those are provided by the repository (that's the whole point :) ). You 
should add the repository for the method you are using (yum, or whatever 
else you are using). See the following for more details

http://en.opensuse.org/Build_Service/User

You can do it manually too, by downloading all the rpms from here:

http://download.opensuse.org/repositories/home:/ashigabou/Fedora_Extras_5/repodata/

For scipy, you need at least lapack3-3.0-devel, lapack3-3.0, refblas3 
and refblas3-devel.
>
> And as I pointed out, I have numpy installed already...
>
> Also, when running the setup script absolutely nothing gets installed. 
> This might be related to the warning printed out:
>
> /usr/lib/python2.4/distutils/dist.py:236: UserWarning: Unknown 
> distribution option: 'configuration'
>   warnings.warn(msg)
>
> Do I need a new version of 'distutils'? Any other idea in order to 
> install scipy?
No. Which numpy are you using ? The last release (1.0.3) is 
unfortunately unusable to build last release 0.5.2 at the moment (but 
the rpms above backport some changes from SVN to make it work, one more 
reason to use them :)).

I think it would be much easier to help you getting the rpm installed, 
if you want to get released versions working.

David


From david at ar.media.kyoto-u.ac.jp  Tue Jul 31 22:00:08 2007
From: david at ar.media.kyoto-u.ac.jp (David Cournapeau)
Date: Wed, 01 Aug 2007 11:00:08 +0900
Subject: [SciPy-user] Another error while installing scipy
In-Reply-To: <46AFB1E1.5080402@gmail.com>
References: <9cf809a00707311458u26b2e5cehd3fd0a22e4a74010@mail.gmail.com>
	<46AFB1E1.5080402@gmail.com>
Message-ID: <46AFE928.7040207@ar.media.kyoto-u.ac.jp>

Robert Kern wrote:
> Alexander Dietz wrote:
>   
>> Hi,
>>
>> while trying to install scipy on my laptop (FC5, numpy 1.0.3) I
>> encountered finally an error never occured before. I would appreciate to
>> resolve this problem in order to get scipy installed.
>>     
>
>   
>>   File "Lib/odr/setup.py", line 9, in ?
>>     from numpy.distutils.misc_util import get_path, Configuration, dot_join
>> ImportError: cannot import name get_path
>>
>>
>> As far as I know I have all these lapack, blas and atlas packages
>> installed. Any ideas?
>>     
>
> I goofed before the 1.0.3 release. I thought that get_path() was only used
> internally; it *is* currently deprecated in favor of using the Configuration
> object. Lib/odr/setup.py is fixed in scipy SVN, though.
>
>   
Alexander, to install numpy and scipy, you should really try the rpms of 
the ashigabou repository, they take care of all those problems. It is 
just a matter to add the repository to your package manager, and then 
everything is taken care for you. The above problem is taken care of, 
for example :) (I backported some changes from svn to make it work).

David


From rsamurti at airtelbroadband.in  Tue Jul 31 23:28:03 2007
From: rsamurti at airtelbroadband.in (R S Ananda Murthy)
Date: Wed, 01 Aug 2007 08:58:03 +0530
Subject: [SciPy-user] Error in SciPy install on Zenwalk-4.6.1 system.
Message-ID: <46AFFDC3.8060905@airtelbroadband.in>

Hello,

I am trying to make package of SciPy on Zenwalk-4.6.1. I have already 
installed numpy, lapack, fftw. When I do

python setup.py install --prefix=/usr --root=$dest

setup runs for some time, and then I get the following message:

  File "/usr/lib/python2.5/site-packages/numpy/distutils/misc_util.py",
line 687, in _get_configuration_from_setup_py
    ('.py', 'U', 1))
  File "Lib/odr/setup.py", line 9, in <module>
    from numpy.distutils.misc_util import get_path, Configuration, dot_join
ImportError: cannot import name get_path


Why am I getting this ImportError? How to correct this?

Thanks for your help,

Anand





From robert.kern at gmail.com  Tue Jul 31 23:33:12 2007
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 31 Jul 2007 22:33:12 -0500
Subject: [SciPy-user] Error in SciPy install on Zenwalk-4.6.1 system.
In-Reply-To: <46AFFDC3.8060905@airtelbroadband.in>
References: <46AFFDC3.8060905@airtelbroadband.in>
Message-ID: <46AFFEF8.9080108@gmail.com>

R S Ananda Murthy wrote:
> Hello,
> 
> I am trying to make package of SciPy on Zenwalk-4.6.1. I have already 
> installed numpy, lapack, fftw. When I do
> 
> python setup.py install --prefix=/usr --root=$dest
> 
> setup runs for some time, and then I get the following message:
> 
>   File "/usr/lib/python2.5/site-packages/numpy/distutils/misc_util.py",
> line 687, in _get_configuration_from_setup_py
>     ('.py', 'U', 1))
>   File "Lib/odr/setup.py", line 9, in <module>
>     from numpy.distutils.misc_util import get_path, Configuration, dot_join
> ImportError: cannot import name get_path
> 
> 
> Why am I getting this ImportError?

Because I screwed up just before the numpy 1.0.3 release and removed that
deprecated function without realizing that scipy 0.5.2 still used it.

> How to correct this?

This is fixed in the SVN scipy.

-- 
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