From rkern at ucsd.edu Thu Sep 1 01:02:23 2005 From: rkern at ucsd.edu (Robert Kern) Date: Wed, 31 Aug 2005 22:02:23 -0700 Subject: [SciPy-user] getting the "best" two eigenvectors for a PCA analysis with a power method In-Reply-To: <1125484655.4251.97.camel@sandwi.ch.cam.ac.uk> References: <20050829072328.GA5148@itb.biologie.hu-berlin.de> <1125484655.4251.97.camel@sandwi.ch.cam.ac.uk> Message-ID: <43168B5F.6070307@ucsd.edu> Noel O'Boyle wrote: > R calculates the PCs using singular value decomposition, instead of > using the eigenvalues of the covariance matrix. Another reason to use the SVD instead of an eigenvector decomposition of the covariance matrix: strictly speaking, there's no such thing as an eigenvector decomposition of a covariance matrix. The eigenvector decomposition is only defined for linear automorphisms, which map a space onto itself. Covariance matrices map a space onto its dual. -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From d.howey at imperial.ac.uk Thu Sep 1 10:43:01 2005 From: d.howey at imperial.ac.uk (Howey, David A) Date: Thu, 1 Sep 2005 15:43:01 +0100 Subject: [SciPy-user] Matplotlib fonts Message-ID: <056D32E9B2D93B49B01256A88B3EB218766EED@icex2.ic.ac.uk> Does anyone know any easy way to set the 'global' font sizes for eg. tick mark labels, titles, xlabels, ylabels for matplotlib? I would rather not use the matplotlibrc file since I want the code to be platform independent. Can I set something in my code that will then affect all subsequent figures? Dave From jdhunter at ace.bsd.uchicago.edu Thu Sep 1 11:05:50 2005 From: jdhunter at ace.bsd.uchicago.edu (John Hunter) Date: Thu, 01 Sep 2005 10:05:50 -0500 Subject: [SciPy-user] Matplotlib fonts In-Reply-To: <056D32E9B2D93B49B01256A88B3EB218766EED@icex2.ic.ac.uk> ("Howey, David A"'s message of "Thu, 1 Sep 2005 15:43:01 +0100") References: <056D32E9B2D93B49B01256A88B3EB218766EED@icex2.ic.ac.uk> Message-ID: <877je0j0a9.fsf@peds-pc311.bsd.uchicago.edu> >>>>> "Howey," == Howey, David A writes: David> Does anyone know any easy way to set the 'global' font David> sizes for eg. tick mark labels, titles, xlabels, ylabels David> for matplotlib? I would rather not use the matplotlibrc David> file since I want the code to be platform independent. Can David> I set something in my code that will then affect all David> subsequent figures? This question is more appropriate on matplotlib-users (attempting to set followups). You can subscribe at http://lists.sourceforge.net/mailman/listinfo/matplotlib-users . You can also set the rc params from within a script, and this will affect all subsequent figures. There are two ways to do it. * Use rcParams; this is a dictionary where the keys are the same as in the rc file from matplotlib import rcParams rcParams.update( { 'font.family' : 'serif', 'font.style' : 'normal', 'font.variant' : 'normal', 'font.weight' : 'bold', 'font.stretch' : 'normal', 'font.size' : 'large', }) from pylab import * plot([1,2,3]) show() or use the "rc" command, which may be syntactically nicer but is equivalent; see http://matplotlib.sourceforge.net/matplotlib.pylab.html#-rc from matplotlib import rc rc('font', family='serif', style='normal', variant='normal', weight='bold', stretch='normal', size='large') from pylab import * plot([1,2,3]) show() Note that changing these in the rc file is almost as portable. matplotlib respects per directory rc files, so you can make your customizations in the rc file and distribute it with your script/application and this will override the user's default configuration. JDH From d.howey at imperial.ac.uk Thu Sep 1 11:09:45 2005 From: d.howey at imperial.ac.uk (Howey, David A) Date: Thu, 1 Sep 2005 16:09:45 +0100 Subject: [SciPy-user] Matplotlib fonts Message-ID: <056D32E9B2D93B49B01256A88B3EB218766EEE@icex2.ic.ac.uk> Wicked. This is just what I needed. Thanks. I will try and join the matplotlib list now! Dave -----Original Message----- From: scipy-user-bounces at scipy.net [mailto:scipy-user-bounces at scipy.net] On Behalf Of John Hunter Sent: 01 September 2005 16:06 To: SciPy Users List Subject: Re: [SciPy-user] Matplotlib fonts >>>>> "Howey," == Howey, David A writes: David> Does anyone know any easy way to set the 'global' font David> sizes for eg. tick mark labels, titles, xlabels, ylabels David> for matplotlib? I would rather not use the matplotlibrc David> file since I want the code to be platform independent. Can David> I set something in my code that will then affect all David> subsequent figures? This question is more appropriate on matplotlib-users (attempting to set followups). You can subscribe at http://lists.sourceforge.net/mailman/listinfo/matplotlib-users . You can also set the rc params from within a script, and this will affect all subsequent figures. There are two ways to do it. * Use rcParams; this is a dictionary where the keys are the same as in the rc file from matplotlib import rcParams rcParams.update( { 'font.family' : 'serif', 'font.style' : 'normal', 'font.variant' : 'normal', 'font.weight' : 'bold', 'font.stretch' : 'normal', 'font.size' : 'large', }) from pylab import * plot([1,2,3]) show() or use the "rc" command, which may be syntactically nicer but is equivalent; see http://matplotlib.sourceforge.net/matplotlib.pylab.html#-rc from matplotlib import rc rc('font', family='serif', style='normal', variant='normal', weight='bold', stretch='normal', size='large') from pylab import * plot([1,2,3]) show() Note that changing these in the rc file is almost as portable. matplotlib respects per directory rc files, so you can make your customizations in the rc file and distribute it with your script/application and this will override the user's default configuration. JDH _______________________________________________ SciPy-user mailing list SciPy-user at scipy.net http://www.scipy.net/mailman/listinfo/scipy-user From giovanni.samaey at cs.kuleuven.ac.be Fri Sep 2 03:01:16 2005 From: giovanni.samaey at cs.kuleuven.ac.be (Giovanni Samaey) Date: Fri, 02 Sep 2005 09:01:16 +0200 Subject: [SciPy-user] subversion and installation instructions? Message-ID: <4317F8BC.8020708@cs.kuleuven.ac.be> Dear all, I have tried to install scipy on ournew university-wide cluster, but I am unsure as to whether the install was completely successful. I checked out scipy from svn, and because that alone gave me "scipy_base not found", I needed to look through the mail on this list to find out that there exists a separate "scipy_core" module, which also has to be checked out and installed separately. Now both of these are installed. However, I only get 246 tests on level=10. I remember that this was much more in the past. Did I forget to install something? I also get one failure, which is listed below. I must say that I am not exactly running on a standard system... We run on a cluster of Opterons, use pathscale as compiler, and the BLAS and LAPACK are from the AMD Core Math Library instead of ATLAS. Once this system is working and validated, I will post detailed installation instructions for this setup. Best, Giovanni ====================================================================== FAIL: check_dot (scipy.lib.blas.test_blas.test_fblas1_simple) ---------------------------------------------------------------------- Traceback (most recent call last): File "/data/home/u0038151/lib/python2.4/site-packages/scipy/lib/blas/tests/test_blas.py", line 75, in check_dot assert_almost_equal(f([3j,-4,3-4j],[2,3,1]),-9+2j) File "/data/home/u0038151/lib/python2.4/site-packages/scipy_test/testing.py", line 662, in assert_almost_equal assert round(abs(desired - actual),decimal) == 0, msg AssertionError: Items are not equal: DESIRED: (-9+2j) ACTUAL: (-9+2.1839129395899578e-317j) ---------------------------------------------------------------------- Ran 246 tests in 1.658s From oliphant at ee.byu.edu Fri Sep 2 17:32:15 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Fri, 02 Sep 2005 15:32:15 -0600 Subject: [SciPy-user] New Lyx Version eas(ier) to install on Windows. Message-ID: <4318C4DF.2090805@ee.byu.edu> There is a new Lyx Version that works on Windows quite well -- windows is now officially supported. So, those of you who have been hesitant to try LyX because it supposedly doesn't work on Windows, have no more excuse ;-) An executable installer is available at http://www.lyx.org The setup process guides you through the other programs that are needed (ImageMagick, Ghostscript, a minimal shell like minsys, Python, and optionally Perl) to run lyx well. -Travis From aisaac at american.edu Fri Sep 2 20:38:15 2005 From: aisaac at american.edu (Alan G Isaac) Date: Fri, 2 Sep 2005 20:38:15 -0400 Subject: [SciPy-user] Re: [SciPy-dev] scipy core (Numeric3) win32 binaries to play with In-Reply-To: <4318E60D.4040304@ee.byu.edu> References: <4318E60D.4040304@ee.byu.edu> Message-ID: On Fri, 02 Sep 2005, Travis Oliphant apparently wrote: > http://numeric.scipy.org/files/scipy_core-0.4.0.win32-py2.4.exe So far so good. Thanks! Alan Isaac From yaroslavvb at gmail.com Fri Sep 2 21:25:23 2005 From: yaroslavvb at gmail.com (Yaroslav Bulatov) Date: Fri, 2 Sep 2005 18:25:23 -0700 Subject: [SciPy-user] strange behaviour with arrays of PyObjects In-Reply-To: <056D32E9B2D93B49B01256A88B3EB218766EEE@icex2.ic.ac.uk> References: <056D32E9B2D93B49B01256A88B3EB218766EEE@icex2.ic.ac.uk> Message-ID: from scipy import * import scipy from scipy.special import poly1d If I do array([poly1d]*2,PyObject)[1] it works as expected, but if I do array([poly1d([])]*2,PyObject)[1] I get "index out of bounds"...why? From aisaac at american.edu Sat Sep 3 02:23:57 2005 From: aisaac at american.edu (Alan G Isaac) Date: Sat, 3 Sep 2005 02:23:57 -0400 Subject: [SciPy-user] Re: [SciPy-dev] New Lyx Version eas(ier) to install on Windows. In-Reply-To: <4318C4DF.2090805@ee.byu.edu> References: <4318C4DF.2090805@ee.byu.edu> Message-ID: On Fri, 02 Sep 2005, Travis Oliphant apparently wrote: > those of you who have been hesitant to try LyX because it supposedly > doesn't work on Windows, have no more excuse ;-) > An executable installer is available at http://www.lyx.org > The setup process guides you through the other programs that are needed > (ImageMagick, Ghostscript, a minimal shell like minsys, Python, and > optionally Perl) to run lyx well. 1. A question: what ImageMagick file is needed? http://sourceforge.net/project/showfiles.php?group_id=24099 (I.e., what do you get out of "dynamic"?) Comment: ImageMagick seems very RAM intensive. 2. I do not see the shell requirment. Is that just your recommendation? 3. Are you using it on Windows? If so, how's it going? Thanks, Alan Isaac From rkern at ucsd.edu Sat Sep 3 07:23:30 2005 From: rkern at ucsd.edu (Robert Kern) Date: Sat, 03 Sep 2005 04:23:30 -0700 Subject: [SciPy-user] Mersenne Twister PRNG Message-ID: <431987B2.4080902@ucsd.edu> I've made a stab at replacing the RANLIB-based PRNG that we currently use in scipy.stats with the superior-in-every-conceivable-way Mersenne Twister. It's currently a separate module, and it has only been tested on Numeric 24.0b2 not the new scipy.base, but it should be able to be almost a drop-in replacement for the module that's currently there. The only distribution I haven't covered is the multivariate normal because I didn't really feel like doing linear algebra in C. But that's easy enough to do at the Python level. The core Mersenne Twister routines come from Jean-Sebastien Roy's RandomKit. Thank you, JS! http://www.jeannot.org/~js/code/index.en.html I would appreciate people putting it through its paces. I've done visual checking of the distributions via Q-Q plots; the script is included as the tests/qq.py and requires matplotlib. I haven't implemented real unit tests via K-S significance tests or Kullback-Leibler distance, but I guess I ought to eventually. The sampling algorithms were implemented from scratch by me from the original literature, so if you see funky numbers, it's probably my fault. AFAICT right now, the numbers are no funkier than those produced by RANLIB. http://cheeseshop.python.org/pypi/mtrand/0.1 -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From skip at pobox.com Sat Sep 3 19:55:30 2005 From: skip at pobox.com (skip at pobox.com) Date: Sat, 3 Sep 2005 18:55:30 -0500 Subject: [SciPy-user] Mersenne Twister PRNG In-Reply-To: <431987B2.4080902@ucsd.edu> References: <431987B2.4080902@ucsd.edu> Message-ID: <17178.14322.594916.467829@montanaro.dyndns.org> Robert> I've made a stab at replacing the RANLIB-based PRNG that we Robert> currently use in scipy.stats with the superior-in-every- Robert> conceivable-way Mersenne Twister. Python has used the Mersenne Twister as the basis for its random number generator since at least 2.2. Is there a reason to incorporate it separately? -- Skip Montanaro skip at pobox.com From rkern at ucsd.edu Sat Sep 3 20:15:22 2005 From: rkern at ucsd.edu (Robert Kern) Date: Sat, 03 Sep 2005 17:15:22 -0700 Subject: [SciPy-user] Mersenne Twister PRNG In-Reply-To: <17178.14322.594916.467829@montanaro.dyndns.org> References: <431987B2.4080902@ucsd.edu> <17178.14322.594916.467829@montanaro.dyndns.org> Message-ID: <431A3C9A.7070009@ucsd.edu> skip at pobox.com wrote: > Robert> I've made a stab at replacing the RANLIB-based PRNG that we > Robert> currently use in scipy.stats with the superior-in-every- > Robert> conceivable-way Mersenne Twister. > > Python has used the Mersenne Twister as the basis for its random number > generator since at least 2.2. Is there a reason to incorporate it > separately? Yes! We need to get lots of random numbers stuffed into arrays on the C level. AFAICT, _random doesn't expose a C API. -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From gnata at obs.univ-lyon1.fr Mon Sep 5 05:03:42 2005 From: gnata at obs.univ-lyon1.fr (Xavier Gnata) Date: Mon, 05 Sep 2005 11:03:42 +0200 Subject: [SciPy-user] Re: [SciPy-dev] scipy core (Numeric3) win32 binaries to play with In-Reply-To: References: <4318E60D.4040304@ee.byu.edu> Message-ID: <431C09EE.7010105@obs.univ-lyon1.fr> Alan G Isaac wrote: >On Fri, 02 Sep 2005, Travis Oliphant apparently wrote: > > >>http://numeric.scipy.org/files/scipy_core-0.4.0.win32-py2.4.exe >> >> > >So far so good. > >Thanks! >Alan Isaac > >_______________________________________________ >SciPy-user mailing list >SciPy-user at scipy.net >http://www.scipy.net/mailman/listinfo/scipy-user > > > Hi, That's great news! :) Where are the sources corresponding with this windows release (I would like to test that under linux asap)? Is there any beta version documentation? Thanks. Xavier. From jdc at uwo.ca Mon Sep 5 12:42:42 2005 From: jdc at uwo.ca (Dan Christensen) Date: Mon, 05 Sep 2005 12:42:42 -0400 Subject: [SciPy-user] SciPy 2005 Conference: Preliminary Schedule of Presentations In-Reply-To: <42FD2CD6.9020904@enthought.com> References: <42FD2CD6.9020904@enthought.com> Message-ID: <87u0gzzcsd.fsf@uwo.ca> "Travis N. Vaught" writes: > I've posted a preliminary schedule of presentations at: > > http://www.scipy.org/wikis/scipy05/ConferenceSchedule > > Please feel free to provide comments or corrections--either to the list > or directly to me. The abstract links all point back to the schedule. If abstracts are available, I'd be curious to see them. Dan From yaroslavvb at gmail.com Mon Sep 5 18:11:03 2005 From: yaroslavvb at gmail.com (Yaroslav Bulatov) Date: Mon, 5 Sep 2005 15:11:03 -0700 Subject: [SciPy-user] scipy can't find Numeric header files during install In-Reply-To: <430BF53F.3080209@ee.byu.edu> References: <430BF53F.3080209@ee.byu.edu> Message-ID: > >I've installed Numeric and headers went to " ~/usr/include/python2.3/Numeric/" > >But now when I install SciPy by doing "python setup.py install > >--prefix=~/usr", it says > > > >Lib/cluster/src/vq_wrap.cpp:563:33: Numeric/arrayobject.h: No such > >file or directory > > > >How can I tell it to look in "~/usr/include/python2.3/Numeric/" for > >Numeric headers? > > > > > > > > Good question. I'm not sure exactly how the include directories are > found. You may need to install Numeric first with the same prefix (do > you have a python installation with that prefix?) > > But, you could try some things. > > Doing... > > python setup.py --help config > > I got the message > > Global options: > --verbose (-v) run verbosely (default) > --quiet (-q) run quietly (turns verbosity off) > --dry-run (-n) don't actually do anything > --help (-h) show detailed help message > > Options for 'config' command: > --compiler specify the compiler type > --cc specify the compiler executable > --include-dirs (-I) list of directories to search for header files > --define (-D) C preprocessor macros to define > --undef (-U) C preprocessor macros to undefine > --libraries (-l) external C libraries to link with > --library-dirs (-L) directories to search for external C libraries > --noisy show every action (compile, link, run, ...) taken > --dump-source dump generated source files before attempting to > compile them > --fcompiler specify the Fortran compiler type > > usage: setup.py [global_opts] cmd1 [cmd1_opts] [cmd2 [cmd2_opts] ...] > or: setup.py --help [cmd1 cmd2 ...] > or: setup.py --help-commands > or: setup.py cmd --help > > > So, I suspect that > > python setup.py --prefix=~/usr config --include-dirs > /usr/include/python2.3 install > > might work. > > You might have to run the build command first and then the install > command. OK that worked thanks From yaroslavvb at gmail.com Mon Sep 5 18:28:32 2005 From: yaroslavvb at gmail.com (Yaroslav Bulatov) Date: Mon, 5 Sep 2005 15:28:32 -0700 Subject: [SciPy-user] No module named __cvs_version__ Message-ID: I get that warning when I execute scipy.special.chebyt(1) The error comes from somewhere inside of gen_roots_and_weights What might be causing this? Also, how can I go about tracking down the source? I tried by pasting various bits of code from various scipy files and seeing if error is reproduced, but is there a more efficient method? -- Yaroslav Bulatov bulatov at cs.oregonstate.edu Dearborn 102 Oregon State University Corvallis, OR From nvf at MIT.EDU Tue Sep 6 12:03:53 2005 From: nvf at MIT.EDU (Nick Fotopoulos) Date: Tue, 06 Sep 2005 12:03:53 -0400 Subject: [SciPy-user] Upgrading scipy.io.loadmat Message-ID: <431DBDE9.6030901@mit.edu> Greetings, A couple months ago, I posted to scipy-user asking if anyone knew how to fix scipy.io.loadmat to read Matlab 7 (R14) mat-files. Receiving no response, I hacked the mio.py myself and can now read in the new version of mat-files. While I was tinkering, I fixed the in-place-but-broken support for complex arrays; it returned only the real parts before. scipy-users, this is just letting you know that if you need these features, there is a working solution. scipy-devs, what is the proper channel for submitting patches? Take care, Nick Fotopoulos From oliphant at ee.byu.edu Tue Sep 6 13:43:21 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Tue, 06 Sep 2005 11:43:21 -0600 Subject: [SciPy-user] Upgrading scipy.io.loadmat In-Reply-To: <431DBDE9.6030901@mit.edu> References: <431DBDE9.6030901@mit.edu> Message-ID: <431DD539.1090406@ee.byu.edu> Nick Fotopoulos wrote: > Greetings, > > A couple months ago, I posted to scipy-user asking if anyone knew how to > fix scipy.io.loadmat to read Matlab 7 (R14) mat-files. Receiving no > response, I hacked the mio.py myself and can now read in the new version > of mat-files. While I was tinkering, I fixed the in-place-but-broken > support for complex arrays; it returned only the real parts before. Excellent, your R14 contribution is greatly appreciated. By the way, I believe the complex array problem was fixed (which version of SciPy did you start from?). At any rate, I would be glad to apply your patches. There is a bug-tracker at sourceforge where you can store a patch to make sure it gets recorded. But, make sure you let scipy-dev know about the patch so we can apply it. Finally, scipy is undergoing some changes and so don't be discouraged if your patch doesn't get applied immediately. It's on the queue.... -Travis O. From stephen.walton at csun.edu Tue Sep 6 18:29:31 2005 From: stephen.walton at csun.edu (Stephen Walton) Date: Tue, 06 Sep 2005 15:29:31 -0700 Subject: [SciPy-user] subversion and installation instructions? In-Reply-To: <4317F8BC.8020708@cs.kuleuven.ac.be> References: <4317F8BC.8020708@cs.kuleuven.ac.be> Message-ID: <431E184B.5060407@csun.edu> Giovanni Samaey wrote: > I have tried to install scipy on ournew university-wide cluster, but I > am unsure as to whether the install was completely successful. Coming back to this after an absence to do some other things: I believe that, at present, it is still not possible to build scipy from the SVN archive on a machine on which scipy was never before installed. If I do 'rm -rf scipy* weave gui_thread' in /usr/lib/python2.4, and then try: svn co http://svn.scipy.org/svn/scipy/trunk scipy cd scipy setup.py build it immediately complains that there is no scipy_distutils module. The best thing to do, still, seems to be to download Fernando Perez's snapshot of the last CVS archive at http://ipython.scipy.org/tmp/scipy_cvs_2005-07-29.tgz and build that. Even using that archive, I'm finding that "python setup.py bdist_rpm" in scipy/scipy_core doesn't work any more; I think it used to. -------------- next part -------------- A non-text attachment was scrubbed... Name: stephen.walton.vcf Type: text/x-vcard Size: 323 bytes Desc: not available URL: From stephen.walton at csun.edu Tue Sep 6 18:40:00 2005 From: stephen.walton at csun.edu (Stephen Walton) Date: Tue, 06 Sep 2005 15:40:00 -0700 Subject: [SciPy-user] subversion and installation instructions? In-Reply-To: <431E184B.5060407@csun.edu> References: <4317F8BC.8020708@cs.kuleuven.ac.be> <431E184B.5060407@csun.edu> Message-ID: <431E1AC0.2070407@csun.edu> Stephen Walton wrote: > Giovanni Samaey wrote: > >> I have tried to install scipy on ournew university-wide cluster, but >> I am unsure as to whether the install was completely successful. > > > Coming back to this after an absence to do some other things: and now I see Nils Wagner's previous post on this issue, which is to check out scipy_core and scipy both from SVN. My apologies. From skip at pobox.com Tue Sep 6 22:40:07 2005 From: skip at pobox.com (skip at pobox.com) Date: Tue, 6 Sep 2005 21:40:07 -0500 Subject: [SciPy-user] Mersenne Twister PRNG In-Reply-To: <431A3C9A.7070009@ucsd.edu> References: <431987B2.4080902@ucsd.edu> <17178.14322.594916.467829@montanaro.dyndns.org> <431A3C9A.7070009@ucsd.edu> Message-ID: <17182.21255.710639.417352@montanaro.dyndns.org> Robert> AFAICT, _random doesn't expose a C API. It shouldn't be hard to add. Take a look at how it's done for the datetime module (PyDateTime_CAPI struct, PyDateTime_IMPORT macro, etc). Skip From rkern at ucsd.edu Wed Sep 7 00:20:14 2005 From: rkern at ucsd.edu (Robert Kern) Date: Tue, 06 Sep 2005 21:20:14 -0700 Subject: [SciPy-user] Mersenne Twister PRNG In-Reply-To: <17182.21255.710639.417352@montanaro.dyndns.org> References: <431987B2.4080902@ucsd.edu> <17178.14322.594916.467829@montanaro.dyndns.org> <431A3C9A.7070009@ucsd.edu> <17182.21255.710639.417352@montanaro.dyndns.org> Message-ID: <431E6A7E.4070706@ucsd.edu> skip at pobox.com wrote: > Robert> AFAICT, _random doesn't expose a C API. > > It shouldn't be hard to add. Take a look at how it's done for the datetime > module (PyDateTime_CAPI struct, PyDateTime_IMPORT macro, etc). Sure, and I might do just that later. However, the PRNG in scipy really needs to be updated before Python 2.5 will be released or widely adopted. In any case, the amount of code duplication is small, just the MT19937 algorithm and initialization code which is tiny and from the same reference implementation. I don't think anyone would derive much benefit from our using the exported version. -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From giovanni.samaey at cs.kuleuven.ac.be Wed Sep 7 02:49:24 2005 From: giovanni.samaey at cs.kuleuven.ac.be (Giovanni Samaey) Date: Wed, 07 Sep 2005 08:49:24 +0200 Subject: [SciPy-user] subversion and installation instructions? In-Reply-To: <431E1AC0.2070407@csun.edu> References: <4317F8BC.8020708@cs.kuleuven.ac.be> <431E184B.5060407@csun.edu> <431E1AC0.2070407@csun.edu> Message-ID: <431E8D74.1050905@cs.kuleuven.ac.be> > and now I see Nils Wagner's previous post on this issue, which is to > check out scipy_core and scipy both from SVN. My apologies. > Yes this works -- my question was if it is normal to have so many tests missing, and this one failure that I mentioned. > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user From alexander.borghgraef.rma at gmail.com Wed Sep 7 09:00:54 2005 From: alexander.borghgraef.rma at gmail.com (Alexander Borghgraef) Date: Wed, 7 Sep 2005 15:00:54 +0200 Subject: [SciPy-user] Setting include dirs Message-ID: <9e8c52a205090706002f380088@mail.gmail.com> Hi all, I'm trying to install SciPy 0.3.2, but it gets stuck at the compilation of build/src/fortranobject.c. The problem seems to be that it doesn't find Numeric/arrayobject.h which is quite understandable since on the line before the error message I see the following: compile options: '-Ibuild/src -I/usr/include/python2.3 -c' Since I don't have root access to this machine, I installed Numeric locally, so the compile options should contain -I$HOME/local/inlcude/python2.3 . I did add the local directory to the PYTHON_INCLUDES environment variable, and the numpy info line at the beginning of the install run seems to find it too: numpy_info: FOUND: define_macros = [('NUMERIC_VERSION', '"\\"24.0b2\\""')] include_dirs = ['/homes/morlet/aborghgr/local/include/python2.3'] Strangely, it does not get included for compilation somehow. Any ideas on how to fix this? -- Alex Borghgraef -------------- next part -------------- An HTML attachment was scrubbed... URL: From yaroslavvb at gmail.com Wed Sep 7 13:28:05 2005 From: yaroslavvb at gmail.com (Yaroslav Bulatov) Date: Wed, 7 Sep 2005 10:28:05 -0700 Subject: [SciPy-user] Numerical stability and Legendre polynomials Message-ID: I'm trying to approximate a function using an expansion over few dozen of Legendre polynomials, but I'm running into a large numerical error. For instance the following special.legendre(30).integ()(1)-special.legendre(30).integ()(-1) evaluates to 2e-4, instead of 0 So my question is -- is such numerical error normal? Is it realistic to try to approximate a function using 30 Legendre polynomials? -- Yaroslav Bulatov bulatov at cs.oregonstate.edu Dearborn 102 Oregon State University Corvallis, OR From Fernando.Perez at colorado.edu Wed Sep 7 14:37:22 2005 From: Fernando.Perez at colorado.edu (Fernando Perez) Date: Wed, 07 Sep 2005 12:37:22 -0600 Subject: [SciPy-user] Numerical stability and Legendre polynomials In-Reply-To: References: Message-ID: <431F3362.9060200@colorado.edu> Yaroslav Bulatov wrote: > I'm trying to approximate a function using an expansion over few dozen > of Legendre polynomials, but I'm running into a large numerical error. > > For instance the following > special.legendre(30).integ()(1)-special.legendre(30).integ()(-1) > evaluates to 2e-4, instead of 0 > > So my question is -- is such numerical error normal? Is it realistic > to try to approximate a function using 30 Legendre polynomials? The problem, I think, is that special.legendre returns a poly1d object. The coefficients are reasonably well computed, as this comparison (for P_30) with Mathematica shows: Python: In [31]: p30coefs Out[31]: [-0.14446444809436668, 1.5679191456676283e-14, 67.175968363880514, -2.180105233807014e-12, -5172.5495640188101, 7.7713427093987573e-11, 156900.67010857066, 4.5831526974104543e-09, -2487996.3402931187, 3.887890884521149e-07, 23718898.444125757, 3.1658444632926091e-06, -147344672.15291381, -1.590865831210945e-05, 626619649.70533192, -0.00028178808880156549, -1879858949.11516, -0.00073104158643097776, 4042311073.5892482, 3.7885858888136625e-05, -6254944503.3432274, 0.0010140695072848572, 6904808867.3253412, 0.00041825387498683678, -5303693767.6564827, 6.9383401445695963e-05, 2692644528.1947165, 8.5971433363986932e-06, -812067397.39206791, 2.6902196481091132e-07, 110142474.588809] Mathematica: N[LegendreP[30, x], 20] \!\(\(-0.14446444809436798095703125`20. \) + 67.17596836388111114501953125`20. \ x\^2 - 5172.54956401884555816650390625`20. \ x\^4 + 156900.67010857164859771728515625`20. \ x\^6 - 2.48799634029306471347808837890625`20.*^6\ x\^8 + 2.371889844412721693515777587890625`20.*^7\ x\^10 - 1.4734467215291149914264678955078125`20.*^8\ x\^12 + 6.2661964970523901283740997314453125`20.*^8\ x\^14 - 1.87985894911571703851222991943359375`20.*^9\ x\^16 + 4.04231107358869872987270355224609375`20.*^9\ x\^18 - 6.25494450334251277148723602294921875`20.*^9\ x\^20 + 6.90480886732615046203136444091796875`20.*^9\ x\^22 - 5.30369376765631847083568572998046875`20.*^9\ x\^24 + 2.69264452819474630057811737060546875`20.*^9\ x\^26 - 8.1206739739206634461879730224609375`20.*^8\ x\^28 + 1.1014247458880899846553802490234375`20.*^8\ x\^30\) But the problem is that this is an unstable polynomial to compute directly, because of the alternation in signs. Since poly1d doesn't know anything about this fact, it happily computes an unstable quantity. You'll also notice that the odd coefficients, which are analytically zero (for even order Legendre polys), are not quite zero in this construction, another source of error. If you need accurate results with Legendre polynomials as the order increases, your best bet is to use the standard 3-term recursion, which is stable. A braindead implmenentation of the algorithm, as taken from any textbook, gives this: In [47]: p30_end = utils.plegnv(array([-1.0,1.0]),30) In [48]: for n,ends in enumerate(p30_end): ....: print ends[0]-((-1)**n)*ends[-1] ....: 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 This is checking all 30 by subtracting the endpoints, accounting for even/oddness. Here's the same calculation using scipy: In [33]: for n in range(30): ....: pn = scipy.special.legendre(n) ....: print pn(-1.0)-((-1)**n)*pn(1.0) ....: 0.0 0.0 0.0 0.0 -1.11022302463e-15 -4.55191440096e-15 -1.40998324127e-14 -4.4408920985e-15 -4.41868763801e-14 -3.41948691585e-14 -2.05613304161e-13 -1.99495975295e-12 2.8537172625e-12 -3.97726296342e-12 -6.78088696304e-11 1.88647986121e-10 3.10480530175e-10 -9.75348690702e-11 -5.69281288776e-09 -6.65902566421e-09 -7.10117353808e-08 1.19151672973e-07 7.9948702647e-07 1.12660495866e-06 -3.17713138309e-06 1.19201396691e-05 8.83580242022e-05 -4.29196349421e-05 0.000923915025271 0.000167352152224 As you can see, the 3-term recursion fares much, much better. HTH, f From alexander.borghgraef.rma at gmail.com Thu Sep 8 02:49:49 2005 From: alexander.borghgraef.rma at gmail.com (Alexander Borghgraef) Date: Thu, 8 Sep 2005 08:49:49 +0200 Subject: [SciPy-user] Setting include dirs Message-ID: <9e8c52a20509072349490cb0d6@mail.gmail.com> Hi all, I'm trying to install SciPy 0.3.2, but it gets stuck at the compilation of build/src/fortranobject.c. The problem seems to be that it doesn't find Numeric/arrayobject.h which is quite understandable since on the line before the error message I see the following: compile options: '-Ibuild/src -I/usr/include/python2.3 -c' Since I don't have root access to this machine, I installed Numeric locally, so the compile options should contain -I$HOME/local/inlcude/python2.3 . I did add the local directory to the PYTHON_INCLUDES environment variable, and the numpy info line at the beginning of the install run seems to find it too: numpy_info: FOUND: define_macros = [('NUMERIC_VERSION', '"\\"24.0b2\\""')] include_dirs = ['/homes/morlet/aborghgr/local/include/python2.3'] Strangely, it does not get included for compilation somehow. Any ideas on how to fix this? -- Alex Borghgraef -------------- next part -------------- An HTML attachment was scrubbed... URL: From alexander.borghgraef.rma at gmail.com Thu Sep 8 04:00:45 2005 From: alexander.borghgraef.rma at gmail.com (Alexander Borghgraef) Date: Thu, 8 Sep 2005 10:00:45 +0200 Subject: [SciPy-user] More installation troubles Message-ID: <9e8c52a205090801005dbc0485@mail.gmail.com> Some more problems installing SciPy: firstly the installer doesn't seem to find the fftw library. I did set an environment variable FFTW pointing to the local lib and include directories where libfftw3.a, libfftw3.la , fftw3.f and fftw3.h are to be found. Still I get: fftw_info: NOT AVAILABLE dfftw_info: NOT AVAILABLE FFTW (http://www.fftw.org/) libraries not found. Directories to search for the libraries can be specified in the scipy_distutils/site.cfg file (section [fftw]) or by setting the FFTW environment variable. djbfft_info: NOT AVAILABLE DJBFFT (http://cr.yp.to/djbfft.html) libraries not found. Directories to search for the libraries can be specified in the scipy_distutils/site.cfg file (section [djbfft]) or by setting the DJBFFT environment variable. Any idea why this is? Then it goes on to find blas, lapack and atlas, gives a warning that lapack is probably incomplete (which is strange, since I compiled it from source following the instructions on the 'building scipy' page), finds x11, and crashes with the following error: ATLAS version 3.6.0 Traceback (most recent call last): File "setup.py", line 111, in ? setup_package(ignore_packages) File "setup.py", line 85, in setup_package ignore_packages = ignore_packages) File "scipy_core/scipy_distutils/misc_util.py", line 475, in get_subpackages config = setup_module.configuration(*args) File "/home/morlet/aborghgr/incoming/SciPy_complete-0.3.2/scipy_core/scipy_base/setup_scipy_base.py", line 18, in configuration numpy_info = get_info('numpy') File "scipy_core/scipy_distutils/system_info.py", line 192, in get_info return cl().get_info(notfound_action) File "scipy_core/scipy_distutils/system_info.py", line 1111, in __init__ module = __import__(self.modulename) File "/homes/morlet/aborghgr/local//lib/python2.3/site-packages/Numeric/Numeric.py", line 358, in ? from dotblas import dot, innerproduct, vdot File "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/dotblas.py", line 5, in ? import generic as _gen File "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/generic.py", line 13, in ? import numerictypes as _nt File "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/numerictypes.py", line 168, in ? Byte = _register("Byte", Int8) File "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/numerictypes.py", line 68, in _register raise ValueError("Type %s has already been registered" % name) ValueError: Type Byte has already been registered I don't have a clue as to what this means. Anyone care to help? P.S. Sorry about the double message, apparently I don't get my own posts, so I figured it lost. -- Alex Borghgraef -------------- next part -------------- An HTML attachment was scrubbed... URL: From pearu at scipy.org Thu Sep 8 04:44:59 2005 From: pearu at scipy.org (Pearu Peterson) Date: Thu, 8 Sep 2005 03:44:59 -0500 (CDT) Subject: [SciPy-user] More installation troubles In-Reply-To: <9e8c52a205090801005dbc0485@mail.gmail.com> References: <9e8c52a205090801005dbc0485@mail.gmail.com> Message-ID: On Thu, 8 Sep 2005, Alexander Borghgraef wrote: > Some more problems installing SciPy: firstly the installer doesn't seem to > find the fftw library. I did set an environment variable > FFTW pointing to the local lib and include directories where libfftw3.a, > libfftw3.la , fftw3.f and fftw3.h are to be found. Still > I get: > > fftw_info: > NOT AVAILABLE > > dfftw_info: > NOT AVAILABLE > > > FFTW (http://www.fftw.org/) libraries not found. > Directories to search for the libraries can be specified in the > scipy_distutils/site.cfg file (section [fftw]) or by setting > the FFTW environment variable. > djbfft_info: > NOT AVAILABLE > > > DJBFFT (http://cr.yp.to/djbfft.html) libraries not found. > Directories to search for the libraries can be specified in the > scipy_distutils/site.cfg file (section [djbfft]) or by setting > the DJBFFT environment variable. > > Any idea why this is? Then it goes on to find blas, lapack and atlas, gives > a warning that lapack is probably incomplete (which is strange, > since I compiled it from source following the instructions on the 'building > scipy' page), finds x11, and crashes with the following error: scipy 0.3.2 does not have fftw3 support, it's only available in CVS/SVN version of scipy. You can either ignore it (since scipy uses fortran fftpack instead fftw), or install fftw2, or get scipy/scipy_core from SVN repository. > ATLAS version 3.6.0 > Traceback (most recent call last): > File "setup.py", line 111, in ? > setup_package(ignore_packages) > File "setup.py", line 85, in setup_package > ignore_packages = ignore_packages) > File "scipy_core/scipy_distutils/misc_util.py", line 475, in get_subpackages > config = setup_module.configuration(*args) > File "/home/morlet/aborghgr/incoming/SciPy_complete-0.3.2/scipy_core/scipy_base/setup_scipy_base.py", > line 18, in configuration > numpy_info = get_info('numpy') > File "scipy_core/scipy_distutils/system_info.py", line 192, in get_info > return cl().get_info(notfound_action) > File "scipy_core/scipy_distutils/system_info.py", line 1111, in __init__ > module = __import__(self.modulename) > File > "/homes/morlet/aborghgr/local//lib/python2.3/site-packages/Numeric/Numeric.py", > line 358, in ? > from dotblas import dot, innerproduct, vdot > File > "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/dotblas.py", > line 5, in ? > import generic as _gen > File > "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/generic.py", > line 13, in ? > import numerictypes as _nt > File > "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/numerictypes.py", > line 168, in ? > Byte = _register("Byte", Int8) > File > "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/numerictypes.py", > line 68, in _register > raise ValueError("Type %s has already been registered" % name) > ValueError: Type Byte has already been registered > > I don't have a clue as to what this means. Anyone care to help? This looks like a numarray/Numeric issue. For some reasons Numeric imports dotblas from numarray package. HTH, Pearu From alexander.borghgraef.rma at gmail.com Thu Sep 8 08:18:36 2005 From: alexander.borghgraef.rma at gmail.com (Alexander Borghgraef) Date: Thu, 8 Sep 2005 14:18:36 +0200 Subject: [SciPy-user] More installation troubles In-Reply-To: References: <9e8c52a205090801005dbc0485@mail.gmail.com> Message-ID: <9e8c52a2050908051838abe874@mail.gmail.com> On 9/8/05, Pearu Peterson wrote: > > > scipy 0.3.2 does not have fftw3 support, it's only available in CVS/SVN > version of scipy. You can either ignore it (since scipy uses fortran > fftpack instead fftw), or install fftw2, or get scipy/scipy_core from SVN > repository. Ok, I'll stick to fftpack then, I installed fftw3 for another library, I'm not going to mess with that. This looks like a numarray/Numeric issue. For some reasons Numeric imports > dotblas from numarray package. > Yep. Removing numarray solved that problem. Now I'm back to the include problem described in my other post. -- Alex Borghgraef -------------- next part -------------- An HTML attachment was scrubbed... URL: From pearu at scipy.org Thu Sep 8 07:43:29 2005 From: pearu at scipy.org (Pearu Peterson) Date: Thu, 8 Sep 2005 06:43:29 -0500 (CDT) Subject: [SciPy-user] More installation troubles In-Reply-To: <9e8c52a2050908051838abe874@mail.gmail.com> References: <9e8c52a205090801005dbc0485@mail.gmail.com> <9e8c52a2050908051838abe874@mail.gmail.com> Message-ID: On Thu, 8 Sep 2005, Alexander Borghgraef wrote: > Yep. Removing numarray solved that problem. Now I'm back to the include > problem described in my other post. Try python setup.py build_ext -I/homes/morlet/aborghgr/local/include/python2.3 build HTH, Pearu From alexander.borghgraef.rma at gmail.com Thu Sep 8 09:31:46 2005 From: alexander.borghgraef.rma at gmail.com (Alexander Borghgraef) Date: Thu, 8 Sep 2005 15:31:46 +0200 Subject: [SciPy-user] More installation troubles In-Reply-To: References: <9e8c52a205090801005dbc0485@mail.gmail.com> <9e8c52a2050908051838abe874@mail.gmail.com> Message-ID: <9e8c52a205090806313acbb688@mail.gmail.com> On 9/8/05, Pearu Peterson wrote: > > > > On Thu, 8 Sep 2005, Alexander Borghgraef wrote: > > > > Yep. Removing numarray solved that problem. Now I'm back to the include > > problem described in my other post. > > Try > > python setup.py build_ext -I/homes/morlet/aborghgr/local/include/python2.3 > build > Then I get the following errors: running build_ext Traceback (most recent call last): File "setup.py", line 111, in ? setup_package(ignore_packages) File "setup.py", line 98, in setup_package url = "http://www.scipy.org", File "scipy_core/scipy_distutils/core.py", line 73, in setup return old_setup(**new_attr) File "/usr/lib/python2.3/distutils/core.py", line 149, in setup dist.run_commands() File "/usr/lib/python2.3/distutils/dist.py", line 907, in run_commands self.run_command(cmd) File "/usr/lib/python2.3/distutils/dist.py", line 927, in run_command cmd_obj.run() File "scipy_core/scipy_distutils/command/build_ext.py", line 47, in run raise TypeError,'Extension "%s" sources contains unresolved'\ TypeError: Extension "scipy.xxx.spam" sources contains unresolved items (call build_src before build_ext). I also tried python setup.py config --I/homes/morlet/aborghgr/local/include/python2.3 and then running build, but that didn't do the trick either. I find this very weird: I set the PYTHONINCLUDE variable correctly, tried the config options, and still it uses the wrong compile options. -- Alex Borghgraef -------------- next part -------------- An HTML attachment was scrubbed... URL: From pearu at scipy.org Thu Sep 8 08:37:17 2005 From: pearu at scipy.org (Pearu Peterson) Date: Thu, 8 Sep 2005 07:37:17 -0500 (CDT) Subject: [SciPy-user] More installation troubles In-Reply-To: <9e8c52a205090806313acbb688@mail.gmail.com> References: <9e8c52a205090801005dbc0485@mail.gmail.com> <9e8c52a205090806313acbb688@mail.gmail.com> Message-ID: On Thu, 8 Sep 2005, Alexander Borghgraef wrote: > On 9/8/05, Pearu Peterson wrote: >> >> >> >> On Thu, 8 Sep 2005, Alexander Borghgraef wrote: >> >> >>> Yep. Removing numarray solved that problem. Now I'm back to the include >>> problem described in my other post. >> >> Try >> >> python setup.py build_ext -I/homes/morlet/aborghgr/local/include/python2.3 >> build >> > > Then I get the following errors: > > running build_ext > TypeError: Extension "scipy.xxx.spam" sources contains unresolved items > (call build_src before build_ext). Ok, but did you try python setup.py build_src build_ext -I/homes/morlet/aborghgr/local/include/python2.3 build as suggested in error message? > I also tried > > python setup.py config --I/homes/morlet/aborghgr/local/include/python2.3 > > and then running build, but that didn't do the trick either. I find this > very weird: I set the PYTHONINCLUDE variable correctly, > tried the config options, and still it uses the wrong compile options. scipy_distutils does not use PYTHONINCLUDE or similar environment variables because we have `build_ext -I...`. Pearu From alexander.borghgraef.rma at gmail.com Thu Sep 8 10:17:47 2005 From: alexander.borghgraef.rma at gmail.com (Alexander Borghgraef) Date: Thu, 8 Sep 2005 16:17:47 +0200 Subject: [SciPy-user] More installation troubles In-Reply-To: References: <9e8c52a205090801005dbc0485@mail.gmail.com> <9e8c52a205090806313acbb688@mail.gmail.com> Message-ID: <9e8c52a2050908071762548f6b@mail.gmail.com> On 9/8/05, Pearu Peterson wrote: > > > Ok, but did you try > > python setup.py build_src build_ext > -I/homes/morlet/aborghgr/local/include/python2.3 build > > as suggested in error message? No, I did try running python setup.py build_src separately, but that didn't do it. But now I tried this one, and it works. Compiled and installed all the way. Thanks! -- Alex Borghgraef -------------- next part -------------- An HTML attachment was scrubbed... URL: From yaroslavvb at gmail.com Thu Sep 8 12:23:29 2005 From: yaroslavvb at gmail.com (Yaroslav Bulatov) Date: Thu, 8 Sep 2005 09:23:29 -0700 Subject: [SciPy-user] Setting include dirs In-Reply-To: <9e8c52a20509072349490cb0d6@mail.gmail.com> References: <9e8c52a20509072349490cb0d6@mail.gmail.com> Message-ID: I had a similar problem which was fixed by using "python setup.py install --include-dirs dirs" format Here's Travis' post about this http://www.scipy.org/mailinglists/mailman?fn=scipy-user/2005-September/005199.html On 9/7/05, Alexander Borghgraef wrote: > Hi all, > > I'm trying to install SciPy 0.3.2, but it gets stuck at the compilation of > build/src/fortranobject.c. > The problem seems to be that it doesn't find Numeric/arrayobject.h which is > quite understandable since on the line before the error message I see the > following: > > compile options: '-Ibuild/src -I/usr/include/python2.3 -c' > > Since I don't have root access to this machine, I installed Numeric > locally, so the compile options should contain > -I$HOME/local/inlcude/python2.3 . I did add the local directory to the > PYTHON_INCLUDES environment variable, and the numpy info line at the > beginning of the install run seems to find it too: > > numpy_info: > FOUND: > define_macros = [('NUMERIC_VERSION', '"\\"24.0b2\\""')] > include_dirs = > ['/homes/morlet/aborghgr/local/include/python2.3'] > > Strangely, it does not get included for compilation somehow. Any ideas on > how to fix this? > > > -- > Alex Borghgraef > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user > > > -- Yaroslav Bulatov bulatov at cs.oregonstate.edu Dearborn 102 Oregon State University Corvallis, OR From youjins_scipy at hotmail.com Thu Sep 8 17:26:59 2005 From: youjins_scipy at hotmail.com (Jinsheng you) Date: Thu, 08 Sep 2005 16:26:59 -0500 Subject: [SciPy-user] Gamma distribution questions Message-ID: I have some projects related to the gamma distribution. would you please help me with these: 1) Assuming that I have a gamma distribution dataset, how can I estimate parameters such as alpha, beta from the dataset. 2) assuming that I know the alpha and beta of the distribution, how can I get the value for a given probability. 3) assuming that I know the alpha and beta of the distribution, how can I get the probability for a given x. Thank you so much. _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today - it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/ From rkern at ucsd.edu Thu Sep 8 18:07:57 2005 From: rkern at ucsd.edu (Robert Kern) Date: Thu, 08 Sep 2005 15:07:57 -0700 Subject: [SciPy-user] Gamma distribution questions In-Reply-To: References: Message-ID: <4320B63D.1070401@ucsd.edu> Jinsheng you wrote: > I have some projects related to the gamma distribution. would you > please help me with these: First, please read the documentation in Lib/stats/continuous.pdf in the source distribution. The following probably won't make sense otherwise. > 1) Assuming that I have a gamma distribution dataset, how can I estimate > parameters such as alpha, beta from the dataset. I assume for the rest of this post that alpha is the shape parameter and beta is the scale parameter. Currently, the fit() method of distributions is broken. There's a problem with the way it passes arguments to the nnlf() method; that problem seems to apply to all distributions. There is also a problem with distributions like Gamma which are intrinsically positive; all of the distribution objects take a loc parameter. For intrinsically positive variates like Gamma, this really should be fixed to 0 all of the time. So you are going to have to do a little bit of this manually. from scipy import * def f(params, data): return -sum(log(stats.gamma.pdf(data, params[0], scale=params[1])) And then you can use one of the minimizers in scipy.optimize to minimize f(). That's called the "maximum likelihood method," which may or may not be appropriate for what you want to do. > 2) assuming that I know the alpha and beta of the distribution, how can > I get the value for a given probability. Probability of what? > 3) assuming that I know the alpha and beta of the distribution, how can > I get the probability for a given x. By definition, the probability for any given point value is 0. You can get the value of the probability *density function* by using the pdf() method of scipy.stats.gamma . If you want the probability of getting a value <= x, then the cdf() method will give you that. If you want the probability of getting a value >= x, then 1-cdf(). -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From jmiller at stsci.edu Fri Sep 9 11:12:36 2005 From: jmiller at stsci.edu (Todd Miller) Date: Fri, 09 Sep 2005 11:12:36 -0400 Subject: [SciPy-user] More installation troubles In-Reply-To: References: <9e8c52a205090801005dbc0485@mail.gmail.com> <9e8c52a2050908051838abe874@mail.gmail.com> Message-ID: <1126278755.27040.282.camel@halloween.stsci.edu> On Thu, 2005-09-08 at 07:43, Pearu Peterson wrote: > On Thu, 8 Sep 2005, Alexander Borghgraef wrote: > > > > Yep. Removing numarray solved that problem. Now I'm back to the include > > problem described in my other post. I had some comments and questions about this: 1. numarray and Numeric/scipy coexist fine for me. numarray can't yet replace Numeric in scipy but using both has not been a problem. 2. I can't reproduce the problem with Numeric importing numarray's dotblas during scipy setup but if there is one, we want to fix it. What's your OS and compiler toolset? What versions of numarray/Numeric/scipy did you install? 3. Has anyone else ever seen a problem with numarray's dotblas conflicting with Numeric's dotblas during scipy installation? Thanks, Todd From yaroslavvb at gmail.com Fri Sep 9 14:29:18 2005 From: yaroslavvb at gmail.com (Yaroslav Bulatov) Date: Fri, 9 Sep 2005 11:29:18 -0700 Subject: [SciPy-user] Numerical stability and Legendre polynomials In-Reply-To: <431F3362.9060200@colorado.edu> References: <431F3362.9060200@colorado.edu> Message-ID: Thanks for the insight. The 3-term recursion indeed does much better. I did some experiments to measure exactly how much better. Here's the graph of the number of bits needed to represent maximum error when you project f(x)=0.5 onto Legendre polynomial basis. http://web.engr.oregonstate.edu/~bulatov/research/reports/iet/legendre/numerical-error.png Green graph is the old method using scipy's functions, and blue graph uses 3-term recursion And here's just the graph of 3-term's recursion until 250'th order http://web.engr.oregonstate.edu/~bulatov/research/reports/iet/legendre/numerical-error2.png For some reason the error explodes around 210th order Here's the code used to generate error values k = 30 vals = zeros(200, Float) errors = zeros(k, Float) for j in range(k): ff = lambda x: eval_poly_3term(j, x) coef = (2.*j+1)/2*(integrate.quad(ff, -1, 1))[0] vals+=coef*eval_poly_3term(j, arange(-1,1,0.01)) errors[j] = log(max(max(abs(vals-1)),2**-53))/log(2)+53 > In [33]: for n in range(30): > ....: pn = scipy.special.legendre(n) > ....: print pn(-1.0)-((-1)**n)*pn(1.0) > ....: > 0.0 > 0.0 > 0.0 > 0.0 > -1.11022302463e-15 For some reason when I execute the same code I get 0.0 0.0 0.0 8.881784197e-016 -3.10862446895e-015 I'm using 64 bit arithmetic (floats), scipy version 0.3.2, why do you get higher accuracy? -- Yaroslav Bulatov bulatov at cs.oregonstate.edu Dearborn 102 Oregon State University Corvallis, OR From Fernando.Perez at colorado.edu Fri Sep 9 14:58:59 2005 From: Fernando.Perez at colorado.edu (Fernando Perez) Date: Fri, 09 Sep 2005 12:58:59 -0600 Subject: [SciPy-user] Numerical stability and Legendre polynomials In-Reply-To: References: <431F3362.9060200@colorado.edu> Message-ID: <4321DB73.6020307@colorado.edu> Yaroslav Bulatov wrote: > I'm using 64 bit arithmetic (floats), scipy version 0.3.2, why do you > get higher accuracy? I'm not really. Relative differences O(1e-15) in floating point (these are absoulute, but we're comparing against zero) are essentially numerical noise. Compiler differences, libc differences, whatever. For all intents and purposes, the results are quite similar and I would not worry about the discrepancy. Cheers, f From youjins_scipy at hotmail.com Sat Sep 10 16:15:29 2005 From: youjins_scipy at hotmail.com (Jinsheng you) Date: Sat, 10 Sep 2005 15:15:29 -0500 Subject: [SciPy-user] Gamma distribution questions In-Reply-To: <4320B63D.1070401@ucsd.edu> Message-ID: Dear. Robert Kern, I greatly appreciate your help. I will try this and also read the documentation before any further question. Thanks, Jinsheng You >From: Robert Kern >Reply-To: SciPy Users List >To: SciPy Users List >Subject: Re: [SciPy-user] Gamma distribution questions >Date: Thu, 08 Sep 2005 15:07:57 -0700 > >Jinsheng you wrote: > > I have some projects related to the gamma distribution. would you > > please help me with these: > >First, please read the documentation in Lib/stats/continuous.pdf in the >source distribution. The following probably won't make sense otherwise. > > > 1) Assuming that I have a gamma distribution dataset, how can I estimate > > parameters such as alpha, beta from the dataset. > >I assume for the rest of this post that alpha is the shape parameter and >beta is the scale parameter. > >Currently, the fit() method of distributions is broken. There's a >problem with the way it passes arguments to the nnlf() method; that >problem seems to apply to all distributions. There is also a problem >with distributions like Gamma which are intrinsically positive; all of >the distribution objects take a loc parameter. For intrinsically >positive variates like Gamma, this really should be fixed to 0 all of >the time. > >So you are going to have to do a little bit of this manually. > >from scipy import * >def f(params, data): > return -sum(log(stats.gamma.pdf(data, params[0], scale=params[1])) > >And then you can use one of the minimizers in scipy.optimize to minimize >f(). That's called the "maximum likelihood method," which may or may not >be appropriate for what you want to do. > > > 2) assuming that I know the alpha and beta of the distribution, how can > > I get the value for a given probability. > >Probability of what? > > > 3) assuming that I know the alpha and beta of the distribution, how can > > I get the probability for a given x. > >By definition, the probability for any given point value is 0. You can >get the value of the probability *density function* by using the pdf() >method of scipy.stats.gamma . If you want the probability of getting a >value <= x, then the cdf() method will give you that. If you want the >probability of getting a value >= x, then 1-cdf(). > >-- >Robert Kern >rkern at ucsd.edu > >"In the fields of hell where the grass grows high > Are the graves of dreams allowed to die." > -- Richard Harter > >_______________________________________________ >SciPy-user mailing list >SciPy-user at scipy.net >http://www.scipy.net/mailman/listinfo/scipy-user _________________________________________________________________ Is your PC infected? Get a FREE online computer virus scan from McAfee? Security. http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 From jaroslav.stark at imperial.ac.uk Mon Sep 12 07:24:28 2005 From: jaroslav.stark at imperial.ac.uk (J. Stark) Date: Mon, 12 Sep 2005 12:24:28 +0100 Subject: [SciPy-user] Correct syntax for optimize.fmin_powell Message-ID: Can someone point me to the correct syntax for optimize.fmin_powell when the objective function takes side arguments. I have tried every possibility I can think of, most of which work fine for optimize.fmin, but keep getting error messages, mainly due to an incorrect number of arguments. I have looked at the source, but that hasn't helped. I am typically trying to something like def test_arg((x,y),(a,b)): return (x-a)**2 + (y-b)**2 min = optimize.fmin_powell(test_arg,(1.0,2.0), ((5.0,6.0),)) This works with optimize.fmin, but fails with optimize.fmin_powell. Thanks Jaroslav From rkern at ucsd.edu Mon Sep 12 15:56:06 2005 From: rkern at ucsd.edu (Robert Kern) Date: Mon, 12 Sep 2005 12:56:06 -0700 Subject: [SciPy-user] Correct syntax for optimize.fmin_powell In-Reply-To: References: Message-ID: <4325DD56.5000707@ucsd.edu> J. Stark wrote: > Can someone point me to the correct syntax for optimize.fmin_powell when > the objective function takes side arguments. I have tried every > possibility I can think of, most of which work fine for optimize.fmin, > but keep getting error messages, mainly due to an incorrect number of > arguments. I have looked at the source, but that hasn't helped. > > I am typically trying to something like > > def test_arg((x,y),(a,b)): > return (x-a)**2 + (y-b)**2 > > min = optimize.fmin_powell(test_arg,(1.0,2.0), ((5.0,6.0),)) > > This works with optimize.fmin, but fails with optimize.fmin_powell. Works for me. In [2]: def test_arg((x,y),(a,b)): ...: return (x-a)**2 + (y-b)**2 ...: In [3]: min = optimize.fmin_powell(test_arg,(1.0,2.0), ((5.0,6.0),)) Optimization terminated successfully. Current function value: 0.000000 Iterations: 2 Function evaluations: 22 -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From fonnesbeck at gmail.com Mon Sep 12 20:42:42 2005 From: fonnesbeck at gmail.com (Chris Fonnesbeck) Date: Mon, 12 Sep 2005 20:42:42 -0400 Subject: [SciPy-user] sorting arrays with complex numbers Message-ID: <723eb69305091217425a382ac3@mail.gmail.com> This may not be the appropriate place to post this question, but I'm sure someone has the answer. I am trying to extract the eigenvector with the largest corresponding eigenvalue from an array of eigenvectors. As you would expect, the array of eigenvalues contains complex numbers: >>> d array([ 0.6 +0.j , -0.16037672+0.27778062j, -0.16037672-0.27778062j, 0.32075343+0.j ]) When I try and call max() on this array, I get the following error: >>> max(d) Traceback (most recent call last): File "", line 1, in ? TypeError: no ordering relation is defined for complex numbers What is the appropriate method for extracting the maximum value from this array? Thanks, Chris From rkern at ucsd.edu Mon Sep 12 20:47:09 2005 From: rkern at ucsd.edu (Robert Kern) Date: Mon, 12 Sep 2005 17:47:09 -0700 Subject: [SciPy-user] sorting arrays with complex numbers In-Reply-To: <723eb69305091217425a382ac3@mail.gmail.com> References: <723eb69305091217425a382ac3@mail.gmail.com> Message-ID: <4326218D.1070703@ucsd.edu> Chris Fonnesbeck wrote: > This may not be the appropriate place to post this question, but I'm > sure someone has the answer. I am trying to extract the eigenvector > with the largest corresponding eigenvalue from an array of > eigenvectors. As you would expect, the array of eigenvalues contains > complex numbers: > >>>>d > > array([ 0.6 +0.j , -0.16037672+0.27778062j, > -0.16037672-0.27778062j, 0.32075343+0.j ]) > > When I try and call max() on this array, I get the following error: > >>>>max(d) > > Traceback (most recent call last): > File "", line 1, in ? > TypeError: no ordering relation is defined for complex numbers > > What is the appropriate method for extracting the maximum value from this array? d[argmax(absolute(d))] -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From jdc at uwo.ca Mon Sep 12 22:23:36 2005 From: jdc at uwo.ca (Dan Christensen) Date: Mon, 12 Sep 2005 22:23:36 -0400 Subject: [SciPy-user] Correct syntax for optimize.fmin_powell In-Reply-To: <4325DD56.5000707@ucsd.edu> References: <4325DD56.5000707@ucsd.edu> Message-ID: <871x3tk8nr.fsf@uwo.ca> Robert Kern writes: > J. Stark wrote: > >> I am typically trying to something like >> >> def test_arg((x,y),(a,b)): >> return (x-a)**2 + (y-b)**2 >> >> min = optimize.fmin_powell(test_arg,(1.0,2.0), ((5.0,6.0),)) >> >> This works with optimize.fmin, but fails with optimize.fmin_powell. > > Works for me. > > In [2]: def test_arg((x,y),(a,b)): > ...: return (x-a)**2 + (y-b)**2 > ...: > > In [3]: min = optimize.fmin_powell(test_arg,(1.0,2.0), ((5.0,6.0),)) Doesn't work for me. Debian, python 2.4.1, Debian's scipy packages: ii python2.4-scipy 0.3.2-6 ii python2.4-scipy-core 0.3.2-3 exceptions.TypeError Traceback (most recent call last) /home/jdc/ /usr/lib/python2.4/site-packages/scipy/optimize/optimize.py in fmin_powell(func, x0, args, xtol, ftol, maxiter, maxfun, full_output, disp, retall) 1495 direc1 = direc[i] 1496 fx2 = fval -> 1497 fval, x, direc1 = _linesearch_powell(func, x, direc1, args=args, tol=xtol*100) 1498 if (fx2 - fval) > delta: 1499 delta = fx2 - fval /usr/lib/python2.4/site-packages/scipy/optimize/optimize.py in _linesearch_powell(func, p, xi, args, tol) 1424 extra_args = (func, p, xi) + args 1425 alpha_min, fret, iter, num = brent(_myfunc, args=extra_args, -> 1426 full_output=1, tol=tol) 1427 xi = alpha_min*xi 1428 _powell_funcalls += num /usr/lib/python2.4/site-packages/scipy/optimize/optimize.py in brent(func, args, brack, tol, full_output, maxiter) 1205 _cg = 0.3819660 1206 if brack is None: -> 1207 xa,xb,xc,fa,fb,fc,funcalls = bracket(func, args=args) 1208 elif len(brack) == 2: 1209 xa,xb,xc,fa,fb,fc,funcalls = bracket(func, xa=brack[0], xb=brack[1], args=args) /usr/lib/python2.4/site-packages/scipy/optimize/optimize.py in bracket(func, xa, xb, args, grow_limit) 1358 _gold = 1.618034 1359 _verysmall_num = 1e-21 -> 1360 fa = apply(func, (xa,)+args) 1361 fb = apply(func, (xb,)+args) 1362 if (fa < fb): # Switch so fa > fb /usr/lib/python2.4/site-packages/scipy/optimize/optimize.py in _myfunc(alpha, func, x0, direc, args) 1415 def _myfunc(alpha, func, x0, direc, args=()): 1416 funcargs = (x0 + alpha * direc,)+args -> 1417 return func(*funcargs) 1418 1419 def _linesearch_powell(func, p, xi, args=(), tol=1e-3): TypeError: test_arg() takes exactly 2 arguments (3 given) Dan From rkern at ucsd.edu Tue Sep 13 00:48:43 2005 From: rkern at ucsd.edu (Robert Kern) Date: Mon, 12 Sep 2005 21:48:43 -0700 Subject: [SciPy-user] Correct syntax for optimize.fmin_powell In-Reply-To: <871x3tk8nr.fsf@uwo.ca> References: <4325DD56.5000707@ucsd.edu> <871x3tk8nr.fsf@uwo.ca> Message-ID: <43265A2B.8040702@ucsd.edu> Dan Christensen wrote: > Robert Kern writes: > > >>J. Stark wrote: >> >> >>>I am typically trying to something like >>> >>>def test_arg((x,y),(a,b)): >>> return (x-a)**2 + (y-b)**2 >>> >>>min = optimize.fmin_powell(test_arg,(1.0,2.0), ((5.0,6.0),)) >>> >>>This works with optimize.fmin, but fails with optimize.fmin_powell. >> >>Works for me. >> >>In [2]: def test_arg((x,y),(a,b)): >> ...: return (x-a)**2 + (y-b)**2 >> ...: >> >>In [3]: min = optimize.fmin_powell(test_arg,(1.0,2.0), ((5.0,6.0),)) > > > Doesn't work for me. Debian, python 2.4.1, Debian's scipy packages: [snip] > TypeError: test_arg() takes exactly 2 arguments (3 given) Could you also copy-and-paste the code that you ran to yield this error? Because it looks to me like you may have executed min = optimize.fmin_powell(test_arg,(1.0,2.0), (5.0,6.0)) instead. -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From prabhu_r at users.sf.net Tue Sep 13 13:50:40 2005 From: prabhu_r at users.sf.net (prabhu_r at users.sf.net) Date: Tue, 13 September 2005 13:50:40 IST Subject: [SciPy-user] ANN: MayaVi-1.5 released Message-ID: Hi, This is to announce the availability of the MayaVi Data Visualizer version 1.5. MayaVi is a free, easy to use, scientific data visualizer. It is written in Python, uses the Visualization Toolkit (VTK) for the graphics and provides a GUI written using Tkinter. MayaVi is distributed under a BSD license. It is also cross platform and should run on any platform where both Python and VTK are available (this includes Linux, Windows and Mac OSX). Also bundled with MayaVi is a VTK pipeline browser written in Python and a utility module that makes using VTK easier from the Python interpreter. For more information, sources, binaries, screenshots, installation instructions, documentation etc. visit the MayaVi home page at: Home page: http://mayavi.sourceforge.net SourceForge: http://sourceforge.net/projects/mayavi Download: http://sourceforge.net/project/showfiles.php?group_id=27020&release_id=356146 New in this release ------------------- * Added new LUT editor contributed by Gerald Knizia and Arnd Baecker. The LUT editor allows one to edit the colors in RGB or HSV mode. The editor can also be used standalone. * Added support to easily add user defined sources (works just like the user defined modules and filters). Contributed by Ross Harder. * Support for VolumeProMapper contributed by Daniel James White and Heikki Uuksulainen. * Added a Vorticity filter (contributed by Gareth Clay) that computes the vorticity of input vector data. * Updated MayaVi and vtkPipeline to work with VTK-CVS and VTK-5.x. Support for VTK versions < 4.0 has been removed. MayaVi and the pipeline browser will continue to work with VTK-4.x. * Added a utility vtk_doc.py script that allows one to search the VTK class docs using a GUI. * Fixed severe bug in opening data files belonging to a time-series under Windows. * Fixed various other bugs (details in full changelog). Acknowledgements ---------------- Many thanks to SourceForge for their continued support in hosting MayaVi. Thanks also to various the users who provided feature improvements, patches, bug reports, support and suggestions. Have fun! prabhu From alexander.borghgraef.rma at gmail.com Tue Sep 13 07:50:51 2005 From: alexander.borghgraef.rma at gmail.com (Alexander Borghgraef) Date: Tue, 13 Sep 2005 13:50:51 +0200 Subject: [SciPy-user] More installation troubles In-Reply-To: <1126278755.27040.282.camel@halloween.stsci.edu> References: <9e8c52a205090801005dbc0485@mail.gmail.com> <9e8c52a2050908051838abe874@mail.gmail.com> <1126278755.27040.282.camel@halloween.stsci.edu> Message-ID: <9e8c52a20509130450500bcfc@mail.gmail.com> On 9/9/05, Todd Miller wrote: > > On Thu, 2005-09-08 at 07:43, Pearu Peterson wrote: > > On Thu, 8 Sep 2005, Alexander Borghgraef wrote: > > > > > > > Yep. Removing numarray solved that problem. Now I'm back to the > include > > > problem described in my other post. > > I had some comments and questions about this: > > 1. numarray and Numeric/scipy coexist fine for me. numarray can't yet > replace Numeric in scipy but using both has not been a problem. > > 2. I can't reproduce the problem with Numeric importing numarray's > dotblas during scipy setup but if there is one, we want to fix it. > What's your OS and compiler toolset? Fedora core 2, gcc/++/77 version 3.4.4, python 2.3.4 What versions of > numarray/Numeric/scipy did you install? I installed numarray 1.3.3, Numeric 24.0b2, scipy 0.3.2 from source. 3. Has anyone else ever seen a problem with numarray's dotblas > conflicting with Numeric's dotblas during scipy installation? > Well, I tried installing numarray again, and the conflict doesn't even need scipy. When running the line "import Numeric" in idle I got the following errors: Traceback (most recent call last): File "", line 1, in -toplevel- import Numeric File "/homes/morlet/aborghgr/local//lib/python2.3/site-packages/Numeric/Numeric.py", line 358, in -toplevel- from dotblas import dot, innerproduct, vdot File "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/dotblas.py", line 5, in -toplevel- import generic as _gen File "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/generic.py", line 13, in -toplevel- import numerictypes as _nt File "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/numerictypes.py", line 168, in -toplevel- Byte = _register("Byte", Int8) File "/homes/morlet/aborghgr/local/lib/python2.3/site-packages/numarray/numerictypes.py", line 68, in _register raise ValueError("Type %s has already been registered" % name) ValueError: Type Byte has already been registered -- Alex Borghgraef -------------- next part -------------- An HTML attachment was scrubbed... URL: From jdc at uwo.ca Tue Sep 13 08:27:06 2005 From: jdc at uwo.ca (Dan Christensen) Date: Tue, 13 Sep 2005 08:27:06 -0400 Subject: [SciPy-user] Correct syntax for optimize.fmin_powell In-Reply-To: <43265A2B.8040702@ucsd.edu> References: <4325DD56.5000707@ucsd.edu> <871x3tk8nr.fsf@uwo.ca> <43265A2B.8040702@ucsd.edu> Message-ID: <87wtlli25h.fsf@uwo.ca> Robert Kern writes: > Dan Christensen wrote: > >> TypeError: test_arg() takes exactly 2 arguments (3 given) > > Could you also copy-and-paste the code that you ran to yield this error? > Because it looks to me like you may have executed > > min = optimize.fmin_powell(test_arg,(1.0,2.0), (5.0,6.0)) > > instead. Nope, I copy-and-pasted the code from your message. And I've also had this bug affect me in completely different code as well. Just like the original poster, args works fine for all the other fmin* routines, but not for fmin_powell. Dan From nwagner at mecha.uni-stuttgart.de Tue Sep 13 08:32:20 2005 From: nwagner at mecha.uni-stuttgart.de (Nils Wagner) Date: Tue, 13 Sep 2005 14:32:20 +0200 Subject: [SciPy-user] Correct syntax for optimize.fmin_powell In-Reply-To: <87wtlli25h.fsf@uwo.ca> References: <4325DD56.5000707@ucsd.edu> <871x3tk8nr.fsf@uwo.ca> <43265A2B.8040702@ucsd.edu> <87wtlli25h.fsf@uwo.ca> Message-ID: <4326C6D4.70601@mecha.uni-stuttgart.de> Dan Christensen wrote: >Robert Kern writes: > > >>Dan Christensen wrote: >> >> >>>TypeError: test_arg() takes exactly 2 arguments (3 given) >>> >>Could you also copy-and-paste the code that you ran to yield this error? >>Because it looks to me like you may have executed >> >> min = optimize.fmin_powell(test_arg,(1.0,2.0), (5.0,6.0)) >> >>instead. >> > >Nope, I copy-and-pasted the code from your message. > >And I've also had this bug affect me in completely different code >as well. Just like the original poster, args works fine for all >the other fmin* routines, but not for fmin_powell. > >Dan > >_______________________________________________ >SciPy-user mailing list >SciPy-user at scipy.net >http://www.scipy.net/mailman/listinfo/scipy-user > python -i powell.py results in Optimization terminated successfully. Current function value: 0.000000 Iterations: 2 Function evaluations: 22 >>> >>> min array([ 5., 6.]) BTW >>> scipy.__version__ '0.3.3_309.4626' Nils -------------- next part -------------- A non-text attachment was scrubbed... Name: powell.py Type: text/x-python Size: 137 bytes Desc: not available URL: From meesters at uni-mainz.de Tue Sep 13 11:17:30 2005 From: meesters at uni-mainz.de (Christian Meesters) Date: Tue, 13 Sep 2005 17:17:30 +0200 Subject: [SciPy-user] scipy on SUSE with Python 2.4? Message-ID: <303851131d702ddb7dc8a0d0ea5b66d6@uni-mainz.de> Hi Anybody scipy running on SUSE (or a similar distribution) with Python 2.4? How was this accomplished? Any hints on how to compile scipy (including ATLAS)? (I have no idea how to install it on SUSE - it's my first time - and I'm anticipating pitfalls ...) TIA Christian From nwagner at mecha.uni-stuttgart.de Tue Sep 13 11:21:42 2005 From: nwagner at mecha.uni-stuttgart.de (Nils Wagner) Date: Tue, 13 Sep 2005 17:21:42 +0200 Subject: [SciPy-user] scipy on SUSE with Python 2.4? In-Reply-To: <303851131d702ddb7dc8a0d0ea5b66d6@uni-mainz.de> References: <303851131d702ddb7dc8a0d0ea5b66d6@uni-mainz.de> Message-ID: <4326EE86.8040506@mecha.uni-stuttgart.de> Christian Meesters wrote: > Hi > > Anybody scipy running on SUSE (or a similar distribution) with Python > 2.4? How was this accomplished? Any hints on how to compile scipy > (including ATLAS)? > (I have no idea how to install it on SUSE - it's my first time - and > I'm anticipating pitfalls ...) > > TIA > Christian > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user Please follow the instructions available at http://www.scipy.org/documentation/buildscipy.txt It is straightforward :-) Nils From novak at ucolick.org Wed Sep 14 02:41:46 2005 From: novak at ucolick.org (Greg Novak) Date: Tue, 13 Sep 2005 23:41:46 -0700 (PDT) Subject: [SciPy-user] Scipy Platform Message-ID: Ah, yes, I nearly forgot that I promised to get back to the list about what platforms people were using for scientific computing. Mandrake, Gentoo, Fedora, Windows, Quantian, and OS X were mentioned. Debian was the big winner, getting twice as many votes as the other distributions (ie, two votes instead of one). Cheers, Greg From falted at pytables.org Wed Sep 14 14:16:39 2005 From: falted at pytables.org (Francesc Altet) Date: Wed, 14 Sep 2005 20:16:39 +0200 Subject: [SciPy-user] [ANN] PyTables 1.1.1 Message-ID: <200509142016.39351.falted@pytables.org> ========================== Announcing PyTables 1.1.1 ========================== This is a maintenance release of PyTables. In it, several optimizations and bug fixes have been made. As some of the fixed bugs were quite important, it's strongly recommended for users to upgrade. Go to the PyTables web site for downloading the beast: http://pytables.sourceforge.net/ or keep reading for more info about the improvements and bugs fixed. Changes more in depth ===================== Improvements: - Optimized the opening of files with a large number of objects. Now, files with table objects open a 50% faster, and files with arrays open more than twice as fast (up to 2000 objects/s on a Pentium 4 at 2GHz). Hence, a file with a combination of both kinds of objects opens between a 50% and 100% faster than in 1.1. - Optimized the creation of ``NestedRecArray`` objects using ``NumArray`` objects as columns, so that filling a table with the ``Table.append()`` method achieves a performance similar to PyTables pre-1.1 releases. Bug fixes: - ``Table.readCoordinates()`` now converts the coords parameter into ``Int64`` indices automatically. - Fixed a bug that prevented appending to tables (though ``Table.append()``) using a list of ``NumArray`` objects. - ``Int32`` attributes are handled correctly in 64-bit platforms now. - Correction for accepting lists of numarrays as input for ``NestedRecArrays``. - Fixed a problem when creating rank 1 multi-dimensional string columns in ``Table`` objects. Closes SF bug #1269023. - Avoid errors when unpickling objects stored in attributes. See the section ``AttributeSet`` in the reference chapter of the User's Manual for more information. Closes SF bug #1254636. - Assignment for ``*Array`` slices has been improved in order to solve some issues with shapes. Closes SF bug #1288792. - The indexation properties were lost in case the table was closed before an index was created. Now, these properties are saved even in this case. Known bugs: - Classes inheriting from ``IsDescription`` subclasses do not inherit columns defined in the super-class. See SF bug #1207732 for more info. - Time datatypes are non-portable between big-endian and little-endian architectures. This is ultimately a consequence of a HDF5 limitation. See SF bug #1234709 for more info. Backward-incompatible changes: - None (that we are aware of). Important note for MacOSX users =============================== UCL compressor works badly on MacOSX platforms. Recent investigation seems to point to a bug in the development tools in MacOSX. Until the problem is isolated and eventually solved, UCL support will not be compiled by default on MacOSX platforms, even if the installer finds it in the system. However, if you still want to get UCL support on MacOSX, you can use the ``--force-ucl`` flag in ``setup.py``. Important note for Python 2.4 and Windows users =============================================== If you are willing to use PyTables with Python 2.4 in Windows platforms, you will need to get the HDF5 library compiled for MSVC 7.1, aka .NET 2003. It can be found at: ftp://ftp.ncsa.uiuc.edu/HDF/HDF5/current/bin/windows/5-164-win-net.ZIP Users of Python 2.3 on Windows will have to download the version of HDF5 compiled with MSVC 6.0 available in: ftp://ftp.ncsa.uiuc.edu/HDF/HDF5/current/bin/windows/5-164-win.ZIP What it is ========== **PyTables** is a package for managing hierarchical datasets and designed to efficiently cope with extremely large amounts of data (with support for full 64-bit file addressing). It features an object-oriented interface that, combined with C extensions for the performance-critical parts of the code, makes it a very easy-to-use tool for high performance data storage and retrieval. PyTables runs on top of the HDF5 library and numarray (Numeric is also supported) package for achieving maximum throughput and convenient use. Besides, PyTables I/O for table objects is buffered, implemented in C and carefully tuned so that you can reach much better performance with PyTables than with your own home-grown wrappings to the HDF5 library. PyTables sports indexing capabilities as well, allowing doing selections in tables exceeding one billion of rows in just seconds. Platforms ========= This version has been extensively checked on quite a few platforms, like Linux on Intel32 (Pentium), Win on Intel32 (Pentium), Linux on Intel64 (Itanium2), FreeBSD on AMD64 (Opteron), Linux on PowerPC and MacOSX on PowerPC. For other platforms, chances are that the code can be easily compiled and run without further problems. Please, contact us in case you are experiencing problems. Resources ========= Go to the PyTables web site for more details: http://pytables.sourceforge.net/ About the HDF5 library: http://hdf.ncsa.uiuc.edu/HDF5/ About numarray: http://www.stsci.edu/resources/software_hardware/numarray To know more about the company behind the PyTables development, see: http://www.carabos.com/ Acknowledgments =============== Thanks to various the users who provided feature improvements, patches, bug reports, support and suggestions. See THANKS file in 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, a big thank you to THG (http://www.hdfgroup.org/) for sponsoring many of the new features recently introduced in PyTables. Share your experience ===================== Let us know of any bugs, suggestions, gripes, kudos, etc. you may have. ---- **Enjoy data!** -- The PyTables Team From ckkart at hoc.net Thu Sep 15 10:59:41 2005 From: ckkart at hoc.net (Christian Kristukat) Date: Thu, 15 Sep 2005 16:59:41 +0200 Subject: [SciPy-user] io.read_array slow Message-ID: <43298C5D.3080804@hoc.net> Hi, I noticed that io.read_array is really slow compared to a python while/readline/split loop (about 5 times slower). I believe to remember that at some time it was written in C but when I looked at the source of scipy 0.3.2 it seems to be pure python. Is there a evident reason why it must be that slow or are there maybe plans to rewrite it in C? Regards, Christian From oliphant at ee.byu.edu Thu Sep 15 12:59:40 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Thu, 15 Sep 2005 10:59:40 -0600 Subject: [SciPy-user] io.read_array slow In-Reply-To: <43298C5D.3080804@hoc.net> References: <43298C5D.3080804@hoc.net> Message-ID: <4329A87C.80903@ee.byu.edu> Christian Kristukat wrote: > > Hi, > I noticed that io.read_array is really slow compared to a python > while/readline/split loop (about 5 times slower). I believe to > remember that at some time it was written in C but when I looked at > the source of scipy 0.3.2 it seems to be pure python. Is there a > evident reason why it must be that slow or are there maybe plans to > rewrite it in C? It's slower because it does more, than that simple loop. The old complexity-speed trade off. No doubt there are optimizations that can be done. If you find any, please let us know... Notice that the new scipy core ndarray object has a builtin (simple) ascii file reader for each type. This will be very fast for simple cases, but still is not perfect. -Travis From ckkart at hoc.net Fri Sep 16 07:11:46 2005 From: ckkart at hoc.net (Christian Kristukat) Date: Fri, 16 Sep 2005 13:11:46 +0200 Subject: [SciPy-user] io.read_array slow In-Reply-To: <4329A87C.80903@ee.byu.edu> References: <43298C5D.3080804@hoc.net> <4329A87C.80903@ee.byu.edu> Message-ID: <432AA872.8060302@hoc.net> Travis Oliphant wrote: > Christian Kristukat wrote: > >> >> Hi, >> I noticed that io.read_array is really slow compared to a python >> while/readline/split loop (about 5 times slower). I believe to >> remember that at some time it was written in C but when I looked at >> the source of scipy 0.3.2 it seems to be pure python. Is there a >> evident reason why it must be that slow or are there maybe plans to >> rewrite it in C? > > > It's slower because it does more, than that simple loop. The old > complexity-speed trade off. No doubt there are optimizations that can > be done. If you find any, please let us know... I try to do so if I find the time. But where did the old C-code go? Or do I remember wrong? > Notice that the new scipy core ndarray object has a builtin (simple) > ascii file reader for each type. This will be very fast for simple > cases, but still is not perfect. Does 'the new scipy_core' mean it should be in the current svn version? I didn't find an ndarray there. Regards, Christian From jr at sun.ac.za Fri Sep 16 10:44:58 2005 From: jr at sun.ac.za (Johann Rohwer) Date: Fri, 16 Sep 2005 16:44:58 +0200 Subject: [SciPy-user] scipy.gplt PNG output without plotting on screen Message-ID: <200509161644.58206.jr@sun.ac.za> Hi Is there a way using scipy.gplt to output your graphs to files without having them plotted to screen first? I have a script that generates lots of plots, and currently run a whole sequence of commands to build up every plot, e.g.: >>> scipy.gplt.plot(x,y) >>> scipy.gplt.xtitle('xxx') >>> scipy.gplt.ytitle('yyy') >>> scipy.gplt.logx('on') >>> scipy.gplt.save('file.png') I find that updating the plot on screen takes up a lot of time which almost doubles the total execution time of my script. Johann From haase at msg.ucsf.edu Fri Sep 16 16:45:14 2005 From: haase at msg.ucsf.edu (Sebastian Haase) Date: Fri, 16 Sep 2005 13:45:14 -0700 Subject: [SciPy-user] licence: GNU or not GNU for new image analysis platform Message-ID: <200509161345.14702.haase@msg.ucsf.edu> Prabhu, I noticed that you changed MayaVi's license from GPL to BSD. Why did you think this was needed ? I have developed (rather: put together) a platform for (interactive) multidimensional image analysis and algorithm development. It's a graphical shell window based on wxPython's PyShell where I added a nice, fast OpenGL-based 2d image viewer. But the main aspect of all this was that it comes pre-compiled and ready to use ! (Including numarray, wxPython ... for Linux or Windows or Mac-OSX) Now: before I can give this to anyone my university requires me to attach a license to this. I think I like the "copyleft" idea of GPL or maybe better LGPL. My understanding is that this would allow a company to sell it, but they would have "make their changes publicly available" (under LGPL again) I think Fernando's IPython is LGPL - and I think my "platform" is quite a similar spirit... All comments are most welcome Thanks, Sebastian Haase UCSF, Sedat Lab From Fernando.Perez at colorado.edu Fri Sep 16 17:11:44 2005 From: Fernando.Perez at colorado.edu (Fernando Perez) Date: Fri, 16 Sep 2005 15:11:44 -0600 Subject: [SciPy-user] licence: GNU or not GNU for new image analysis platform In-Reply-To: <200509161345.14702.haase@msg.ucsf.edu> References: <200509161345.14702.haase@msg.ucsf.edu> Message-ID: <432B3510.6020201@colorado.edu> Sebastian Haase wrote: > Now: before I can give this to anyone my university requires me to attach a > license to this. I think I like the "copyleft" idea of GPL or maybe better > LGPL. > My understanding is that this would allow a company to sell it, but they would > have "make their changes publicly available" (under LGPL again) > > I think Fernando's IPython is LGPL - and I think my "platform" is quite a > similar spirit... Nope, BSD as well. It used to be LGPL, but I changed to BSD a year ago. My reasoning was multiple-fold (I'm speaking for myself here, NOT for Prabhu): - Python itself is BSD-like. I'd rather have a license that's as python-compatible as possible. - It lowers the 'barrier to adoption', since some people/institutions have blanket no-GPL policies. - scipy is BSD, and I like scipy :) More seriously, ipython is so heavily used by scipy users, that I wanted to harmonize with it as much as possible. - matplotlib is BSD, and I also want very tight integration with matplotlib. Compatible licenses make this much easier. - Even if this opens the possibility of someone 'stealing' ipython, I don't care. I suspect that in the long run, it's more beneficial for people who may use it to contribute back their changes. There is a substantial cost in maintaining a branched copy of a project without contributing, because the burden of integrating the project's mainline development falls on you. Unless you are trying to flat out fork a project, in most cases it's easier to just send your changes back to the mainline. - John Hunter's experience with matplotlib supports that last point: he's had private companies contribute back mpl enhancements, simply because it's not worth their time to keep them separate and have to manually track matplotlib in a private branch. You can see the original messages on this in the ipython lists: http://www.scipy.net/pipermail/ipython-dev/2004-October/000270.html http://www.scipy.net/pipermail/ipython-user/2004-October/000494.html I sent these to everyone who had ever contributed to ipython, and got no negative replies, so I didn't have to rewrite any code myself (though I was prepared to if I had to). I think there is a place for the GPL, and I'm not fundamentally anti-GPL. But in my case, I felt the benefits of a BSD approach were significant, and I went with it. Ultimately the decsion is yours, though keep in mind that given that scipy, mayavi, matplotlib, ipython (and the new wx-based graphical ipython) are all BSD, the room for integration with other tools for your project will grow with BSD. A *GPL license will to some extent 'wall off' your project from these others. The decision of what to do is fully yours. Cheers, f ps - Just to be clear: I am NOT engaging into a license discussion/war. I won't reply to any message in that direction. I was simply clarifying for Sebastian's benefit the reasoning behind ipython's licensing, since he was under the impression (historically correct, but currently not anymore) that ipython was LGPL. If you feel very strongly about this topic, there are many places on the net where you can discuss it. From haase at msg.ucsf.edu Fri Sep 16 17:27:09 2005 From: haase at msg.ucsf.edu (Sebastian Haase) Date: Fri, 16 Sep 2005 14:27:09 -0700 Subject: [SciPy-user] licence: GNU or not GNU for new image analysis platform In-Reply-To: <432B3510.6020201@colorado.edu> References: <200509161345.14702.haase@msg.ucsf.edu> <432B3510.6020201@colorado.edu> Message-ID: <200509161427.09654.haase@msg.ucsf.edu> Fernando, Thanks a lot for the reply: My ONLY reasoning was that it might be easier to get permission from our "patent"-people to use LGPL than BSD (but after all: UCSF is part of UC; just like Berkeley - where BSD is from !?) No need to worry about license war - but I'm just wondering how other people deal with their employers/universities if they want to give away (some) results of their work: at least here, everybody seems very concerned about someone ELSE getting very rich !! ;-) Thanks again, Sebastian On Friday 16 September 2005 14:11, Fernando Perez wrote: > Sebastian Haase wrote: > > Now: before I can give this to anyone my university requires me to attach > > a license to this. I think I like the "copyleft" idea of GPL or maybe > > better LGPL. > > My understanding is that this would allow a company to sell it, but they > > would have "make their changes publicly available" (under LGPL again) > > > > I think Fernando's IPython is LGPL - and I think my "platform" is quite > > a similar spirit... > > Nope, BSD as well. It used to be LGPL, but I changed to BSD a year ago. > My reasoning was multiple-fold (I'm speaking for myself here, NOT for > Prabhu): > > - Python itself is BSD-like. I'd rather have a license that's as > python-compatible as possible. > > - It lowers the 'barrier to adoption', since some people/institutions have > blanket no-GPL policies. > > - scipy is BSD, and I like scipy :) More seriously, ipython is so heavily > used by scipy users, that I wanted to harmonize with it as much as > possible. > > - matplotlib is BSD, and I also want very tight integration with > matplotlib. Compatible licenses make this much easier. > > - Even if this opens the possibility of someone 'stealing' ipython, I don't > care. I suspect that in the long run, it's more beneficial for people who > may use it to contribute back their changes. > > There is a substantial cost in maintaining a branched copy of a project > without contributing, because the burden of integrating the project's > mainline development falls on you. Unless you are trying to flat out fork > a project, in most cases it's easier to just send your changes back to the > mainline. > > - John Hunter's experience with matplotlib supports that last point: he's > had private companies contribute back mpl enhancements, simply because it's > not worth their time to keep them separate and have to manually track > matplotlib in a private branch. > > You can see the original messages on this in the ipython lists: > > http://www.scipy.net/pipermail/ipython-dev/2004-October/000270.html > http://www.scipy.net/pipermail/ipython-user/2004-October/000494.html > > I sent these to everyone who had ever contributed to ipython, and got no > negative replies, so I didn't have to rewrite any code myself (though I was > prepared to if I had to). > > I think there is a place for the GPL, and I'm not fundamentally anti-GPL. > But in my case, I felt the benefits of a BSD approach were significant, and > I went with it. > > Ultimately the decsion is yours, though keep in mind that given that scipy, > mayavi, matplotlib, ipython (and the new wx-based graphical ipython) are > all BSD, the room for integration with other tools for your project will > grow with BSD. A *GPL license will to some extent 'wall off' your project > from these others. The decision of what to do is fully yours. > > Cheers, > > f > > ps - Just to be clear: I am NOT engaging into a license discussion/war. I > won't reply to any message in that direction. I was simply clarifying for > Sebastian's benefit the reasoning behind ipython's licensing, since he was > under the impression (historically correct, but currently not anymore) that > ipython was LGPL. If you feel very strongly about this topic, there are > many places on the net where you can discuss it. > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user From Fernando.Perez at colorado.edu Fri Sep 16 17:38:28 2005 From: Fernando.Perez at colorado.edu (Fernando Perez) Date: Fri, 16 Sep 2005 15:38:28 -0600 Subject: [MayaVi-users] Re: [SciPy-user] licence: GNU or not GNU for new image analysis platform In-Reply-To: <200509161427.09654.haase@msg.ucsf.edu> References: <200509161345.14702.haase@msg.ucsf.edu> <432B3510.6020201@colorado.edu> <200509161427.09654.haase@msg.ucsf.edu> Message-ID: <432B3B54.9010404@colorado.edu> Sebastian Haase wrote: > Fernando, > Thanks a lot for the reply: > My ONLY reasoning was that it might be easier to get permission from our > "patent"-people to use LGPL than BSD (but after all: UCSF is part of UC; just > like Berkeley - where BSD is from !?) > > No need to worry about license war - but I'm just wondering how other people > deal with their employers/universities if they want to give away (some) > results of their work: at least here, everybody seems very concerned about > someone ELSE getting very rich !! ;-) My comment wasn't directed at you specifically :) Just pre-empting a discussion that is never-ending if you let it start, and that I don't have time for right now. Good luck with your legal people! Best, f From aisaac at american.edu Fri Sep 16 18:07:56 2005 From: aisaac at american.edu (Alan G Isaac) Date: Fri, 16 Sep 2005 18:07:56 -0400 Subject: =?ISO-8859-1?Q?Re[2]:=20[SciPy-user]=20licence:=20GNU=20or=20not=20GNU=20f?==?ISO-8859-1?Q?or=20new=20image=20analysis=09platform?= In-Reply-To: <200509161427.09654.haase@msg.ucsf.edu> References: <200509161345.14702.haase@msg.ucsf.edu><432B3510.6020201@colorado.edu><200509161427.09654.haase@msg.ucsf.edu> Message-ID: On Fri, 16 Sep 2005, Sebastian Haase apparently wrote: > I'm just wondering how other people deal with their > employers/universities if they want to give away (some) > results of their work: at least here, everybody seems very > concerned about someone ELSE getting very rich! This corruption of academe is partly due to the well intentioned but misguided Bayh-Dole Act, as well as a paradigm shift in academe which has undermined its culture of open exchange. Fortunately a backlash is gathering. Remind your employers what a university is. Maybe ask them if they think Pitt should have forced Salk to patent his polio vaccine. If they say yes, walk away. Cheers, Alan Isaac From perry at stsci.edu Fri Sep 16 18:24:48 2005 From: perry at stsci.edu (Perry Greenfield) Date: Fri, 16 Sep 2005 18:24:48 -0400 Subject: [SciPy-user] licence: GNU or not GNU for new image analysis platform In-Reply-To: <200509161427.09654.haase@msg.ucsf.edu> References: <200509161345.14702.haase@msg.ucsf.edu> <432B3510.6020201@colorado.edu> <200509161427.09654.haase@msg.ucsf.edu> Message-ID: On Sep 16, 2005, at 5:27 PM, Sebastian Haase wrote: > Fernando, > Thanks a lot for the reply: > My ONLY reasoning was that it might be easier to get permission from > our > "patent"-people to use LGPL than BSD (but after all: UCSF is part of > UC; just > like Berkeley - where BSD is from !?) > > No need to worry about license war - but I'm just wondering how other > people > deal with their employers/universities if they want to give away > (some) > results of their work: at least here, everybody seems very concerned > about > someone ELSE getting very rich !! ;-) > > Thanks again, > Sebastian > If I'm not mistaken, nothing prevents you from taking BSD software, adding stuff to it and relicensing it as GPL (that's what the whole Python 2.0/GPL compatibility hubbub was all about, allowing that very thing). BSD permits that (but IANAL). Having said that, I agree with Fernando, the worry about others getting rich off of this sort of software is usually way overblown (And even if they did, does it matter? It doesn't mean you or your organization would have made a cent off of it anyway. There's usually a lot more involved to making that money than just writing the code.) Perry From alford at wuphys.wustl.edu Fri Sep 16 18:28:05 2005 From: alford at wuphys.wustl.edu (Mark Alford) Date: Fri, 16 Sep 2005 17:28:05 -0500 (CDT) Subject: [SciPy-user] relaxation solver Message-ID: I want to solve a boundary value ODE problem, where I know some boundary conditions at each end of the interval. As far as I know, odeint is just for initial value problems where you know boundary conditions at one end of the interval. Is there a relaxation solver in SciPy? From prabhu_r at users.sf.net Sat Sep 17 00:08:51 2005 From: prabhu_r at users.sf.net (Prabhu Ramachandran) Date: Sat, 17 Sep 2005 09:38:51 +0530 Subject: [SciPy-user] licence: GNU or not GNU for new image analysis platform In-Reply-To: <200509161345.14702.haase@msg.ucsf.edu> References: <200509161345.14702.haase@msg.ucsf.edu> Message-ID: <17195.38611.78174.707975@monster.linux.in> >>>>> "Sebastian" == Sebastian Haase writes: Sebastian> Prabhu, I noticed that you changed MayaVi's license Sebastian> from GPL to BSD. Why did you think this was needed ? That was a long time ago. Here are some messages from the mayavi-users m/l regarding the change. Initial thoughts: http://sourceforge.net/mailarchive/message.php?msg_id=109733 After more thought: http://sourceforge.net/mailarchive/message.php?msg_id=1527518 The clincher (links to a detailed essay on GPL vs. BSD): http://sourceforge.net/mailarchive/message.php?msg_id=1609833 My original reason for using GPL was to ensure that devious commercial interests did not abuse MayaVi's freedom. My reasons were that this might be the best for the community. However, I was proved wrong by companies like Enthought and Kitware who believe in OSS and release code as BSD. They have helped MayaVi enormously. To the point that the *entire* TVTK and MayaVi2 effort is funded and supported by Enthought! If not for Kitware's VTK, MayaVi would not exist. The main reasons are: 1. If someone decides to be nasty and illegally "steals" from a GPL project, the only thing you have is the ability to sue the person/company. Do you have the time for that, or do you really care? Unless you are a company, not a chance. I think we have much better things to do. So, if someone did steal, there is little you can do, so why not just make it free? 2. When you write BSD software you allow companies to get involved in your project without being worried about hassles with licensing. This makes it easy for them to get involved. Once they are involved, it is usually much easier for them to support the project rather than fork the code (as has been eloquently explained by others). Net result is that everyone gains. Unless you write software for money and not the community, you usually are happy if a company tells you "hey your code is cool, we used it in this product". Finally, all this applies to software that is of utility to the scientific community. It may or may not apply to all software. Ultimately, you have to decide and see how comfortable you are with it. All I can say is that I am very happy with my decision to switch MayaVi from GPL to BSD. Sebastian> I have developed (rather: put together) a platform for Sebastian> (interactive) multidimensional image analysis and Sebastian> algorithm development. It's a graphical shell window Sebastian> based on wxPython's PyShell where I added a nice, fast Sebastian> OpenGL-based 2d image viewer. But the main aspect of Sebastian> all this was that it comes pre-compiled and ready to Sebastian> use ! (Including numarray, wxPython ... for Linux or Sebastian> Windows or Mac-OSX) Very cool! If you have the chance do take a look at Enthought's Envisage application framework. If you will be there at SciPy06, you could probably see it in action. Good luck! cheers, prabhu From alopez at imim.es Sun Sep 18 04:07:24 2005 From: alopez at imim.es (LOPEZ GARCIA DE LOMANA, ADRIAN) Date: Sun, 18 Sep 2005 10:07:24 +0200 Subject: [SciPy-user] using odeint Message-ID: <66373AD054447F47851FCC5EB49B36113E27A1@basquet.imim.es> Hi people, I'm using odeint from SciPy to solve a a system of ordinary differential equations. I've found a quite a strange error while taking different parameters on the equations: 0.0 0.0 0.0 0.0 0.0 Traceback (most recent call last): File "/home/alopez/tmp/news/output/scratch/modelF.integ.py", line 23, in func xdot[7] = + (v_NACT * x[5]**m_NACT) / (x[5]**m_NACT + k_NACT**m_NACT) - k_deg_h * x[7] ValueError: negative number cannot be raised to a fractional power odepack.error: Error occured while calling the Python function named func 0.0 Traceback (most recent call last): File "/home/alopez/tmp/news/output/scratch/modelF.integ.py", line 23, in func xdot[7] = + (v_NACT * x[5]**m_NACT) / (x[5]**m_NACT + k_NACT**m_NACT) - k_deg_h * x[7] ValueError: negative number cannot be raised to a fractional power odepack.error: Error occured while calling the Python function named func 5.03700957163e-99 Traceback (most recent call last): File "/home/alopez/tmp/news/output/scratch/modelF.integ.py", line 23, in func xdot[7] = + (v_NACT * x[5]**m_NACT) / (x[5]**m_NACT + k_NACT**m_NACT) - k_deg_h * x[7] ValueError: negative number cannot be raised to a fractional power odepack.error: Error occured while calling the Python function named func 7.37964481865e-99 Traceback (most recent call last): those numbers in between the error messages are prints of x[5]. Why is telling me about negative numbers? It's not, actually, close to zero, but positive. k_NACT is also positive. k_NACT = 2.994997603 Some ideas? Thanks for your time, Adri?n. From rubencharles at gmail.com Mon Sep 19 13:09:44 2005 From: rubencharles at gmail.com (Ruben Charles) Date: Mon, 19 Sep 2005 13:09:44 -0400 Subject: [SciPy-user] plotting questions (gplt and plt) Message-ID: Hi, I'm using Scipy on Debian Gnu/Linux. (python2.3-scipy) in a program that have to show plots but while i'm testing capabilities of the three possible options (gplt, plt an xplt) a couple of questions comes up. -------------------------------------------------------------------------------------------------------------- plt: If a follow the steps of the plotting tutorial everything seems ok in the first try. >>> import gui_thread >>> gui_thread.start() # or gui_thread.wxPython_thread() >>> from scipy import plt >>> plt.plot((1,2,3)) But if i close the Plot frame (manually with the cursor) and try to plot again >>> plt.plot((1,2,3)) The interpreter shows a message: "The C++ part of the code has been deleted" and i can?t plot again without exit the interpreter and back again. How i can solve that? ---------------------------------------------------------------------------------------------------------------- gplt (gnuplot): If i create a plot with the Gnuplot program the resulting plot has responses to mouse events and shows coordinates in the window but if i create a plot in python gplt.plot((1,2,3)) The window have not the same response to mouse events (no response at all). How i can give the same responses? ___________________________________________________________________ Thanks From ckkart at hoc.net Tue Sep 20 04:33:40 2005 From: ckkart at hoc.net (Christian Kristukat) Date: Tue, 20 Sep 2005 10:33:40 +0200 Subject: [SciPy-user] plotting questions (gplt and plt) In-Reply-To: References: Message-ID: <432FC964.8@hoc.net> Ruben Charles wrote: > Hi, > > I'm using Scipy on Debian Gnu/Linux. (python2.3-scipy) in a program > that have to show plots but while i'm testing capabilities of the > three possible options (gplt, plt an xplt) a couple of questions comes > up. > > -------------------------------------------------------------------------------------------------------------- > plt: > > If a follow the steps of the plotting tutorial everything seems ok in > the first try. > > >>>>import gui_thread >>>>gui_thread.start() # or gui_thread.wxPython_thread() >>>>from scipy import plt >>>>plt.plot((1,2,3)) > > > > But if i close the Plot frame (manually with the cursor) and try to plot again > > >>>>plt.plot((1,2,3)) > > > > The interpreter shows a message: "The C++ part of the code has been > deleted" and i can?t plot again without exit the interpreter and back > again. > > > How i can solve that? > > ---------------------------------------------------------------------------------------------------------------- > > gplt (gnuplot): > > If i create a plot with the Gnuplot program the resulting plot has > responses to mouse events and shows coordinates in the window but if i > create a plot in python > > gplt.plot((1,2,3)) > > The window have not the same response to mouse events (no response at all). > > How i can give the same responses? Just in case you haven't heard about matplotlib (matplotlib.sf.net) you should try that too. There's heavy development going on and it has a very active mailing list. Christian From B.P.S.Thurin at city.ac.uk Tue Sep 20 04:29:13 2005 From: B.P.S.Thurin at city.ac.uk (Brice Thurin) Date: Tue, 20 Sep 2005 10:29:13 +0200 Subject: [SciPy-user] plotting questions (gplt and plt) In-Reply-To: References: Message-ID: <432FC859.4060000@city.ac.uk> Hi, When I was running into problem with plt, xplt and gnuplot, someone suggest to me to use matplotlib, which I did... and I am quite satisfied. Ruben Charles wrote: >Hi, > >I'm using Scipy on Debian Gnu/Linux. (python2.3-scipy) in a program >that have to show plots but while i'm testing capabilities of the >three possible options (gplt, plt an xplt) a couple of questions comes >up. > >-------------------------------------------------------------------------------------------------------------- >plt: > >If a follow the steps of the plotting tutorial everything seems ok in >the first try. > > > >>>>import gui_thread >>>>gui_thread.start() # or gui_thread.wxPython_thread() >>>>from scipy import plt >>>>plt.plot((1,2,3)) >>>> >>>> > > >But if i close the Plot frame (manually with the cursor) and try to plot again > > > >>>>plt.plot((1,2,3)) >>>> >>>> > > >The interpreter shows a message: "The C++ part of the code has been >deleted" and i can?t plot again without exit the interpreter and back >again. > > >How i can solve that? > >---------------------------------------------------------------------------------------------------------------- > >gplt (gnuplot): > >If i create a plot with the Gnuplot program the resulting plot has >responses to mouse events and shows coordinates in the window but if i >create a plot in python > >gplt.plot((1,2,3)) > >The window have not the same response to mouse events (no response at all). > >How i can give the same responses? >___________________________________________________________________ > > > > >Thanks > >_______________________________________________ >SciPy-user mailing list >SciPy-user at scipy.net >http://www.scipy.net/mailman/listinfo/scipy-user > > > > From rubencharles at gmail.com Tue Sep 20 10:27:07 2005 From: rubencharles at gmail.com (Ruben Charles) Date: Tue, 20 Sep 2005 10:27:07 -0400 Subject: [SciPy-user] plotting questions (gplt and plt) In-Reply-To: <432FC859.4060000@city.ac.uk> References: <432FC859.4060000@city.ac.uk> Message-ID: Thanks, i'm going to try... But the plt "problem" hasn't been resolved yet? On 9/20/05, Brice Thurin wrote: > Hi, > > When I was running into problem with plt, xplt and gnuplot, someone > suggest to me to use matplotlib, which I did... and I am quite satisfied. > > > > Ruben Charles wrote: > > >Hi, > > > >I'm using Scipy on Debian Gnu/Linux. (python2.3-scipy) in a program > >that have to show plots but while i'm testing capabilities of the > >three possible options (gplt, plt an xplt) a couple of questions comes > >up. > > > >-------------------------------------------------------------------------------------------------------------- > >plt: > > > >If a follow the steps of the plotting tutorial everything seems ok in > >the first try. > > > > > > > >>>>import gui_thread > >>>>gui_thread.start() # or gui_thread.wxPython_thread() > >>>>from scipy import plt > >>>>plt.plot((1,2,3)) > >>>> > >>>> > > > > > >But if i close the Plot frame (manually with the cursor) and try to plot again > > > > > > > >>>>plt.plot((1,2,3)) > >>>> > >>>> > > > > > >The interpreter shows a message: "The C++ part of the code has been > >deleted" and i can?t plot again without exit the interpreter and back > >again. > > > > > >How i can solve that? > > > >---------------------------------------------------------------------------------------------------------------- > > > >gplt (gnuplot): > > > >If i create a plot with the Gnuplot program the resulting plot has > >responses to mouse events and shows coordinates in the window but if i > >create a plot in python > > > >gplt.plot((1,2,3)) > > > >The window have not the same response to mouse events (no response at all). > > > >How i can give the same responses? > >___________________________________________________________________ > > > > > > > > > >Thanks > > > >_______________________________________________ > >SciPy-user mailing list > >SciPy-user at scipy.net > >http://www.scipy.net/mailman/listinfo/scipy-user > > > > > > > > > > From rkern at ucsd.edu Tue Sep 20 10:58:00 2005 From: rkern at ucsd.edu (Robert Kern) Date: Tue, 20 Sep 2005 07:58:00 -0700 Subject: [SciPy-user] plotting questions (gplt and plt) In-Reply-To: References: <432FC859.4060000@city.ac.uk> Message-ID: <43302378.2030906@ucsd.edu> Ruben Charles wrote: > Thanks, i'm going to try... > > But the plt "problem" hasn't been resolved yet? Not many people use plt, xplt or gplt. They are essentially deprecated, and new users are being directed to matplotlib. xplt occasionally gets some TLC because Travis Oliphant uses it, but other than that, they've pretty much been abandoned. I doubt that they are going to make it through the coming reorganization. -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From rubencharles at gmail.com Tue Sep 20 11:15:54 2005 From: rubencharles at gmail.com (Ruben Charles) Date: Tue, 20 Sep 2005 11:15:54 -0400 Subject: [SciPy-user] plotting questions (gplt and plt) In-Reply-To: <43302378.2030906@ucsd.edu> References: <432FC859.4060000@city.ac.uk> <43302378.2030906@ucsd.edu> Message-ID: Ok, thanks for the sugestion, I asked for gplt, and plt because they are in scipy. The plotting tutorial show only xplt, gplt, and plt. Matplot is the best choice for 2d-plots? Ok... Then it seems to me that the documentation in scypy.org is out of date. On 9/20/05, Robert Kern wrote: > Ruben Charles wrote: > > Thanks, i'm going to try... > > > > But the plt "problem" hasn't been resolved yet? > > Not many people use plt, xplt or gplt. They are essentially deprecated, > and new users are being directed to matplotlib. xplt occasionally gets > some TLC because Travis Oliphant uses it, but other than that, they've > pretty much been abandoned. I doubt that they are going to make it > through the coming reorganization. > > -- > Robert Kern > rkern at ucsd.edu > > "In the fields of hell where the grass grows high > Are the graves of dreams allowed to die." > -- Richard Harter > From rkern at ucsd.edu Tue Sep 20 11:44:38 2005 From: rkern at ucsd.edu (Robert Kern) Date: Tue, 20 Sep 2005 08:44:38 -0700 Subject: [SciPy-user] plotting questions (gplt and plt) In-Reply-To: References: <432FC859.4060000@city.ac.uk> <43302378.2030906@ucsd.edu> Message-ID: <43302E66.8040405@ucsd.edu> Ruben Charles wrote: > Ok, thanks for the sugestion, > > I asked for gplt, and plt because they are in scipy. > The plotting tutorial show only xplt, gplt, and plt. > > Matplot is the best choice for 2d-plots? Ok... > > Then it seems to me that the documentation in scypy.org is out of date. Yes. Quite. I'm working on it. -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From aisaac at american.edu Tue Sep 20 12:04:13 2005 From: aisaac at american.edu (Alan G Isaac) Date: Tue, 20 Sep 2005 12:04:13 -0400 Subject: [SciPy-user] plotting questions (gplt and plt) In-Reply-To: <43302E66.8040405@ucsd.edu> References: <432FC859.4060000@city.ac.uk> <43302378.2030906@ucsd.edu> <43302E66.8040405@ucsd.edu> Message-ID: > Ruben Charles wrote: >> The plotting tutorial show only xplt, gplt, and plt. >> Then it seems to me that the documentation in scypy.org >> is out of date. On Tue, 20 Sep 2005, Robert Kern apparently wrote: > Yes. Quite. I'm working on it. Here are the first three tutorials, redone for Matplotlib: http://www.scipy.org/documentation/mailman?fn=scipy-dev/2005-July/003049.html http://www.scipy.org/documentation/mailman?fn=scipy-dev/2005-July/003050.html http://www.scipy.org/documentation/mailman?fn=scipy-dev/2005-July/003051.html hth, Alan Isaac From sylvain.gerbier at newlogic.fr Wed Sep 21 05:04:25 2005 From: sylvain.gerbier at newlogic.fr (Sylvain Gerbier) Date: Wed, 21 Sep 2005 11:04:25 +0200 Subject: [SciPy-user] Scipy with Python 2.4 on Windows Message-ID: <43312219.5010805@newlogic.fr> Hi everybody, Need help here ... does anybody install scipy on a windows computer with Python 2.4 ? If yes, can he/she tells me how to do it ? I am not so much familiar with the build procedure but can learn ... If nobody have done it, I actually only use the io.savemat function from the scipy library. Does anybody know how I can find a turnaround to save a file with the matlab format without using the scipy library ? Thanks for your help. Sylvain From ryanfedora at comcast.net Wed Sep 21 08:30:39 2005 From: ryanfedora at comcast.net (Ryan Krauss) Date: Wed, 21 Sep 2005 08:30:39 -0400 Subject: [SciPy-user] Scipy with Python 2.4 on Windows In-Reply-To: <43312219.5010805@newlogic.fr> References: <43312219.5010805@newlogic.fr> Message-ID: <4331526F.6050101@comcast.net> Whenever the question of 2.4 and Windows has come up, the answer has been that no one is doing it yet. Ryan Sylvain Gerbier wrote: > Hi everybody, > > Need help here ... does anybody install scipy on a windows computer with > Python 2.4 ? If yes, can he/she tells me how to do it ? I am not so much > familiar with the build procedure but can learn ... > If nobody have done it, I actually only use the io.savemat function from > the scipy library. Does anybody know how I can find a turnaround to save > a file with the matlab format without using the scipy library ? > Thanks for your help. > > Sylvain > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user > From d.howey at imperial.ac.uk Wed Sep 21 08:55:54 2005 From: d.howey at imperial.ac.uk (Howey, David A) Date: Wed, 21 Sep 2005 13:55:54 +0100 Subject: [SciPy-user] Scipy with Python 2.4 on Windows Message-ID: <056D32E9B2D93B49B01256A88B3EB21835E78B@icex2.ic.ac.uk> I just used python 2.3 it works fine on windows. also check out enthought python at http://www.enthought.com/python/ Dave Howey -----Original Message----- From: scipy-user-bounces at scipy.net on behalf of Ryan Krauss Sent: Wed 21/09/2005 13:30 To: SciPy Users List Subject: Re: [SciPy-user] Scipy with Python 2.4 on Windows Whenever the question of 2.4 and Windows has come up, the answer has been that no one is doing it yet. Ryan Sylvain Gerbier wrote: > Hi everybody, > > Need help here ... does anybody install scipy on a windows computer with > Python 2.4 ? If yes, can he/she tells me how to do it ? I am not so much > familiar with the build procedure but can learn ... > If nobody have done it, I actually only use the io.savemat function from > the scipy library. Does anybody know how I can find a turnaround to save > a file with the matlab format without using the scipy library ? > Thanks for your help. > > Sylvain > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user > _______________________________________________ SciPy-user mailing list SciPy-user at scipy.net http://www.scipy.net/mailman/listinfo/scipy-user -------------- next part -------------- A non-text attachment was scrubbed... Name: winmail.dat Type: application/ms-tnef Size: 3034 bytes Desc: not available URL: From jdhunter at ace.bsd.uchicago.edu Thu Sep 22 02:58:06 2005 From: jdhunter at ace.bsd.uchicago.edu (John Hunter) Date: Thu, 22 Sep 2005 01:58:06 -0500 Subject: [SciPy-user] os x vga dongle at scipy conference tomorrow Message-ID: <87psr1bncx.fsf@peds-pc311.bsd.uchicago.edu> If anyone is coming to the SciPy conference tomorrow morning and has a 12 inch powerbook dvi->VGA adaptor for the overhead could you bring it along for the morning session? Apple in their infinite wisdom has two sizes on the dvi side depending on the 12 inch model. I have the fat one. Or I should say, I need the fat one. Thanks, JDH From arnd.baecker at web.de Fri Sep 23 03:34:13 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Fri, 23 Sep 2005 09:34:13 +0200 (CEST) Subject: [SciPy-user] Installation on dual opteron ... Message-ID: Hi, we are trying to install scipy on a dual opteron machine (presently running SUSE Linux 10.0 OSS beta). Because there is no fortran compiler installed (and it seems there is none for that beta?), we started to compile the relevant bits from scratch. There are several problems, so before we start reporting them in detail ;-), I would like to ask some general questions: - I have read the thread started by Stephen Walton http://aspn.activestate.com/ASPN/Mail/Message/scipy-dev/2817873 and would like to know, if the patches by Ravikiran Rajagopa (including the c versions of d1mach.f, r1mach.f and i1mach.f, posted in http://aspn.activestate.com/ASPN/Mail/Message/scipy-dev/2820689) have been applied to svn? - Would you recommend to use - gcc, Branch status 2005-09-13 (frozen for 4.0.2 release) or - or GCC 3.4.4 (released 2005-05-18) instead of gcc 4.0.1? (Would choosing gcc 3.4.4 have any negative impact on 64Bit performance?) - And more generally: is there any progress concerning Arrays >2GB with python/Numeric or numarray (as mentioned by Perry Greenfield http://www.scipy.net/pipermail/scipy-user/2005-February/004079.html The machine has 2x16 GB RAM and it would be nice to be able to use them... Many thanks in advance, Arnd and Jan Short notes on the installation: - gcc 4.0.1 (via http://gcc.gnu.org/) Compile command: ../gcc-4.0.1/configure --prefix=/scr/python --enable-shared \ --enable-threads=posix --enable-__cxa_atexit --enable-clocale=gnu \ --enable-languages=c,c++,f95,objc - export CFLAGS=-fPIC - Python 2.4.1 - ATLAS: atlas3.7.11.tar.bz2 (Linux_HAMMER64SSE3_2) added: full Lapack version (following http://www.scipy.org/documentation/buildatlas4scipy.txt) - BLAS: gfortran -fno-second-underscore -O2 -fPIC -c *.f - LAPACK: make.inc.LINUX, In OPTS variable replace '-fno-f2c -O3' with '-O2 -fPIC' - ATLAS: make xconfig ./xconfig -F f "-fomit-frame-pointer -O -fno-second-underscore -fPIC" -b $BLAS make config - 64 Bit - Posix threads - don't stop - use express setup? [y]: - Enter f77 compiler [g77]: gfortran Enter F77 Flags [-fomit-frame-pointer -O -m64]: - Tune the Level 1 BLAS? [y]: One Problem is that the tests make sanity_test arch=Linux_HAMMER64SSE3_2 make ptsanity_test arch=Linux_HAMMER64SSE3_2 fail - ipython - fftw - f2py2e - scipy: svn co http://svn.scipy.org/svn/scipy/trunk scipy # Checked out revision 1270. cd scipy svn co http://svn.scipy.org/svn/scipy_core/trunk/ scipy_core # Install scipy_distutils: cd scipy_core/scipy_distutils python setup.py install # get system info: python scipy_core/scipy_distutils/system_info.py > ../scipy_systeminfo.txt F77=gfortran python setup.py install | tee ../scipy_install_log.txt From stephen.walton at csun.edu Sat Sep 24 01:20:24 2005 From: stephen.walton at csun.edu (Stephen Walton) Date: Fri, 23 Sep 2005 22:20:24 -0700 Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: References: Message-ID: <4334E218.9020905@csun.edu> Hello, > and would like to know, if the patches > by Ravikiran Rajagopa > (including the c versions of d1mach.f, r1mach.f and i1mach.f, > posted in > http://aspn.activestate.com/ASPN/Mail/Message/scipy-dev/2820689) > have been applied to svn? > > > As far as I can tell, the answer is no. As I did mention, I think the C versions of ?1mach need a lot more testing. There is also a problem I just found with Ravi's patch: gfortran is used to compile and link absoft_version.so but against -lg2c rather than against -lgfortran. >- Would you recommend to use > - gcc, Branch status 2005-09-13 (frozen for 4.0.2 release) > or > - or GCC 3.4.4 (released 2005-05-18) > instead of gcc 4.0.1? > > I would strongly recommend 3.4.4 and g77. Since gfortran is part of the gcc4 suite, I'm not sure what will happen if you use gcc 3.x but gfortran instead of g77. > (Would choosing gcc 3.4.4 have any negative impact on 64Bit > performance?) > > My tests show that gfortran/gcc4 do in fact have some performance advantages over gcc3/g77, but I've only looked at the i386 platform. > F77=gfortran python setup.py install | tee ../scipy_install_log.txt > > I think this should be python setup.py config_fc --fcompiler=gfortran install if you're using Ravi's patches. From oliphant at ee.byu.edu Sat Sep 24 23:42:22 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Sat, 24 Sep 2005 21:42:22 -0600 Subject: [SciPy-user] Scipy with Python 2.4 on Windows In-Reply-To: <4331526F.6050101@comcast.net> References: <43312219.5010805@newlogic.fr> <4331526F.6050101@comcast.net> Message-ID: <43361C9E.5040902@ee.byu.edu> Ryan Krauss wrote: > Whenever the question of 2.4 and Windows has come up, the answer has > been that no one is doing it yet. I'm pretty sure the problem with scipy and windows on Python 2.4 is with the build of numpyio not picking up the same input / output DLL that Python was compiled against. I'm not sure why this would happen suddenly on the change to Python 2.4. It is probably a distutils-related issues. In the mean time, the rest of scipy should be fine you just won't be able to use a lot of the io functions. If anybody has any ideas, please help. Otherwise, this may take a bit to track down as I don't have time to spend on it, and nobody else has volunteered. -Travis From oliphant at ee.byu.edu Sun Sep 25 00:51:27 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Sat, 24 Sep 2005 22:51:27 -0600 Subject: [SciPy-user] Release of scipy core beta will happen next week. Message-ID: <43362CCF.5020100@ee.byu.edu> At the SciPy 2005 conference I announced that I was hoping to get a beta of the new scipy (core) (aka Numeric3 aka Numeric Next Generation) released by the end of the conference. This did not happen. Some last minute features were suggested by Fernando Perez that I think will be relatively easy to add and make the release that much stronger. Look for the beta announcement next week. For the impatient, the svn server is always available: http://svn.scipy.org/svn/scipy_core/branches/newcore -Travis O. From simon.j.hook at jpl.nasa.gov Sun Sep 25 13:45:50 2005 From: simon.j.hook at jpl.nasa.gov (Simon Hook) Date: Sun, 25 Sep 2005 10:45:50 -0700 Subject: [SciPy-user] SciPy 2005 - thanks, accolades and suggestions Message-ID: <4336E24E.7060405@jpl.nasa.gov> Many thanks to Enthought and all the speakers/developers for putting on an excellent SciPy 2005. As a someone who is fairly new to both Python and SciPy, I found this a tremendous opportunity to learn more about what can be done with Python for scientific computing (just about everything!). At the meeting there was a request for suggestions on how increase the usability and audience. Here are more thoughts: 1. The Enthought distribution of Python (Enthon) which includes many of the scientific packages, tested to make sure they work together is extremely valuable to someone like me who does not know what tools are out there and does not want to do dozens of downloads and installs only to find that package "a" conflicts with "b". I realize that such a distribution is a significant contribution and I would be happy to pay a modest sum for such a package. If someone puts together such a package please make the payment easy - i..e. credit card over the net. If you work in/with government there are certain methods that are straighforward to pay for stuff and others that require much more paperwork, e.g. check, paypal, ebay. The best method is credit card over the net. Given the rapid rate of improvement, another approach might be to buy the first distribution, then get free new distributions for a year and then buy the next distribution. Note setting up service contracts is also a hassle so better to have a new distribution to buy once per year or so. 2. I would like to suggest the following packages be added to the Enthon distribution: mxtime (www.egenix.com) ipython (ipython.scipy.org) matplotlib (matplotlib.sourceforge.net) inotebook (shown at scipy - not sure where to download) vision (www.scripps.edu/~sanner/python/viper) 3. Many of the packages have excellent tutorials but they are package centric. What I mean by this is instead of taking a scientific problem and showing how you solve it, each package carefully describes its capabilities largely in isolation. It would be great if there was a simple tutorial(s) that tried to show how the tools could be used together. For example, say you want to read in some air temperature data, produce a plot of the data, create and fit a model to the data, generate an inotebook and then create a vision type model to interact with the data - you would have to know about all the packages and have read all the tutorials for each of the packages - too much time would be spent on the tools and not enough on the problem. Incidentally there is an example of something headed along these lines in the scipy tutorial. I am happy to help try and put something like this together, perhaps it could be a tutorial, perhaps a book - it would benefit from help from each of the people developing the different packages. On that note it would be nice if the package tutorials were on the scipy site under docs and if package developers called their guides "package-name_users_guide" rather than just "users_guide". 4. More examples. I find it much easier to follow examples than read manuals. It would be great if there were more user contributed examples in scipy. Matplotlib does an excellent job with examples. I just sent an example in on using read_array for scipy. 5. If there were some tutorials that showed how the tools could be used to solve scientific problems it may be worthwhile trying to have a morning at scipy for new users that worked them through the tutorials. Alternatively these could be at science conferences. In my research area, before a conference starts there are usually classes which people pay for that teach them new tools. The classes tend to be geared towards a particular research area, in my case geologic remote sensing. I have run a few classes and it would be good way to introduce students to scipy+ in the context of what they need to do. Again many thanks again to the organizers, developers and presenters. Simon -- +++++++++++++++++++++++++++++++++++++++++ Simon J. Hook, MSc, PhD Jet Propulsion Laboratory MS 183-501 Pasadena, CA 91109 Office: 818-354-0974 Fax: 818-354-0966 Email: simon.j.hook at jpl.nasa.gov http://asterweb.jpl.nasa.gov http://masterweb.jpl.nasa.gov http://laketahoe.jpl.nasa.gov +++++++++++++++++++++++++++++++++++++++++ From rkern at ucsd.edu Mon Sep 26 02:55:11 2005 From: rkern at ucsd.edu (Robert Kern) Date: Sun, 25 Sep 2005 23:55:11 -0700 Subject: [SciPy-user] Scipy with Python 2.4 on Windows In-Reply-To: <43361C9E.5040902@ee.byu.edu> References: <43312219.5010805@newlogic.fr> <4331526F.6050101@comcast.net> <43361C9E.5040902@ee.byu.edu> Message-ID: <43379B4F.9030102@ucsd.edu> Travis Oliphant wrote: > Ryan Krauss wrote: > >> Whenever the question of 2.4 and Windows has come up, the answer has >> been that no one is doing it yet. > > I'm pretty sure the problem with scipy and windows on Python 2.4 is with > the build of numpyio not picking up the same input / output DLL that > Python was compiled against. I'm not sure why this would happen > suddenly on the change to Python 2.4. It is probably a > distutils-related issues. Well, the C runtime changed completely with Python 2.4 which is now being compiled with MSVC 7. > In the mean time, the rest of scipy should be fine you just won't be > able to use a lot of the io functions. > > If anybody has any ideas, please help. Otherwise, this may take a bit > to track down as I don't have time to spend on it, and nobody else has > volunteered. As I mentioned the last time this came up (three weeks ago!), John Hunter say that he has had success compiling matplotlib with mingw by making a tiny change to distutils: http://mail.python.org/pipermail/python-list/2004-December/254826.html I'm not sure if matplotlib is passing FILE* pointers between Python and the extension modules, though. It may not be exercising the crash-inducing conditions. -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From sylvain.gerbier at newlogic.fr Mon Sep 26 03:10:55 2005 From: sylvain.gerbier at newlogic.fr (Sylvain Gerbier) Date: Mon, 26 Sep 2005 09:10:55 +0200 Subject: [SciPy-user] Scipy with Python 2.4 on Windows In-Reply-To: <43379B4F.9030102@ucsd.edu> References: <43312219.5010805@newlogic.fr> <4331526F.6050101@comcast.net> <43361C9E.5040902@ee.byu.edu> <43379B4F.9030102@ucsd.edu> Message-ID: <43379EFF.9010708@newlogic.fr> Thanks for all your answers .. I'm not familiar enough with this compilation and build things to look into it .. I'm going to wait and try to find another solution .. Thanks anyway .. Sylvain Robert Kern wrote: > Travis Oliphant wrote: > >>Ryan Krauss wrote: >> >> >>>Whenever the question of 2.4 and Windows has come up, the answer has >>>been that no one is doing it yet. >> >>I'm pretty sure the problem with scipy and windows on Python 2.4 is with >>the build of numpyio not picking up the same input / output DLL that >>Python was compiled against. I'm not sure why this would happen >>suddenly on the change to Python 2.4. It is probably a >>distutils-related issues. > > > Well, the C runtime changed completely with Python 2.4 which is now > being compiled with MSVC 7. > > >>In the mean time, the rest of scipy should be fine you just won't be >>able to use a lot of the io functions. >> >>If anybody has any ideas, please help. Otherwise, this may take a bit >>to track down as I don't have time to spend on it, and nobody else has >>volunteered. > > > As I mentioned the last time this came up (three weeks ago!), John > Hunter say that he has had success compiling matplotlib with mingw by > making a tiny change to distutils: > > http://mail.python.org/pipermail/python-list/2004-December/254826.html > > I'm not sure if matplotlib is passing FILE* pointers between Python and > the extension modules, though. It may not be exercising the > crash-inducing conditions. > From rainman.rocks at gmail.com Mon Sep 26 14:26:30 2005 From: rainman.rocks at gmail.com (rainman) Date: Mon, 26 Sep 2005 22:26:30 +0400 Subject: [SciPy-user] Scipy with Python 2.4 on Windows - solution Message-ID: <1527564003.20050926222630@gmail.com> Hello scipy-user, To succesfully build SciPy with Python 2.4, replace file \SciPy_complete-0.3.2\scipy_core\scipy_distutils\mingw32_support.py with http://home.uic.tula.ru/~s001180/scipy-py24/mingw32ccompiler.py Those who do not believe their spirit is bold enough to face tremendous nightmares of compiling tons of sources, may take a chance with this (beware, downloading is slow): http://home.uic.tula.ru/~s001180/scipy-py24/SciPy_complete-0.3.2.win32-py2.4-atlas3.6.0_WinNT_PIII.exe Note that this is UNofficial bugfix and UNofficial build. Just a quick hack for those who lost all hope. -- Best regards, rainman mailto:rainman.rocks at gmail.com From aisaac at american.edu Mon Sep 26 14:43:36 2005 From: aisaac at american.edu (Alan G Isaac) Date: Mon, 26 Sep 2005 14:43:36 -0400 Subject: [SciPy-user] Scipy with Python 2.4 on Windows - solution In-Reply-To: <1527564003.20050926222630@gmail.com> References: <1527564003.20050926222630@gmail.com> Message-ID: On Mon, 26 Sep 2005, rainman apparently wrote: > http://home.uic.tula.ru/~s001180/scipy-py24/SciPy_complete-0.3.2.win32-py2.4-atlas3.6.0_WinNT_PIII.exe > Note that this is UNofficial bugfix and UNofficial build. Just a > quick hack for those who lost all hope. Does this use the new scipy.core? Thanks, Alan Isaac From oliphant at ee.byu.edu Mon Sep 26 15:57:21 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Mon, 26 Sep 2005 13:57:21 -0600 Subject: [SciPy-user] Scipy with Python 2.4 on Windows - solution In-Reply-To: References: <1527564003.20050926222630@gmail.com> Message-ID: <433852A1.4030802@ee.byu.edu> Alan G Isaac wrote: >On Mon, 26 Sep 2005, rainman apparently wrote: > > >> http://home.uic.tula.ru/~s001180/scipy-py24/SciPy_complete-0.3.2.win32-py2.4-atlas3.6.0_WinNT_PIII.exe >> >> > > > >> Note that this is UNofficial bugfix and UNofficial build. Just a >> quick hack for those who lost all hope. >> >> > >Does this use the new scipy.core? > > No, I'm pretty sure it does not. -Travis From oliphant at ee.byu.edu Mon Sep 26 16:07:18 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Mon, 26 Sep 2005 14:07:18 -0600 Subject: [SciPy-user] Scipy with Python 2.4 on Windows - solution In-Reply-To: <1527564003.20050926222630@gmail.com> References: <1527564003.20050926222630@gmail.com> Message-ID: <433854F6.5050907@ee.byu.edu> rainman wrote: >Hello scipy-user, > > To succesfully build SciPy with Python 2.4, replace file > \SciPy_complete-0.3.2\scipy_core\scipy_distutils\mingw32_support.py > with > http://home.uic.tula.ru/~s001180/scipy-py24/mingw32ccompiler.py > > I've incorporated the changes rainman specified into the svn version of mingw32ccompiler.py Hopefully, this means that the compile will work for anybody with the svn version of scipy. Testers, welcomed. -Travis O. From rainman.rocks at gmail.com Tue Sep 27 00:36:29 2005 From: rainman.rocks at gmail.com (rainman) Date: Tue, 27 Sep 2005 08:36:29 +0400 Subject: [SciPy-user] Re: Scipy with Python 2.4 on Windows - solution In-Reply-To: <1527564003.20050926222630@gmail.com> References: <1527564003.20050926222630@gmail.com> Message-ID: <837163390.20050927083629@gmail.com> Hello rainman, > To succesfully build SciPy with Python 2.4, replace file > \SciPy_complete-0.3.2\scipy_core\scipy_distutils\mingw32_support.py Sorry! of course, \SciPy_complete-0.3.2\scipy_core\scipy_distutils\mingw32ccompiler.py > with http://home.uic.tula.ru/~s001180/scipy-py24/mingw32ccompiler.py -- Best regards, rainman mailto:rainman.rocks at gmail.com From sylvain.gerbier at newlogic.fr Tue Sep 27 02:23:10 2005 From: sylvain.gerbier at newlogic.fr (Sylvain Gerbier) Date: Tue, 27 Sep 2005 08:23:10 +0200 Subject: [SciPy-user] Scipy with Python 2.4 on Windows - solution In-Reply-To: <1527564003.20050926222630@gmail.com> References: <1527564003.20050926222630@gmail.com> Message-ID: <4338E54E.9010503@newlogic.fr> This is a great post and a good news this morning (in France) !! I'm going to test that and will come back to you for the results .. Thanks a lot !! Sylvain rainman wrote: > Hello scipy-user, > > To succesfully build SciPy with Python 2.4, replace file > \SciPy_complete-0.3.2\scipy_core\scipy_distutils\mingw32_support.py > with > http://home.uic.tula.ru/~s001180/scipy-py24/mingw32ccompiler.py > > Those who do not believe their spirit is bold enough to face > tremendous nightmares of compiling tons of sources, may take a > chance with this (beware, downloading is slow): > > http://home.uic.tula.ru/~s001180/scipy-py24/SciPy_complete-0.3.2.win32-py2.4-atlas3.6.0_WinNT_PIII.exe > > Note that this is UNofficial bugfix and UNofficial build. Just a > quick hack for those who lost all hope. > > From arnd.baecker at web.de Tue Sep 27 03:06:04 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Tue, 27 Sep 2005 09:06:04 +0200 (CEST) Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: <4334E218.9020905@csun.edu> References: <4334E218.9020905@csun.edu> Message-ID: Hi, many thanks for your reply! On Fri, 23 Sep 2005, Stephen Walton wrote: [...] > I would strongly recommend 3.4.4 and g77. Since gfortran is part of the > gcc4 suite, I'm not sure what will happen if you use gcc 3.x but > gfortran instead of g77. So I went for 3.4.4 and get stuck with the -fPIC stuff. Presumably I am missing something simple - any help is appreciated (list of questions below ...). `python setup.py install` for scipy leads to: [...] building 'scipy.lib.lapack.flapack' extension compiling C sources gcc options: '-pthread -fno-strict-aliasing -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -fPIC -fPIC' compile options: '-DATLAS_INFO="\"3.7.11\"" -DNUMERIC_VERSION="\"24.0b2\"" -DNUMERIC -I/scr/python3/lib -I/scr/python3/include/python2.4 -Ibuild/src -I/scr/python3/include/python2.4 -c' /scr/python3/bin/g77 -shared build/temp.linux-x86_64-2.4/build/src/build/src/Lib/lib/lapack/flapackmodule.o build/temp.linux-x86_64-2.4/build/src/fortranobject.o -L/scr/python3/lib64 -Lbuild/temp.linux-x86_64-2.4 -llapack -lptf77blas -lptcblas -latlas -lg2c -o build/lib.linux-x86_64-2.4/scipy/lib/lapack/flapack.so /usr/bin/ld: /scr/python3/lib64/liblapack.a(slamch.o): relocation R_X86_64_32S against `a local symbol' can not be used when making a shared object; recompile with -fPIC /scr/python3/lib64/liblapack.a: could not read symbols: Bad value collect2: ld returned 1 exit status error: Command "/scr/python3/bin/g77 -shared build/temp.linux-x86_64-2.4/build/src/build/src/Lib/lib/lapack/flapackmodule.o build/temp.linux-x86_64-2.4/build/src/fortranobject.o -L/scr/python3/lib64 -Lbuild/temp.linux-x86_64-2.4 -llapack -lptf77blas -lptcblas -latlas -lg2c -o build/lib.linux-x86_64-2.4/scipy/lib/lapack/flapack.so" failed with exit status 1/usr/bin/ld: /scr/python3/lib64/liblapack.a(slamch.o): relocation R_X86_64_32S against `a local symbol' can not be used when making a shared object; recompile with -fPIC /scr/python3/lib64/liblapack.a: could not read symbols: Bad value collect2: ld returned 1 exit status - I compiled atlas etc. with -fPIC (at least I tried to enforce that) - Is there a way to check the resulting libraries if they were compiled properly with -fPIC? - It *seems* that Numeric works with the ATLAS I compiled - Is there a way to check (after installation) against which libraries Numeric and scipy are linked (does it use ATLAS, does it use dotblas?). - Below I attach my notes of the steps of the installation. Maybe some expert can spot what I am doing wrong... Many thanks in advance, Arnd ============= Install notes ============= Directory ``Sources`` ===================== gcc-3.4.4.tar.bz2 Python-2.4.1.tar.bz2 ipython.zip (from CVS) atlas3.7.11.tar.bz2 Numerical.zip (from CVS) f2py2e.zip (from CVS) scipy.zip (from svn, includes scipy_core) Environment variables ====================== export PHOME=/scr/python3/ export INFODIR=${PHOME}/info:$INFODIR export INFOPATH=${PHOME}/info:$INFOPATH export MANPATH=${PHOME}/man:$MANPATH export LD_LIBRARY_PATH=${PHOME}/lib64:${PHOME}/lib:$LD_LIBRARY_PATH export LD_RUN_PATH=${PHOME}/lib64:${PHOME}/lib:$LD_RUN_PATH export PYTHONDOCS=${PHOME}/docs/Python-Docs-2.4.1/ export PATH=${PHOME}/bin:${PATH} export PAGER="less -R" export CFLAGS=-fPIC For the installation:: mkdir CompileDir cd CompileDir gcc === tar xjf ../Sources/gcc-3.4.4.tar.bz2 mkdir gcc-build cd gcc-build ../gcc-3.4.4/configure --prefix=${PHOME} --enable-shared --enable-threads=posix --enable-__cxa_atexit --enable-clocale=gnu --enable-languages=c,c++,f77,objc make make install cd /scr/python3/bin/ ln -s gcc /scr/python3/bin/cc cd - Python ====== tar xjf ../Sources/Python-2.4.1.tar.bz2 cd Python-2.4.1 ./configure --prefix=${PHOME} make make install make test 255 tests OK. 35 tests skipped: test_aepack test_al test_applesingle test_bsddb185 test_bsddb3 test_cd test_cl test_codecmaps_cn test_codecmaps_hk test_codecmaps_jp test_codecmaps_kr test_codecmaps_tw test_curses test_dl test_gl test_imageop test_imgfile test_linuxaudiodev test_macfs test_macostools test_nis test_normalization test_ossaudiodev test_pep277 test_plistlib test_rgbimg test_scriptpackages test_socket_ssl test_socketserver test_sunaudiodev test_timeout test_urllib2net test_urllibnet test_winreg test_winsound Those skips are all expected on linux2. cd .. IPython ======= unzip ../Sources/ipython.zip cd ipython python setup.py install cd .. ATLAS and BLAS/LAPACK ===================== See: http://www.scipy.org/documentation/buildatlas4scipy.txt export BLAS_SRC=${HOME}/INSTALL_PYTHON/CompileDir/BLAS export LAPACK_SRC=${HOME}/INSTALL_PYTHON/CompileDir/LAPACK mkdir -p $BLAS_SRC cd $BLAS_SRC wget http://www.netlib.org/blas/blas.tgz tar xzf blas.tgz g77 -fno-second-underscore -O2 -fPIC -c *.f ar r libfblas.a *.o ranlib libfblas.a rm -rf *.o mkdir -p $LAPACK_SRC cd $LAPACK_SRC/.. wget http://www.netlib.org/lapack/lapack.tgz tar xzf lapack.tgz cd $LAPACK_SRC cp INSTALL/make.inc.LINUX make.inc # on LINUX # In OPTS variable replace '-fno-f2c -O3' with '-O2 -fPIC' make lapacklib make clean cd .. ---------- export BLAS=$BLAS_SRC/libfblas.a tar xjf ../Sources/atlas3.7.11.tar.bz2 cd ATLAS make xconfig ./xconfig -F f "-fomit-frame-pointer -O -fno-second-underscore -fPIC" \ -F c " -fomit-frame-pointer -O -mfpmath=387 -m64 -fPIC" \ -F m "-fomit-frame-pointer -O -mfpmath=387 -m64 -fPIC" -b $BLAS - 64 Bit - Posix threads - don't stop - use express setup? [y]: - Enter f77 compiler [g77]: gfortran Enter F77 Flags [-fomit-frame-pointer -O -m64]: - Tune the Level 1 BLAS? [y]: # Gives for the compile commands: F77 = /scr/python3//bin/g77 -fomit-frame-pointer -O -fno-second-underscore -fPIC CC = /scr/python3//bin/gcc -fomit-frame-pointer -O -mfpmath=387 -m64 -fPIC MCC = /scr/python3//bin/gcc -fomit-frame-pointer -O -mfpmath=387 -m64 -fPIC make sanity_test arch=Linux_HAMMER64SSE3_2 make ptsanity_test arch=Linux_HAMMER64SSE3_2 # both work fine * Make optimized LAPACK library:: cd lib//Linux_HAMMER64SSE3_2/ mkdir tmp; cd tmp ar x ../liblapack.a cp ../liblapack.a ../liblapack.a_SAVE cp ../../../../LAPACK/lapack_LINUX.a ../liblapack.a ar r ../liblapack.a *.o cd ..; rm -rf tmp ls -lh liblapack.a -rw-r--r-- 1 arnd users 10M 2005-09-26 14:11 liblapack.a cp *.a $PHOME/lib64 cd ../../ cp include/{cblas.h,clapack.h} $PHOME/include Numerical ========= unzip ../Sources/Numerical.zip cd Numerical emacs customize.py if 1: use_system_lapack = 1 lapack_library_dirs = ['/scr/python3//lib64/'] lapack_libraries = ['lapack', 'cblas', 'f77blas', 'atlas', 'g2c'] if 1: use_dotblas = 1 dotblas_include_dirs = ['/scr/python3/include/atlas'] dotblas_library_dirs = lapack_library_dirs dotblas_libraries = ['cblas', 'atlas', 'g2c'] python setup.py build | tee ../Numeric_install_log.txt python setup.py install | tee ../Numeric_install_log.txt cd .. scipy ===== # Fortran to Python unzip ../Sources/f2py2e.zip cd f2py2e make install cd .. # scipy: unzip ../Sources/scipy.zip cd scipy export ATLAS=/scr/python3/lib64:/scr/python3/include cd scipy_core/scipy_distutils python setup.py install cd ../.. # scipy_systeminfo.txt for report on installed packages python scipy_core/scipy_distutils/system_info.py > ../scipy_systeminfo.txt python setup.py install | tee ../scipy_install_log.txt I also tried: python setup.py config_fc --opt "-fPIC" build | tee ../scipy_install_log.txt but no difference ... From bgoli at sun.ac.za Tue Sep 27 03:54:54 2005 From: bgoli at sun.ac.za (Brett Olivier) Date: Tue, 27 Sep 2005 09:54:54 +0200 Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: References: <4334E218.9020905@csun.edu> Message-ID: <200509270954.54610.bgoli@sun.ac.za> Hi Try the options suggested in http://www.scipy.org/mailinglists/mailman?fn=scipy-user/2005-February/004066.html I used them to compile ATLAS/LAPACK for SciPy on an Intel processor running Mandriva x86_64 after I ran into -fPIC problems. Brett On Tuesday 27 September 2005 09:06, Arnd Baecker wrote: > Hi, > > many thanks for your reply! > > On Fri, 23 Sep 2005, Stephen Walton wrote: > > [...] > > > I would strongly recommend 3.4.4 and g77. Since gfortran is part of the > > gcc4 suite, I'm not sure what will happen if you use gcc 3.x but > > gfortran instead of g77. > > So I went for 3.4.4 and get stuck with the -fPIC stuff. > Presumably I am missing something simple - any help is appreciated > (list of questions below ...). > > `python setup.py install` for scipy leads to: > > [...] > building 'scipy.lib.lapack.flapack' extension > compiling C sources > gcc options: '-pthread -fno-strict-aliasing -DNDEBUG -g -O3 -Wall > -Wstrict-prototypes -fPIC -fPIC' > compile options: '-DATLAS_INFO="\"3.7.11\"" -DNUMERIC_VERSION="\"24.0b2\"" > -DNUMERIC -I/scr/python3/lib -I/scr/python3/include/python2.4 -Ibuild/src > -I/scr/python3/include/python2.4 -c' > /scr/python3/bin/g77 -shared > build/temp.linux-x86_64-2.4/build/src/build/src/Lib/lib/lapack/flapackmodul >e.o build/temp.linux-x86_64-2.4/build/src/fortranobject.o > -L/scr/python3/lib64 -Lbuild/temp.linux-x86_64-2.4 -llapack -lptf77blas > -lptcblas -latlas -lg2c -o > build/lib.linux-x86_64-2.4/scipy/lib/lapack/flapack.so > /usr/bin/ld: /scr/python3/lib64/liblapack.a(slamch.o): relocation > R_X86_64_32S against `a local symbol' can not be used when making a shared > object; recompile with -fPIC > /scr/python3/lib64/liblapack.a: could not read symbols: Bad value > collect2: ld returned 1 exit status > error: Command "/scr/python3/bin/g77 -shared > build/temp.linux-x86_64-2.4/build/src/build/src/Lib/lib/lapack/flapackmodul >e.o build/temp.linux-x86_64-2.4/build/src/fortranobject.o > -L/scr/python3/lib64 -Lbuild/temp.linux-x86_64-2.4 -llapack -lptf77blas > -lptcblas -latlas -lg2c -o > build/lib.linux-x86_64-2.4/scipy/lib/lapack/flapack.so" failed with exit > status 1/usr/bin/ld: /scr/python3/lib64/liblapack.a(slamch.o): relocation > R_X86_64_32S against `a local symbol' can not be used when making a shared > object; recompile with -fPIC > /scr/python3/lib64/liblapack.a: could not read symbols: Bad value > collect2: ld returned 1 exit status -- Brett G. Olivier Postdoctoral Fellow Triple-J Group for Molecular Cell Physiology Stellenbosch University bgoli at sun dot ac dot za http://glue.jjj.sun.ac.za/~bgoli Tel +27-21-8082704 Fax +27-21-8085863 Mobile +27-82-7329306 Let us live!!! Let us love!!! Let us share the deepest secrets of our souls!!! You first. From arnd.baecker at web.de Tue Sep 27 08:03:32 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Tue, 27 Sep 2005 14:03:32 +0200 (CEST) Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: <200509270954.54610.bgoli@sun.ac.za> References: <200509270954.54610.bgoli@sun.ac.za> Message-ID: Hi, On Tue, 27 Sep 2005, Brett Olivier wrote: > Hi > > Try the options suggested in > > http://www.scipy.org/mailinglists/mailman?fn=scipy-user/2005-February/004066.html Excellent - I have overlooked this one - it seems the ATLAS (compile) part works now - *many* thanks!! Now the next problem(s) come up: I) building 'scipy.xplt.gistC' extension /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.so when searching for -lX11 /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.a when searching for -lX11 /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.so when searching for -lX11 /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.a when searching for -lX11 /usr/bin/ld: cannot find -lX11 collect2: ld returned 1 exit status /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.so when searching for -lX11 /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.a when searching for -lX11 /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.so when searching for -lX11 /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.a when searching for -lX11 /usr/bin/ld: cannot find -lX11 collect2: ld returned 1 exit status Could this arise because X11 was not compiled with -fPIC? (is there any way to test this in some way ??) (2) Moving the xplt dir out of the way (just a temporary solution for me, as I will need xplt) leads to: [...] gcc: scipy_core/scipy_base/src/display_test.c scipy_core/scipy_base/src/display_test.c:65: warning: initialization from incompatible pointer type gcc -pthread -shared -fPIC build/temp.linux-x86_64-2.4/scipy_core/scipy_base/src/display_test.o -L/usr/X11R6/lib -Lbuild/temp.linux-x86_64-2.4 -lX11 -o build/lib.linux-x86_64-2.4/scipy_base/display_test.so /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.so when searching for -lX11 /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.a when searching for -lX11 /usr/bin/ld: cannot find -lX11 collect2: ld returned 1 exit status /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.so when searching for -lX11 /usr/bin/ld: skipping incompatible /usr/X11R6/lib/libX11.a when searching for -lX11 /usr/bin/ld: cannot find -lX11 collect2: ld returned 1 exit status error: Command "gcc -pthread -shared -fPIC build/temp.linux-x86_64-2.4/scipy_core/scipy_base/src/display_test.o -L/usr/X11R6/lib -Lbuild/temp.linux-x86_64-2.4 -lX11 -o build/lib.linux-x86_64-2.4/scipy_base/display_test.so" failed with exit status 1 Does this mean I should also build my own X11 for this machine? If so, any recommendations for suitable flags? Many thanks in advance for suggestions (and any cheering up as well ;-), Arnd From arnd.baecker at web.de Tue Sep 27 08:25:14 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Tue, 27 Sep 2005 14:25:14 +0200 (CEST) Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: References: Message-ID: as a quick follow-up: removing the display_test part from scipy_core/scipy_base/setup_scipy_base.py gives me build which finishes without further errors!. scipy.test(10) gives: ====================================================================== FAIL: check_round (scipy.special.basic.test_basic.test_round) ---------------------------------------------------------------------- Traceback (most recent call last): File "/scr/python3/lib/python2.4/site-packages/scipy/special/tests/test_basic.py", line 1789, in check_round assert_array_equal(rnd,rndrl) File "/scr/python3/lib/python2.4/site-packages/scipy_test/testing.py", line 715, in assert_array_equal assert cond,\ AssertionError: Arrays are not equal (mismatch 25.0%): Array 1: [10 10 11 11] Array 2: [10 10 10 11] ---------------------------------------------------------------------- Ran 1279 tests in 35.080s FAILED (failures=1) Out[4]: To me this looks very good, in particular that no ATLAS related errors show up. So only the above error and the X11 problem remain until full happiness. Once this works, should I post (more complete) installation notes at some point? Best, Arnd From alford at wuphys.wustl.edu Tue Sep 27 10:21:28 2005 From: alford at wuphys.wustl.edu (alford at wuphys.wustl.edu) Date: Tue, 27 Sep 2005 09:21:28 -0500 (CDT) Subject: [SciPy-user] making odeint shut up Message-ID: Does anyone know how to stop odeint from displaying error messages like this: lsoda-- at current t (=r1), mxstep (=i1) steps taken on this call before reaching tout in above message, i1 = 500 in above message, r1 = 0.8840344508730E-01 Excess work done on this call (perhaps wrong Dfun type). I would like to tell it to send them to a different output channel (so they don't get mixed up with my print statements), or just suppress them altogether. From noel.oboyle2 at mail.dcu.ie Tue Sep 27 10:49:29 2005 From: noel.oboyle2 at mail.dcu.ie (Noel O'Boyle) Date: Tue, 27 Sep 2005 15:49:29 +0100 Subject: [SciPy-user] Selecting items from an array Message-ID: <1127832569.16760.260.camel@sandwi.ch.cam.ac.uk> Newbie question: I'm trying to select items from a vector using a mask. For example: X = array([100,56,44,99]) mask = ([0,1,0,1]) I want to create: newX = array([56,99]) i.e. only those elements of X for which mask is true. scipy.take will do this but I need to find the indices of all of the 1's in mask. It must be possible to do this with scipy, but how? Regards, Noel From ckkart at hoc.net Tue Sep 27 11:19:19 2005 From: ckkart at hoc.net (Christian Kristukat) Date: Tue, 27 Sep 2005 17:19:19 +0200 Subject: [SciPy-user] Selecting items from an array In-Reply-To: <1127832569.16760.260.camel@sandwi.ch.cam.ac.uk> References: <1127832569.16760.260.camel@sandwi.ch.cam.ac.uk> Message-ID: <433962F7.4050602@hoc.net> Noel O'Boyle wrote: > Newbie question: > > I'm trying to select items from a vector using a mask. > > For example: > X = array([100,56,44,99]) > mask = ([0,1,0,1]) > > I want to create: > newX = array([56,99]) > i.e. only those elements of X for which mask is true. > > scipy.take will do this but I need to find the indices of all of the 1's > in mask. It must be possible to do this with scipy, but how? compress(mask==1, X) or use the MA (masked array) module of Numeric. Regards, Christian From stephen.walton at csun.edu Tue Sep 27 17:31:09 2005 From: stephen.walton at csun.edu (Stephen Walton) Date: Tue, 27 Sep 2005 14:31:09 -0700 Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: References: Message-ID: <4339BA1D.3060805@csun.edu> Arnd Baecker wrote: >scipy.test(10) gives: > >====================================================================== >FAIL: check_round (scipy.special.basic.test_basic.test_round) >---------------------------------------------------------------------- >Traceback (most recent call last): > File >"/scr/python3/lib/python2.4/site-packages/scipy/special/tests/test_basic.py", >line 1789, in check_round > assert_array_equal(rnd,rndrl) > File "/scr/python3/lib/python2.4/site-packages/scipy_test/testing.py", >line 715, in assert_array_equal > assert cond,\ >AssertionError: >Arrays are not equal (mismatch 25.0%): > Array 1: [10 10 11 11] > Array 2: [10 10 10 11] > > I'm getting this same error on my i386 systems using Absoft Fortran and Fedora Core 4, so this problem is not unique to the Opteron architecture. I can't help with the X11 link problem, I'm afraid, as I don't see it on my i386 systems. From rkern at ucsd.edu Tue Sep 27 17:46:49 2005 From: rkern at ucsd.edu (Robert Kern) Date: Tue, 27 Sep 2005 14:46:49 -0700 Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: <4339BA1D.3060805@csun.edu> References: <4339BA1D.3060805@csun.edu> Message-ID: <4339BDC9.6050807@ucsd.edu> Stephen Walton wrote: > Arnd Baecker wrote: > >> scipy.test(10) gives: >> >> ====================================================================== >> FAIL: check_round (scipy.special.basic.test_basic.test_round) >> ---------------------------------------------------------------------- >> Traceback (most recent call last): >> File >> "/scr/python3/lib/python2.4/site-packages/scipy/special/tests/test_basic.py", >> >> line 1789, in check_round >> assert_array_equal(rnd,rndrl) >> File "/scr/python3/lib/python2.4/site-packages/scipy_test/testing.py", >> line 715, in assert_array_equal >> assert cond,\ >> AssertionError: >> Arrays are not equal (mismatch 25.0%): >> Array 1: [10 10 11 11] >> Array 2: [10 10 10 11] >> > I'm getting this same error on my i386 systems using Absoft Fortran and > Fedora Core 4, so this problem is not unique to the Opteron > architecture. I can't help with the X11 link problem, I'm afraid, as I > don't see it on my i386 systems. We've run into this several times before. The test is right according to the documentation. In [5]: round? Type: ufunc String Form: Namespace: Interactive Docstring: y=Returns the nearest integer to x as a double precision floating point result. If x ends in 0.5 exactly, the nearest even integer is chosen. That's the round() from scipy.special which ought to be the one being called by this check_round(). You might want to add a "print round" statement to the check_round() method just to make sure that it is the ufunc from scipy.special and not the alias to around() from Numeric which follows Python's round-up-on-0.5 behavior. The last time we came across it, I had Fernando check a C program that tries the underlying Cephes routine, and it gives the expected answer. -- Robert Kern rkern at ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter From mantha at chem.unr.edu Tue Sep 27 22:52:02 2005 From: mantha at chem.unr.edu (Jordan H. Mantha) Date: Tue, 27 Sep 2005 19:52:02 -0700 (PDT) Subject: [SciPy-user] Re: io.read_array slow Message-ID: Christian Kristukat wrote: >Hi, > I noticed that io.read_array is really slow compared to a python >while/readline/split loop (about 5 times slower). I believe to remember >that at some time it was written in C but when I looked at the source of >scipy 0.3.2 it seems to be pure python. Is there a evident reason why it >must be that slow or are there maybe plans to rewrite it in C? > Regards, >Christian I have noticed this too. I was wrote a fitting script that reads in a 3 column data file that has 502,000 rows. I first started with io.read_array and it took > 10 min. to get the array in. I was reading "Python Scripting for Computational Science" by Hans Petter Langtangen and used the following: data = array([float(x) for x in infile.read().split()], Float) data.shape = (len(data)/3,3) and it cut down the time to read in the data to ~ 5 s. This was a dramatic speed up for me. I know that io.read_array can handle more complicated cases, but it doesn't always seem to be the best solution for reading in data. --Jordan Mantha From arnd.baecker at web.de Wed Sep 28 02:30:16 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Wed, 28 Sep 2005 08:30:16 +0200 (CEST) Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: <4339BDC9.6050807@ucsd.edu> References: <4339BA1D.3060805@csun.edu> <4339BDC9.6050807@ucsd.edu> Message-ID: Hi, On Tue, 27 Sep 2005, Robert Kern wrote: > Stephen Walton wrote: > > Arnd Baecker wrote: > > > >> scipy.test(10) gives: > >> > >> ====================================================================== > >> FAIL: check_round (scipy.special.basic.test_basic.test_round) > >> ---------------------------------------------------------------------- > >> Traceback (most recent call last): > >> File > >> "/scr/python3/lib/python2.4/site-packages/scipy/special/tests/test_basic.py", > >> > >> line 1789, in check_round > >> assert_array_equal(rnd,rndrl) > >> File "/scr/python3/lib/python2.4/site-packages/scipy_test/testing.py", > >> line 715, in assert_array_equal > >> assert cond,\ > >> AssertionError: > >> Arrays are not equal (mismatch 25.0%): > >> Array 1: [10 10 11 11] > >> Array 2: [10 10 10 11] > >> > > I'm getting this same error on my i386 systems using Absoft Fortran and > > Fedora Core 4, so this problem is not unique to the Opteron > > architecture. I can't help with the X11 link problem, I'm afraid, as I > > don't see it on my i386 systems. > We've run into this several times before. Indeed, I should have done a proper search (but from home with modem ...;-( ) > The test is right according to > the documentation. > > In [5]: round? > Type: ufunc > String Form: > Namespace: Interactive > Docstring: > y=Returns the nearest integer to x as a double precision > floating point result. If x ends in 0.5 exactly, the > nearest even integer is chosen. > > That's the round() from scipy.special which ought to be the one being > called by this check_round(). You might want to add a "print round" > statement to the check_round() method just to make sure that it is the > ufunc from scipy.special and not the alias to around() from Numeric > which follows Python's round-up-on-0.5 behavior. It gives: Direct invocation gives In [25]: scipy.special.round(10.5) Out[25]: 11.0 Just to be sure: In [26]: repr(scipy.special.round) Out[26]: "" So this is not the expected behaviour... > The last time we came across it, I had Fernando check a C program that > tries the underlying Cephes routine, and it gives the expected answer. OK, I took the one from http://www.scipy.net/pipermail/scipy-user/2005-June/004577.html This looks like the compiler matters a) with gcc -v Using built-in specs. Target: x86_64-suse-linux Configured with: ../configure --enable-threads=posix --prefix=/usr --with-local-prefix=/usr/local --infodir=/usr/share/info --mandir=/usr/share/man --libdir=/usr/lib64 --libexecdir=/usr/lib64 --enable-languages=c,c++,objc,f95,java,ada --enable-checking --with-gxx-include-dir=/usr/include/c++/4.0.2 --enable-java-awt=gtk --disable-libjava-multilib --with-slibdir=/lib64 --with-system-zlib --enable-shared --enable-__cxa_atexit --without-system-libunwind --host=x86_64-suse-linux Thread model: posix gcc version 4.0.2 20050826 (prerelease) (SUSE Linux) gcc round.c -lm ./a.out 10.4 -> 10 10.5 -> 11 10.6 -> 11 11.4 -> 11 11.5 -> 12 11.6 -> 12 So this result is different from the one you obtain(ed) on your Mac. b) with gcc -v (the one which I thought I used for compiling scipy) Reading specs from /scr/python3/bin/../lib/gcc/x86_64-unknown-linux-gnu/3.4.4/specs Configured with: ../gcc-3.4.4/configure --prefix=/scr/python3/ --enable-shared --enable-threads=posix --enable-__cxa_atexit --enable-clocale=gnu --enable-languages=c,c++,f77,objc Thread model: posix gcc version 3.4.4 ./a.out 10.4 -> 10 10.5 -> 10 10.6 -> 11 11.4 -> 11 11.5 -> 12 11.6 -> 12 Looks alright. What I don't understand: I am pretty sure, that I compiled scipy with gcc 3.4.4 (Is there a way to check this afterwards?). If you have any further ideas what I could do so that we can get rid of this problem, just let me know! Best, and many thanks, Arnd From arnd.baecker at web.de Wed Sep 28 03:52:27 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Wed, 28 Sep 2005 09:52:27 +0200 (CEST) Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: <4339BA1D.3060805@csun.edu> References: <4339BA1D.3060805@csun.edu> Message-ID: On Tue, 27 Sep 2005, Stephen Walton wrote: > I can't help with the X11 link problem, I'm afraid, as I > don't see it on my i386 systems. Googling for "ld: skipping incompatible /usr/X11R6/lib/libX11.so" gives quite a few hits. It looks like a 64Bit vs. 32 Bit thing. And the solution is to put the following into scipy_core/scipy_distutils/site.cfg [x11] library_dirs = /usr/X11R6/lib64 include_dirs = /usr/X11R6/include Then scipy builds and installs fine, but ... In [1]: from scipy import * In [2]: x=arange(10.0) In [3]: xplt.plg(x) --------------------------------------------------------------------------- exceptions.ImportError Traceback (most recent call last) /home/abaecker/ /scr/python3/lib/python2.4/site-packages/scipy_base/ppimport.py in __getattr__(self, name) 301 module = self.__dict__['_ppimport_module'] 302 except KeyError: --> 303 module = self._ppimport_importer() 304 return getattr(module, name) 305 /scr/python3/lib/python2.4/site-packages/scipy_base/ppimport.py in _ppimport_importer(self) 256 module = sys.modules[name] 257 except KeyError: --> 258 raise ImportError,self.__dict__.get('_ppimport_exc_info')[1] 259 if module is not self: 260 exc_info = self.__dict__.get('_ppimport_exc_info') ImportError: Array can not be safely cast to required type Any idea what to do about this one ? Many thanks, Arnd From Fernando.Perez at colorado.edu Wed Sep 28 04:32:12 2005 From: Fernando.Perez at colorado.edu (Fernando Perez) Date: Wed, 28 Sep 2005 02:32:12 -0600 Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: References: <4339BA1D.3060805@csun.edu> Message-ID: <433A550C.4000804@colorado.edu> Arnd Baecker wrote: > On Tue, 27 Sep 2005, Stephen Walton wrote: > > >>I can't help with the X11 link problem, I'm afraid, as I >>don't see it on my i386 systems. > > > Googling for "ld: skipping incompatible /usr/X11R6/lib/libX11.so" > gives quite a few hits. > It looks like a 64Bit vs. 32 Bit thing. > > And the solution is to put the following into > scipy_core/scipy_distutils/site.cfg > > [x11] > library_dirs = /usr/X11R6/lib64 > include_dirs = /usr/X11R6/include > > Then scipy builds and installs fine, but ... > > In [1]: from scipy import * > In [2]: x=arange(10.0) > In [3]: xplt.plg(x) Do you get this error in non-*plt code? If so, we should worry about it, but if it's only being triggered by *plt code, I'm not sure it warrants spending preciously scarce resources, giving that in the new scipy, *plt are going to disappear. There is pretty much consensus that matplotlib will take care of plotting and scipy will drop the old, unmaintained *plt modules. Cheers, f From arnd.baecker at web.de Wed Sep 28 04:34:32 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Wed, 28 Sep 2005 10:34:32 +0200 (CEST) Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: References: Message-ID: On Wed, 28 Sep 2005, Arnd Baecker wrote: [...] > Then scipy builds and installs fine, but ... > > In [1]: from scipy import * > In [2]: x=arange(10.0) > In [3]: xplt.plg(x) > --------------------------------------------------------------------------- > exceptions.ImportError Traceback (most It seems that this is an old-fellow again, http://www.scipy.net/pipermail/scipy-user/2004-October/003496.html related to slice3.py in scipy.xplt and Numeric, as already from scipy.xplt import * fails (see below for the full error message). To test this: import Numeric print Numeric.__version__ import arrayfns x=Numeric.array([[0, 0, 0, 1], [1, 1, 1, 0]]) y=Numeric.array([[0, 1], [0, 1], [1, 0], [1, 0]]) print x.typecode() print y.typecode() arrayfns.find_mask(x, y) gives on my normal desktop machine: In [1]:import Numeric In [2]:print Numeric.__version__ 23.8 In [3]:import arrayfns In [4]:x=Numeric.array([[0, 0, 0, 1], [1, 1, 1, 0]]) In [5]:y=Numeric.array([[0, 1], [0, 1], [1, 0], [1, 0]]) In [6]:print x.typecode() l In [7]:print y.typecode() l In [8]:arrayfns.find_mask(x, y) Out[8]:array([1, 0, 1, 0],'i') However, on the one where `scipy.xplt import *` fails I get: In [1]: import Numeric In [2]: print Numeric.__version__ 24.0b2 In [3]: import arrayfns In [4]: x=Numeric.array([[0, 0, 0, 1], [1, 1, 1, 0]]) In [5]: y=Numeric.array([[0, 1], [0, 1], [1, 0], [1, 0]]) In [6]: print x.typecode() l In [7]: print y.typecode() l In [8]: arrayfns.find_mask(x, y) --------------------------------------------------------------------------- exceptions.TypeError Traceback (most recent call last) /home/abaecker/ TypeError: Array can not be safely cast to required type If one uses instead: In [9]: arrayfns.find_mask(x.astype("b"), y.astype("b")) it works fine. No idea what the reason (64 Bit?) for this changed behaviour of Numeric is. The solution/workaround is: in slice3.py, routine _construct3 change mask = find_mask (below, _node_edges3 [itype]) to mask = find_mask (below.astype("b"), _node_edges3 [itype].astype("b")) With this plotting with scipy.xplt works fine. Best, Arnd In [11]: from scipy.xplt import * --------------------------------------------------------------------------- exceptions.TypeError Traceback (most recent call last) /home/abaecker/ /scr/python3/lib/python2.4/site-packages/scipy/xplt/__init__.py 43 if maxheight is None: 44 maxheight = 768 45 ---> 46 from Mplot import * 47 /scr/python3/lib/python2.4/site-packages/scipy/xplt/Mplot.py 1009 _dheight=6*inches 1010 -> 1011 import colorbar 1012 def imagesc_cb(z,cmin=None,cmax=None,xryr=None,_style='default', 1013 zlabel=None,font='helvetica',fontsize=16,color='black', /scr/python3/lib/python2.4/site-packages/scipy/xplt/colorbar.py 7 from arrayfns import * 8 from gist import * ----> 9 from slice3 import * 10 11 def nice_levels (z, n = 8) : /scr/python3/lib/python2.4/site-packages/scipy/xplt/slice3.py 1579 # ------------------------------------------------------------------------ 1580 -> 1581 _poly_permutations4 = _construct3 (0) 1582 _poly_permutations5 = _construct3 (1) 1583 _poly_permutations6 = _construct3 (2) /scr/python3/lib/python2.4/site-packages/scipy/xplt/slice3.py in _construct3(itype) 1573 # For some reason the node edges for a cell need to be in different order 1574 # here than in slice3 to get the correct results. Hence _node_edges3. -> 1575 mask = find_mask (below, _node_edges3 [itype]) 1576 1577 return construct3 (mask, itype) TypeError: Array can not be safely cast to required type From arnd.baecker at web.de Wed Sep 28 04:57:01 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Wed, 28 Sep 2005 10:57:01 +0200 (CEST) Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: <433A550C.4000804@colorado.edu> References: <433A550C.4000804@colorado.edu> Message-ID: Hi Fernando, On Wed, 28 Sep 2005, Fernando Perez wrote: [...] > Do you get this error in non-*plt code? No, and I think it is solved (though it seems to point to some Numeric problem). > If so, we should worry about it, but > if it's only being triggered by *plt code, I'm not sure it warrants spending > preciously scarce resources, giving that in the new scipy, *plt are going to > disappear. > There is pretty much consensus that matplotlib will take care of > plotting and scipy will drop the old, unmaintained *plt modules. Umpf - has this already been decided? scipy.xplt is very important to me/us as a lot of code used in our group is heavily based on it! I do understand that new users should use matplotlib. But I am not sure if it is fair to say that scipy.xplt is unmaintained as it based on pygist (which is maintained by Michiel de Hoon) and both he and Travis also have been very helpful in fixing bugs for scipy.xplt. Despite the fantastic progress (also in terms of speed) of matplotlib, the people in our group still prefer scipy.xplt for fast, one-off plots to display arrays. Therefore it would be good if scipy.xplt is kept in the new scipy (Maybe with a comment in the documentation: "Only for backwards-compatibility - use matplotlib instead") so that old code can still run with the new scipy. Best, Arnd From oliphant at ee.byu.edu Wed Sep 28 05:10:16 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Wed, 28 Sep 2005 03:10:16 -0600 Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: References: <433A550C.4000804@colorado.edu> Message-ID: <433A5DF8.9090900@ee.byu.edu> Arnd Baecker wrote: > > >>If so, we should worry about it, but >>if it's only being triggered by *plt code, I'm not sure it warrants spending >>preciously scarce resources, giving that in the new scipy, *plt are going to >>disappear. >>There is pretty much consensus that matplotlib will take care of >>plotting and scipy will drop the old, unmaintained *plt modules. >> >> > >Umpf - has this already been decided? >scipy.xplt is very important to me/us as a lot of code used in our group >is heavily based on it! > > Yeah, it's been decided. But, don't panic. I don't think it will be that big of a deal. pygist will still exist and code from scipy.xplt will go into it. There will still be a scipy.xplt package (just not bundled with the new system). All it really means is that there will be an additional package to install. We are moving plotting outside of the scipy system. -Travis From Fernando.Perez at colorado.edu Wed Sep 28 05:11:48 2005 From: Fernando.Perez at colorado.edu (Fernando Perez) Date: Wed, 28 Sep 2005 03:11:48 -0600 Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: References: <433A550C.4000804@colorado.edu> Message-ID: <433A5E54.5000709@colorado.edu> Arnd Baecker wrote: >>If so, we should worry about it, but >>if it's only being triggered by *plt code, I'm not sure it warrants spending >>preciously scarce resources, giving that in the new scipy, *plt are going to >>disappear. >>There is pretty much consensus that matplotlib will take care of >>plotting and scipy will drop the old, unmaintained *plt modules. > > > Umpf - has this already been decided? > scipy.xplt is very important to me/us as a lot of code used in our group > is heavily based on it! > > I do understand that new users should use matplotlib. > But I am not sure if it is fair to say that scipy.xplt is unmaintained > as it based on pygist (which is maintained by Michiel de Hoon) > and both he and Travis also have been very helpful in fixing > bugs for scipy.xplt. > Despite the fantastic progress (also in terms of speed) of matplotlib, > the people in our group still prefer scipy.xplt for fast, one-off plots > to display arrays. > > Therefore it would be good if scipy.xplt is kept in > the new scipy (Maybe with a comment in the documentation: > "Only for backwards-compatibility - use matplotlib instead") > so that old code can still run with the new scipy. Well, I certainly don't maintain it, so it's ultimately not my call to make. It's a matter of manpower. My impression has always been that there isn't enough of that to maintain the scipy.*plt family when John and his team are taking care of matplotlib. And just as much as in the future we hope that the numerical functionality in matplotlib will migrate (gently) towards scipy, I think that scipy's core should equally divest itself from plotting conerns. There's only so much that can be effectively maintained, documented and developed, and dividing responsibility is the first step towards a sustainable model. I worry that by trying to carry *plt around, scipy hurts itself, given the limited human resources available, which are already stretched pretty thin. While matplotlib may not be perfect, I'd rather see Travis' and Robert's time go towards working on core scipy issues, and let John and his group nibble away at matplotlib's remaining problems. There is _plenty_ of work to do in scipy, believe me :) But again: these are not my personal responsibility, so ultimately the call will be made by Travis, Robert and the other core developers. But my vote is for deprecation, even if it has some negative impact on people. I'd rather distribute it as a standalone package _outside_ of scipy, which can be used (and perhaps maintained in the future) by those who actually need it, without it being in the scipy core libraries. Cheers, f From arnd.baecker at web.de Wed Sep 28 06:29:22 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Wed, 28 Sep 2005 12:29:22 +0200 (CEST) Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: <433A5DF8.9090900@ee.byu.edu> References: <433A5DF8.9090900@ee.byu.edu> Message-ID: On Wed, 28 Sep 2005, Travis Oliphant wrote: > Arnd Baecker wrote: > > >>If so, we should worry about it, but > >>if it's only being triggered by *plt code, I'm not sure it warrants spending > >>preciously scarce resources, giving that in the new scipy, *plt are going to > >>disappear. > >>There is pretty much consensus that matplotlib will take care of > >>plotting and scipy will drop the old, unmaintained *plt modules. > >> > > > >Umpf - has this already been decided? > >scipy.xplt is very important to me/us as a lot of code used in our group > >is heavily based on it! > > > Yeah, it's been decided. But, don't panic. I don't think it will be > that big of a deal. > > pygist will still exist and code from scipy.xplt will go into it. > There will still be a scipy.xplt package (just not bundled with the new > system). > > All it really means is that there will be an additional package to install. That sounds very good - in particular if then `scipy.xplt.*` can still be used as before it won't break any old code. ((and if it could still be bundled with the Windows Sumo Package, at least for a while, everyone should be happy ...) Thanks for the clarification, best, Arnd From ckkart at hoc.net Wed Sep 28 08:27:05 2005 From: ckkart at hoc.net (Christian Kristukat) Date: Wed, 28 Sep 2005 14:27:05 +0200 Subject: [SciPy-user] Re: io.read_array slow In-Reply-To: References: Message-ID: <433A8C19.8010708@hoc.net> Jordan H. Mantha wrote: > Christian Kristukat wrote: > > >>Hi, >>I noticed that io.read_array is really slow compared to a python >>while/readline/split loop (about 5 times slower). I believe to remember >>that at some time it was written in C but when I looked at the source of >>scipy 0.3.2 it seems to be pure python. Is there a evident reason why it >>must be that slow or are there maybe plans to rewrite it in C? >>Regards, >>Christian > > > > I have noticed this too. I was wrote a fitting script that reads in a 3 > column data file that has 502,000 rows. I first started with io.read_array > and it took > 10 min. to get the array in. I was reading "Python Scripting > for Computational Science" by Hans Petter Langtangen and used the > following: > > data = array([float(x) for x in infile.read().split()], Float) > data.shape = (len(data)/3,3) > Nice short solution indeed. I think read_array could be cut down to something less versatile such that it can handle different separators and works at C speed. Regards, Christian From arnd.baecker at web.de Wed Sep 28 09:02:13 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Wed, 28 Sep 2005 15:02:13 +0200 (CEST) Subject: [SciPy-user] Installation on dual opteron ... In-Reply-To: <433A5E54.5000709@colorado.edu> References: <433A5E54.5000709@colorado.edu> Message-ID: On Wed, 28 Sep 2005, Fernando Perez wrote: > Arnd Baecker wrote: [...] > > Therefore it would be good if scipy.xplt is kept in > > the new scipy (Maybe with a comment in the documentation: > > "Only for backwards-compatibility - use matplotlib instead") > > so that old code can still run with the new scipy. > > Well, I certainly don't maintain it, so it's ultimately not my call to make. > It's a matter of manpower. My impression has always been that there isn't > enough of that to maintain the scipy.*plt family when John and his team are > taking care of matplotlib. And just as much as in the future we hope that the > numerical functionality in matplotlib will migrate (gently) towards scipy, I > think that scipy's core should equally divest itself from plotting conerns. > > There's only so much that can be effectively maintained, documented and > developed, and dividing responsibility is the first step towards a sustainable > model. I worry that by trying to carry *plt around, scipy hurts itself, given > the limited human resources available, which are already stretched pretty thin. > > While matplotlib may not be perfect, I'd rather see Travis' and Robert's time > go towards working on core scipy issues, and let John and his group nibble > away at matplotlib's remaining problems. There is _plenty_ of work to do in > scipy, believe me :) All valid and true points - and indeed I do believe you that there are loads things to be done... It really depends on how much work maintaining scipy.xplt is, and this has to be judged by someone else, not me, as I am just stating an end-users wish on scipy.xplt. > But again: these are not my personal responsibility, so ultimately the call > will be made by Travis, Robert and the other core developers. But my vote is > for deprecation, even if it has some negative impact on people. Deprecation is fine with me, but it would be good, if it still could be used for backwards-compatibility reason (I would really like to avoid rewriting a lot of code, in particular as long as a couple of things cannot be transferred directly from scipy.xplt to matplotlib) But this is not in contradiction with your point > I'd rather > distribute it as a standalone package _outside_ of scipy, which can be used > (and perhaps maintained in the future) by those who actually need it, without > it being in the scipy core libraries. and Travis solution, to have this as a separate, still easy-to-install package. In particular, when one still can use `scipy.xplt.*`, all the old code will work and everything is fine. Best, Arnd From arnd.baecker at web.de Wed Sep 28 09:06:37 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Wed, 28 Sep 2005 15:06:37 +0200 (CEST) Subject: [SciPy-user] Re: io.read_array slow In-Reply-To: <433A8C19.8010708@hoc.net> References: <433A8C19.8010708@hoc.net> Message-ID: On Wed, 28 Sep 2005, Christian Kristukat wrote: > Jordan H. Mantha wrote: > > Christian Kristukat wrote: > > > >>Hi, > >>I noticed that io.read_array is really slow compared to a python > >>while/readline/split loop (about 5 times slower). I believe to remember > >>that at some time it was written in C but when I looked at the source of > >>scipy 0.3.2 it seems to be pure python. Is there a evident reason why it > >>must be that slow or are there maybe plans to rewrite it in C? > >>Regards, > >>Christian > > > > I have noticed this too. I was wrote a fitting script that reads in a 3 > > column data file that has 502,000 rows. I first started with io.read_array > > and it took > 10 min. to get the array in. I was reading "Python Scripting > > for Computational Science" by Hans Petter Langtangen and used the > > following: > > > > data = array([float(x) for x in infile.read().split()], Float) > > data.shape = (len(data)/3,3) > > > > Nice short solution indeed. I think read_array could be cut down to something > less versatile such that it can handle different separators and works at C speed. You could also have a look at TableIO, http://php.iupui.edu/~mmiller3/python/ Best, Arnd From europa100 at comcast.net Wed Sep 28 09:24:00 2005 From: europa100 at comcast.net (Rob) Date: Wed, 28 Sep 2005 06:24:00 -0700 Subject: [SciPy-user] pyvtk Message-ID: <433A9970.4000006@comcast.net> Is this list a proper place to be discussing pyvtk? I have found no other. Thanks, Rob. From arnd.baecker at web.de Wed Sep 28 09:33:48 2005 From: arnd.baecker at web.de (Arnd Baecker) Date: Wed, 28 Sep 2005 15:33:48 +0200 (CEST) Subject: [SciPy-user] pyvtk In-Reply-To: <433A9970.4000006@comcast.net> References: <433A9970.4000006@comcast.net> Message-ID: On Wed, 28 Sep 2005, Rob wrote: > Is this list a proper place to be discussing pyvtk? I have found no other. maybe the mayavi (http://mayavi.sourceforge.net/) mailing list is an even better place. Arnd From yaroslavvb at gmail.com Wed Sep 28 13:33:24 2005 From: yaroslavvb at gmail.com (Yaroslav Bulatov) Date: Wed, 28 Sep 2005 10:33:24 -0700 Subject: [SciPy-user] 3d plotting question Message-ID: There are so many 3d plotting packages, but I'm looking for something simple that'll resemble 3d plots of Gnuplot/Mathematica, can someone recommend one? -- Yaroslav Bulatov bulatov at cs.oregonstate.edu Dearborn 102 Oregon State University Corvallis, OR From rutt at bmi.osu.edu Wed Sep 28 16:25:20 2005 From: rutt at bmi.osu.edu (Benjamin Rutt) Date: Wed, 28 Sep 2005 16:25:20 -0400 Subject: [SciPy-user] build error on itanium2 Message-ID: <5vwtl1ym6n.fsf@akron.bmi.ohio-state.edu> I'm trying to build scipy from source; I've installed f2py, lapack and blas first, and set BLAS and LAPACK. I seem to be getting the following error, on Itanium2. Reading play/unix/README.fpu revealed some discussion; should I be enabling FPU_IGNORE? And if so, how? Or, could I try not building xplt somehow? Thank you. gcc: Lib/xplt/src/play/unix/fpuset.c Lib/xplt/src/play/unix/fpuset.c:304:2: #error Lib/xplt/src/play/unix/fpuset.c:304:2: #error error: Command "gcc -pthread -fno-strict-aliasing -DNDEBUG -g -O3 -Wall -Wstric t-prototypes -fPIC -Ibuild/temp.linux-ia64-2.4/config_pygist -ILib/xplt/src/gis t -ILib/xplt/src/play -ILib/xplt/src/play/unix -I/usr/X11R6/include -I/usr/X11R 6/include -I/home/ac/rutt/.upak/installed/python-2.4/include/python2.4 -c Lib/x plt/src/play/unix/fpuset.c -o build/temp.linux-ia64-2.4/Lib/xplt/src/play/unix/ fpuset.o -DGISTPATH="\"/home/ac/rutt/.upak/installed/python-2.4/lib/python2.4/s ite-packages/scipy/xplt/gistdata\""" failed with exit status 1 From ryanfedora at comcast.net Wed Sep 28 17:57:41 2005 From: ryanfedora at comcast.net (Ryan Krauss) Date: Wed, 28 Sep 2005 17:57:41 -0400 Subject: [SciPy-user] 3d plotting question In-Reply-To: References: Message-ID: <433B11D5.3040404@comcast.net> You can actually use Gnuplot from Python. I have had the best luck with the Gnuplot.py package (http://gnuplot-py.sourceforge.net/). If it doesn't implement the features you want, you can send pure gnuplot code from your python script to the gnuplot session. Mayavi is pretty good but I think a bit more complicated (http://mayavi.sourceforge.net/). It can do some neat animations. But, it doesn't do as nice a job of generating post-script files as gnuplot. I use mayavi for animations and gnuplot.py for print quality graphics. Ryan Yaroslav Bulatov wrote: > There are so many 3d plotting packages, but I'm looking for something > simple that'll resemble 3d plots of Gnuplot/Mathematica, can someone > recommend one? > > -- > Yaroslav Bulatov > bulatov at cs.oregonstate.edu > Dearborn 102 > Oregon State University > Corvallis, OR > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user > From grante at visi.com Wed Sep 28 21:44:17 2005 From: grante at visi.com (Grant Edwards) Date: Thu, 29 Sep 2005 01:44:17 +0000 (UTC) Subject: [SciPy-user] Re: 3d plotting question References: Message-ID: On 2005-09-28, Yaroslav Bulatov wrote: > There are so many 3d plotting packages, but I'm looking for > something simple that'll resemble 3d plots of > Gnuplot/Mathematica, can someone recommend one? I've found that Gnuplot does a very good job of resembling Gnuplot. ;) http://gnuplot-py.sourceforge.net/ -- Grant Edwards grante Yow! My EARS are GONE!! at visi.com From lanceboyle at cwazy.co.uk Wed Sep 28 23:30:22 2005 From: lanceboyle at cwazy.co.uk (Lance Boyle) Date: Wed, 28 Sep 2005 20:30:22 -0700 Subject: [SciPy-user] 3d plotting question In-Reply-To: <433B11D5.3040404@comcast.net> References: <433B11D5.3040404@comcast.net> Message-ID: I might have posted this little bit of code before, but it's not too much so I hope no one minds seeing it again. It's not mine, so for proper cretids, Google a unique-looking line and it should pop right up. With this, you can get to gnuplot with a minimum of fuss. Note that it sets up AquaTerm (that's a "plotting terminal" for OS X that uses native antialiased graphics with PDF saving, etc. If you don't use it, delete the line or set to your own terminal. import os class gnuplot: def __init__(self): print "opening new gnuplot session..." self.session = os.popen("gnuplot","w") def __del__(self): print "closing gnuplot session..." self.session.close() def send(self, cmd): self.session.write(cmd+'\n') self.session.flush() if __name__=="__main__": g1 = gnuplot() g1.send("set terminal aqua 0") # Your code here g1.send("plot sin(x)") del g1 On Sep 28, 2005, at 2:57 PM, Ryan Krauss wrote: > You can actually use Gnuplot from Python. I have had the best luck > with the Gnuplot.py package (http://gnuplot-py.sourceforge.net/). > If it doesn't implement the features you want, you can send pure > gnuplot code from your python script to the gnuplot session. > > Mayavi is pretty good but I think a bit more complicated (http:// > mayavi.sourceforge.net/). It can do some neat animations. But, it > doesn't do as nice a job of generating post-script files as > gnuplot. I use mayavi for animations and gnuplot.py for print > quality graphics. > > Ryan > > Yaroslav Bulatov wrote: > >> There are so many 3d plotting packages, but I'm looking for something >> simple that'll resemble 3d plots of Gnuplot/Mathematica, can someone >> recommend one? >> -- >> Yaroslav Bulatov >> bulatov at cs.oregonstate.edu >> Dearborn 102 >> Oregon State University >> Corvallis, OR >> _______________________________________________ >> SciPy-user mailing list >> SciPy-user at scipy.net >> http://www.scipy.net/mailman/listinfo/scipy-user >> > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user > From prabhu_r at users.sf.net Thu Sep 29 00:55:50 2005 From: prabhu_r at users.sf.net (Prabhu Ramachandran) Date: Thu, 29 Sep 2005 10:25:50 +0530 Subject: [SciPy-user] 3d plotting question In-Reply-To: <433B11D5.3040404@comcast.net> References: <433B11D5.3040404@comcast.net> Message-ID: <17211.29654.146184.626086@monster.linux.in> >>>>> "Ryan" == Ryan Krauss writes: Ryan> Mayavi is pretty good but I think a bit more complicated Ryan> (http://mayavi.sourceforge.net/). It can do some neat Ryan> animations. But, it doesn't do as nice a job of generating Ryan> post-script files as gnuplot. Ryan> I use mayavi for animations and gnuplot.py for print Ryan> quality graphics. Thanks. Just a clarification. It is not easy to generate vector graphics output for complex visualizations. However, if you build VTK with GL2PS support it should be possible to generate true vector PS/EPS/PDF output. This works pretty well for simple visualizations. IMHO, MayaVi is complicated primarily because you need to create your data as per some format or another. I am hoping to do something about that in the next several months. If you are willing to try and go over a few hoops installing the enthought tool suite (no installers yet) you can get it from SVN and try out the tools inside the TVTK package. Some preliminary information is available here (I'll update the pages over the next week): http://www.enthought.com/enthought/wiki/TVTK There are links to papers and presentations on TVTK there. With TVTK's mlab module (enthought.tvtk.tools.mlab) you can create simple visualizations with pure Numeric/numarray data. It is like mayavi.tools.imv but a lot better with many more features. I'll announce this as soon as I have written a little more documentation on installation of the enthought tool suite and a better TVTK web page. Hopefully that will happen by next week. If you are interested to check it out right away, it might be a good idea to just read the EuroPython2005 paper and presentation at the above site. cheers, prabhu From philopensource at hotmail.com Thu Sep 29 05:42:53 2005 From: philopensource at hotmail.com (philopensource at hotmail.com) Date: Thu, 29 Sep 2005 11:42:53 +0200 Subject: [SciPy-user] How to wrap numarray with pygrid in python? Message-ID: <433BB71D.6040007@hotmail.com> Hi everyone, My problem is : I have a multidimensionnal numarray.array and i need to display it in a spreadsheet (like pygrid). I need to wrap numarray with pygrig. Although i read some tutorial how to create python wrapper. (Perhaps with swig). Can anybody help me or tell me where to find some ressources? Thanks a lot, Philippe Collet From ckkart at hoc.net Thu Sep 29 06:37:59 2005 From: ckkart at hoc.net (Christian Kristukat) Date: Thu, 29 Sep 2005 12:37:59 +0200 Subject: [SciPy-user] 3d plotting question In-Reply-To: <17211.29654.146184.626086@monster.linux.in> References: <433B11D5.3040404@comcast.net> <17211.29654.146184.626086@monster.linux.in> Message-ID: <433BC407.1050609@hoc.net> Prabhu Ramachandran wrote: [...] > that in the next several months. > > If you are willing to try and go over a few hoops installing the > enthought tool suite (no installers yet) you can get it from SVN and > try out the tools inside the TVTK package. Some preliminary > information is available here (I'll update the pages over the next > week): > > http://www.enthought.com/enthought/wiki/TVTK > That sounds very interesting. Do I need a complete enthought installation to be able to use tvtk? Anyway I couldn't build enthought so I would need some hints here. I tried python setup_enthought.py build in a fresh svn copy and I get the following error: Traceback (most recent call last): File "setup_enthought.py", line 143, in ? setup_package() File "setup_enthought.py", line 135, in setup_package url="http://www.scipy.org/", File "/usr/lib/python2.3/site-packages/scipy_distutils/core.py", line 73, in setup return old_setup(**new_attr) File "/usr/lib/python2.3/distutils/core.py", line 149, in setup dist.run_commands() File "/usr/lib/python2.3/distutils/dist.py", line 907, in run_commands self.run_command(cmd) File "/usr/lib/python2.3/distutils/dist.py", line 927, in run_command cmd_obj.run() File "/usr/lib/python2.3/distutils/command/build.py", line 107, in run self.run_command(cmd_name) File "/usr/lib/python2.3/distutils/cmd.py", line 333, in run_command self.distribution.run_command(command) File "/usr/lib/python2.3/distutils/dist.py", line 927, in run_command cmd_obj.run() File "/usr/lib/python2.3/site-packages/scipy_distutils/command/build_src.py", line 129, in run self.build_sources() File "/usr/lib/python2.3/site-packages/scipy_distutils/command/build_src.py", line 215, in build_s ources self.build_extension_sources(ext) File "/usr/lib/python2.3/site-packages/scipy_distutils/command/build_src.py", line 249, in build_e xtension_sources sources = self.generate_sources(sources, ext) File "/usr/lib/python2.3/site-packages/scipy_distutils/command/build_src.py", line 291, in generat e_sources source = func(extension, build_dir) File "/usr/lib/python2.3/site-packages/scipy_distutils/misc_util.py", line 565, in generate_svn_ve rsion_py local_path = extension.local_path AttributeError: Extension instance has no attribute 'local_path' Regards, Christian From ryanfedora at comcast.net Thu Sep 29 09:06:23 2005 From: ryanfedora at comcast.net (Ryan Krauss) Date: Thu, 29 Sep 2005 09:06:23 -0400 Subject: [SciPy-user] 3d plotting question In-Reply-To: <17211.29654.146184.626086@monster.linux.in> References: <433B11D5.3040404@comcast.net> <17211.29654.146184.626086@monster.linux.in> Message-ID: <433BE6CF.80003@comcast.net> Prabhu, I would be really interested in getting eps/pdf output. How hard would it be to get this working in Windows? I don't even mind if the drawing part of the figure is somewhat pixelated, I just think it looks bad when the axes and labels aren't type set well. Ryan Prabhu Ramachandran wrote: >>>>>>"Ryan" == Ryan Krauss writes: > > > Ryan> Mayavi is pretty good but I think a bit more complicated > Ryan> (http://mayavi.sourceforge.net/). It can do some neat > Ryan> animations. But, it doesn't do as nice a job of generating > Ryan> post-script files as gnuplot. > Ryan> I use mayavi for animations and gnuplot.py for print > Ryan> quality graphics. > > Thanks. Just a clarification. It is not easy to generate vector > graphics output for complex visualizations. However, if you build VTK > with GL2PS support it should be possible to generate true vector > PS/EPS/PDF output. This works pretty well for simple visualizations. > > IMHO, MayaVi is complicated primarily because you need to create your > data as per some format or another. I am hoping to do something about > that in the next several months. > > If you are willing to try and go over a few hoops installing the > enthought tool suite (no installers yet) you can get it from SVN and > try out the tools inside the TVTK package. Some preliminary > information is available here (I'll update the pages over the next > week): > > http://www.enthought.com/enthought/wiki/TVTK > > There are links to papers and presentations on TVTK there. With > TVTK's mlab module (enthought.tvtk.tools.mlab) you can create simple > visualizations with pure Numeric/numarray data. It is like > mayavi.tools.imv but a lot better with many more features. I'll > announce this as soon as I have written a little more documentation on > installation of the enthought tool suite and a better TVTK web page. > Hopefully that will happen by next week. If you are interested to > check it out right away, it might be a good idea to just read the > EuroPython2005 paper and presentation at the above site. > > cheers, > prabhu > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user > From prabhu_r at users.sf.net Thu Sep 29 09:15:57 2005 From: prabhu_r at users.sf.net (Prabhu Ramachandran) Date: Thu, 29 Sep 2005 18:45:57 +0530 Subject: [SciPy-user] 3d plotting question In-Reply-To: <433BC407.1050609@hoc.net> References: <433B11D5.3040404@comcast.net> <17211.29654.146184.626086@monster.linux.in> <433BC407.1050609@hoc.net> Message-ID: <17211.59661.320358.1243@monster.linux.in> >>>>> "Christian" == Christian Kristukat writes: >> http://www.enthought.com/enthought/wiki/TVTK Christian> That sounds very interesting. Do I need a complete Christian> enthought installation to be able to use tvtk? Anyway I Christian> couldn't build enthought so I would need some hints Christian> here. I tried python setup_enthought.py build in a Christian> fresh svn copy and I get the following error: Thanks for trying this out. In the next couple of days I'll create a wiki page for installation and also post on this list with a link. For now here are some initial instructions (which will eventually get on the wiki). Requirements ------------ 1. SciPy (and therefore numerix) 2. You will need arrayobject.h (i.e. the dev package for numerix). 3. A working VTK-Python installation. 4. For the UI you will need a working wxPython-2.6.x. 5. *Optionally* (if you want to try chaco, enable, kiva etc.) you will need SWIG to build kiva. This is *not* necessary for TVTK. Note that SWIG is packaged with Enthon (see below). Note that none of this is really tested under Mac OS X. I'll defer to Robert Kern for issues on that platform :). TVTK and MayaVi2 and friends all work on Linux and Win32. On Win32: --------- Under win32 you can simply install this version of Enthon (it works for me on XP) to get *all* the requirements: http://download.enthought.com/enthought_python-2.3.5-1076-wx26-n238.exe In fact, if you install this, you may simply get the enthought svn tree (as shown below) and double click on enthought/src/build_inplace.bat and you should be all set except you'll need to set your PYTHONPATH to point to the enthought directory (also shown below). Building and setup ------------------ Once you have the requirements setup. You can do the following: $ svn co http://www.enthought.com/svn/enthought/trunk enthought $ # Build traits. $ cd enthought/src/lib/enthought/traits $ python setup.py build_ext --inplace $ # Build tvtk $ cd ../tvtk $ python setup.py build_ext --inplace # For want of a working installer, lets just set PYTHONPATH for now. $ export PYTHONPATH=$PYTHONPATH:/path/to/enthought/src/lib/ Then try this: $ cd examples $ python tiny_mesh.py $ python simple.py $ python ivtk_example.py If that works, you are ready to play with mlab.py which is inside enthought/tvtk/tools/mlab.py. You can run the file as it is, to see a spherical harmonic on screen. See the source for some nice examples. If you have come this far, I hope you are rewarded by all the new tools you can play with. :) cheers, prabhu From wjdandreta at att.net Thu Sep 29 09:24:33 2005 From: wjdandreta at att.net (Bill Dandreta) Date: Thu, 29 Sep 2005 09:24:33 -0400 Subject: [SciPy-user] 3d plotting question In-Reply-To: <433BE6CF.80003@comcast.net> References: <433B11D5.3040404@comcast.net> <17211.29654.146184.626086@monster.linux.in> <433BE6CF.80003@comcast.net> Message-ID: <433BEB11.8030902@att.net> Ryan Krauss wrote: > Prabhu, > > I would be really interested in getting eps/pdf output. How hard > would it be to get this working in Windows? Have you tried ReportLab? http://www.reportlab.org/ Bill From ckkart at hoc.net Thu Sep 29 09:35:01 2005 From: ckkart at hoc.net (Christian Kristukat) Date: Thu, 29 Sep 2005 15:35:01 +0200 Subject: [SciPy-user] 3d plotting question In-Reply-To: <17211.59661.320358.1243@monster.linux.in> References: <433B11D5.3040404@comcast.net> <17211.29654.146184.626086@monster.linux.in> <433BC407.1050609@hoc.net> <17211.59661.320358.1243@monster.linux.in> Message-ID: <433BED85.7040403@hoc.net> Prabhu Ramachandran wrote: > If that works, you are ready to play with mlab.py which is inside > enthought/tvtk/tools/mlab.py. You can run the file as it is, to see a > spherical harmonic on screen. See the source for some nice examples. > > If you have come this far, I hope you are rewarded by all the new > tools you can play with. :) Wow!!! That was easy. I'm very impressed. Only the ivtk example isn't working. I'm using vtk python 4.2.6. The traceback follows: Exception occurred in traits notification handler Traceback (most recent call last): File "/home/ck/testarea/trunk/src/lib/enthought/traits/trait_notifiers.py", line 214, in rebind_call_1 getattr( self.object(), self.name )( new ) File "/home/ck/testarea/trunk/src/lib/enthought/pyface/tvtk/scene.py", line 699, in _background_changed self.render() File "/home/ck/testarea/trunk/src/lib/enthought/pyface/tvtk/scene.py", line 205, in render self.control.Render() File "/usr/lib/python2.3/site-packages/vtk_python/vtk/wx/wxVTKRenderWindowInteractor.py", line 339, in Render self._Iren.GetRenderWindow().SetNextWindowInfo(str(self.GetHandle())) AttributeError: SetNextWindowInfo Aborted Regards, Christian From ryanfedora at comcast.net Thu Sep 29 09:36:43 2005 From: ryanfedora at comcast.net (Ryan Krauss) Date: Thu, 29 Sep 2005 09:36:43 -0400 Subject: [SciPy-user] 3d plotting question In-Reply-To: <433BEB11.8030902@att.net> References: <433B11D5.3040404@comcast.net> <17211.29654.146184.626086@monster.linux.in> <433BE6CF.80003@comcast.net> <433BEB11.8030902@att.net> Message-ID: <433BEDEB.7020502@comcast.net> I was speaking of pdf output of 3D visualizations. I would use LaTeX for the actual report generation. Ryan Bill Dandreta wrote: > Ryan Krauss wrote: > >> Prabhu, >> >> I would be really interested in getting eps/pdf output. How hard >> would it be to get this working in Windows? > > > Have you tried ReportLab? > > http://www.reportlab.org/ > > Bill > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user > From prabhu_r at users.sf.net Thu Sep 29 09:44:29 2005 From: prabhu_r at users.sf.net (Prabhu Ramachandran) Date: Thu, 29 Sep 2005 19:14:29 +0530 Subject: [SciPy-user] 3d plotting question In-Reply-To: <433BE6CF.80003@comcast.net> References: <433B11D5.3040404@comcast.net> <17211.29654.146184.626086@monster.linux.in> <433BE6CF.80003@comcast.net> Message-ID: <17211.61373.363438.693541@monster.linux.in> >>>>> "Ryan" == Ryan Krauss writes: Ryan> Prabhu, I would be really interested in getting eps/pdf Ryan> output. How hard would it be to get this working in Ryan> Windows? If you use the updated VTK-Python zip file that is available here: http://mayavi.sourceforge.net/dwnld/vtk/win32/VTK-Python-4.4.zip then gl2ps support should be available. The new MayaVi-1.5 installer does have gl2ps support. The above zip file is made using Enthon's latest VTK dlls. Earlier versions did not support gl2ps but the one I sited in my previous mail does support gl2ps. So, if you install this version of VTK and then save a mayavi image via gl2ps (there is a menu option for this) you should get vector EPS/PDF output. Ryan> I don't even mind if the drawing part of the figure is Ryan> somewhat pixelated, I just think it looks bad when the axes Ryan> and labels aren't type set well. By default vtk's gl2ps exporter dumps every primitive that it can (it can't do image actors though) as a vector graphic. However for complex visualizations you can turn on the 'Write3DPropsAsRasterImage' on then all 3d primitives are dumped as an image and all 2d primitives (like text) will be rendered as vector graphics. For example, try gl2ps on the heart.mv example. It will take a while to write but should work. HTH. cheers, prabhu From pecontal at obs.univ-lyon1.fr Thu Sep 29 09:52:27 2005 From: pecontal at obs.univ-lyon1.fr (emmanuel pecontal) Date: Thu, 29 Sep 2005 15:52:27 +0200 Subject: [SciPy-user] splrep problem... Message-ID: <200509291552.28004.pecontal@obs.univ-lyon1.fr> Hello, I am trying to use the splrep function to fit data with a B-spline. It works well when the knots are computed by the routine (ie task=0). But when I provide the knots myself I have a segmentation error. And if I provide the knots list computed by splrep in a previous call with task=0, I have the same behaviour. For example: a = arange(100)*2*pi/10. tck = interpolate.splrep(a,sin(a),s=1) tck = interpolate.splrep(a,sin(a),task=-1,t=tck[0]) -> segmentation error Does someone has an idea of what is happening? -- Emmanuel P?contal CRAL - Observatoire de Lyon 9, Av. Charles Andre F-69561 Saint Genis Laval Cedex tel (33) (0)4.78.86.83.76 - fax (33) (0)4.78.86.83.86 email : pecontal at obs.univ-lyon1.fr From prabhu_r at users.sf.net Thu Sep 29 10:14:10 2005 From: prabhu_r at users.sf.net (Prabhu Ramachandran) Date: Thu, 29 Sep 2005 19:44:10 +0530 Subject: [SciPy-user] 3d plotting question In-Reply-To: <433BED85.7040403@hoc.net> References: <433B11D5.3040404@comcast.net> <17211.29654.146184.626086@monster.linux.in> <433BC407.1050609@hoc.net> <17211.59661.320358.1243@monster.linux.in> <433BED85.7040403@hoc.net> Message-ID: <17211.63154.449495.387875@monster.linux.in> >>>>> "Christian" == Christian Kristukat writes: Christian> Prabhu Ramachandran wrote: >> If that works, you are ready to play with mlab.py which is >> inside enthought/tvtk/tools/mlab.py. You can run the file as >> it is, to see a spherical harmonic on screen. See the source >> for some nice examples. >> >> If you have come this far, I hope you are rewarded by all the >> new tools you can play with. :) Christian> Wow!!! That was easy. I'm very impressed. Only the ivtk Christian> example isn't working. I'm using vtk python 4.2.6. The Christian> traceback follows: Terrific! VTK-4.4 is worth upgrading to if possible. In any case, I forgot that you can also run all the tests in tvtk/tests to make sure the core functionality works. I do the following: $ cd tvtk/tests $ for i in *.py; do echo $i; python $i; done [...] Christian> "/usr/lib/python2.3/site-packages/vtk_python/vtk/wx/wxVTKRenderWindowInteractor.py", Christian> line 339, in Render Christian> self._Iren.GetRenderWindow().SetNextWindowInfo(str(self.GetHandle())) If you are willing to do a little bit of surgery, just replace that file (wxVTKRenderWindowInteractor.py), with the latest from here: http://public.kitware.com/cgi-bin/viewcvs.cgi/Wrapping/Python/vtk/wx/wxVTKRenderWindowInteractor.py?rev=1.17&view=log Then it should hopefully work provided you have wxPython-2.6.x. Good luck and please let us know how this goes. cheers, prabhu From rcsqtc at iiqab.csic.es Thu Sep 29 10:49:56 2005 From: rcsqtc at iiqab.csic.es (Ramon Crehuet) Date: Thu, 29 Sep 2005 16:49:56 +0200 Subject: [SciPy-user] mdp question Message-ID: <433BFF14.6050200@iiqab.csic.es> Hi all, I am using mdp to perform a PCA on some (dummy) variables. I can get the principal components but I do not know how to get the "explained variance" or change the number of output components. Here is my input: import mdp from scipy import array x=array([[1.51, -2.10, 1.91], [-2.10, 0.45, -1.25], [1.59, -0.51, 0.50], [0.45, -0.93, 1.51], [0.04, -0.49,-0.09], [-0.78, 1.89,-1.55], [-0.62,0.00, -0.42]]) pca=mdp.nodes.PCANode() pca.train(x) pca.stop_training() mat=pca.get_projmatrix() # works fine. x=pca.execute(x,1) #works fine, too. print pca.get_explained_variance() # This gives "None" pca.set_output_dim(1) # and this raises an exception: Traceback (most recent call last): File "/home/ramon/python/pca.py", line 21, in -toplevel- pca.set_output_dim(1) File "/usr/lib/python2.3/site-packages/mdp/nodes/pca_nodes.py", line 32, in set_output_dim raise TrainingFinishedException, errstr NameError: global name 'TrainingFinishedException' is not defined Could someone tell me what I am doing wrong? By the way, does anybody know of an implementation of the partial least squares algorithm? Thanks for your help. Ramon From noel.oboyle2 at mail.dcu.ie Thu Sep 29 11:02:01 2005 From: noel.oboyle2 at mail.dcu.ie (Noel O'Boyle) Date: Thu, 29 Sep 2005 16:02:01 +0100 Subject: [SciPy-user] mdp question In-Reply-To: <433BFF14.6050200@iiqab.csic.es> References: <433BFF14.6050200@iiqab.csic.es> Message-ID: <1128006121.4194.10.camel@sandwi.ch.cam.ac.uk> For PLS, see pychem.sf.net. It's currently being converted to scipy, so you should check out the CVS or post to the pychem mailing list if you have any queries. Noel On Thu, 2005-09-29 at 16:49 +0200, Ramon Crehuet wrote: > Hi all, > I am using mdp to perform a PCA on some (dummy) variables. I can get the > principal components but I do not know how to get the "explained > variance" or change the number of output components. Here is my input: > > import mdp > from scipy import array > x=array([[1.51, -2.10, 1.91], > [-2.10, 0.45, -1.25], > [1.59, -0.51, 0.50], > [0.45, -0.93, 1.51], > [0.04, -0.49,-0.09], > [-0.78, 1.89,-1.55], > [-0.62,0.00, -0.42]]) > pca=mdp.nodes.PCANode() > pca.train(x) > pca.stop_training() > mat=pca.get_projmatrix() > # works fine. > x=pca.execute(x,1) > #works fine, too. > print pca.get_explained_variance() > # This gives "None" > pca.set_output_dim(1) > # and this raises an exception: > > Traceback (most recent call last): > File "/home/ramon/python/pca.py", line 21, in -toplevel- > pca.set_output_dim(1) > File "/usr/lib/python2.3/site-packages/mdp/nodes/pca_nodes.py", line > 32, in set_output_dim > raise TrainingFinishedException, errstr > NameError: global name 'TrainingFinishedException' is not defined > > > Could someone tell me what I am doing wrong? > By the way, does anybody know of an implementation of the partial least > squares algorithm? > Thanks for your help. > > Ramon > > > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user From stephen.walton at csun.edu Thu Sep 29 13:24:33 2005 From: stephen.walton at csun.edu (Stephen Walton) Date: Thu, 29 Sep 2005 10:24:33 -0700 Subject: [SciPy-user] problems installing scipy In-Reply-To: <200510100445.46154.john.russo@poste.it> References: <200510100445.46154.john.russo@poste.it> Message-ID: <433C2351.50204@csun.edu> john russo wrote: >export ~/src/blas/libfblas.a >export ~/src/LAPACK/lapack_LINUX_IFC.a >export ATLAS=None > > > Your first two lines should be export BLAS=~/src/blas/libfblas.a export LAPACK=~/src/LAPACK/lapack_LINUX_IFC.a From ckkart at hoc.net Thu Sep 29 13:46:17 2005 From: ckkart at hoc.net (Christian Kristukat) Date: Thu, 29 Sep 2005 19:46:17 +0200 Subject: [SciPy-user] Re: io.read_array slow In-Reply-To: References: <433A8C19.8010708@hoc.net> Message-ID: <433C2869.2070502@hoc.net> Arnd Baecker wrote: > On Wed, 28 Sep 2005, Christian Kristukat wrote: > > >>Jordan H. Mantha wrote: >> >>>Christian Kristukat wrote: >>> >>> >>>>Hi, >>>>I noticed that io.read_array is really slow compared to a python >>>>while/readline/split loop (about 5 times slower). I believe to remember >>>>that at some time it was written in C but when I looked at the source of >>>>scipy 0.3.2 it seems to be pure python. Is there a evident reason why it >>>>must be that slow or are there maybe plans to rewrite it in C? >>>>Regards, >>>>Christian >>> >>>I have noticed this too. I was wrote a fitting script that reads in a 3 >>>column data file that has 502,000 rows. I first started with io.read_array >>>and it took > 10 min. to get the array in. I was reading "Python Scripting >>>for Computational Science" by Hans Petter Langtangen and used the >>>following: >>> >>> data = array([float(x) for x in infile.read().split()], Float) >>> data.shape = (len(data)/3,3) >>> >> >>Nice short solution indeed. I think read_array could be cut down to something >>less versatile such that it can handle different separators and works at C speed. > > > You could also have a look at TableIO, > http://php.iupui.edu/~mmiller3/python/ That's fast! So we really should have some C coded IO module. TableIO is GPLed, so it's not possible to include it directly according to what I've learned here in the last weeks. Maybe I'll write a replacement some day. Regards, Christian From ckkart at hoc.net Thu Sep 29 13:48:48 2005 From: ckkart at hoc.net (Christian Kristukat) Date: Thu, 29 Sep 2005 19:48:48 +0200 Subject: [SciPy-user] 3d plotting question In-Reply-To: <17211.63154.449495.387875@monster.linux.in> References: <433B11D5.3040404@comcast.net> <17211.29654.146184.626086@monster.linux.in> <433BC407.1050609@hoc.net> <17211.59661.320358.1243@monster.linux.in> <433BED85.7040403@hoc.net> <17211.63154.449495.387875@monster.linux.in> Message-ID: <433C2900.80703@hoc.net> Prabhu Ramachandran wrote: >>>>>>"Christian" == Christian Kristukat writes: > > > Christian> Prabhu Ramachandran wrote: > >> If that works, you are ready to play with mlab.py which is > >> inside enthought/tvtk/tools/mlab.py. You can run the file as > >> it is, to see a spherical harmonic on screen. See the source > >> for some nice examples. > >> > >> If you have come this far, I hope you are rewarded by all the > >> new tools you can play with. :) > > Christian> Wow!!! That was easy. I'm very impressed. Only the ivtk > Christian> example isn't working. I'm using vtk python 4.2.6. The > Christian> traceback follows: > > Terrific! VTK-4.4 is worth upgrading to if possible. In any case, I > forgot that you can also run all the tests in tvtk/tests to make sure > the core functionality works. I do the following: I upgraded to 4.4.2 and now simple.py is no longer working. Is that a TK issue? Traceback (most recent call last): File "simple.py", line 45, in ? bw = tvtk.BoxWidget(interactor=rwi, place_factor=1.25, AttributeError: 'TVTK' object has no attribute 'BoxWidget' And the ivtk example now results in a X error: ** (python:24725): WARNING **: Cannot open font file for font Adobe Helvetica 12 ** (python:24725): WARNING **: Cannot open font file for font Adobe Helvetica 14The program 'python' received an X Window System error. This probably reflects a bug in the program. The error was 'BadWindow (invalid Window parameter)'. (Details: serial 7 error_code 3 request_code 2 minor_code 0) (Note to programmers: normally, X errors are reported asynchronously; that is, you will receive the error a while after causing it. To debug your program, run it with the --sync command line option to change this behavior. You can then get a meaningful backtrace from your debugger if you break on the gdk_x_error() function.) > $ cd tvtk/tests > $ for i in *.py; do echo $i; python $i; done > Most of them work. I've got no numarray installed but I doubt that it's necessary. test_array_ext_na.py Traceback (most recent call last): File "test_array_ext_na.py", line 10, in ? import numarray ImportError: No module named numarray test_array_ext.py .. ---------------------------------------------------------------------- Ran 2 tests in 0.096s OK test_array_handler.py ...... ---------------------------------------------------------------------- Ran 6 tests in 0.574s OK test_class_tree.py ...... ---------------------------------------------------------------------- Ran 6 tests in 0.127s OK test_indenter.py ......... ---------------------------------------------------------------------- Ran 9 tests in 0.019s OK test_messenger.py ..... ---------------------------------------------------------------------- Ran 5 tests in 0.005s OK test_tvtk_base.py ........... ---------------------------------------------------------------------- Ran 11 tests in 0.055s OK test_tvtk.py ....F................ ====================================================================== FAIL: Test the caching of ancestor classes. ---------------------------------------------------------------------- Traceback (most recent call last): File "test_tvtk.py", line 80, in test_cache self.assertEqual(tvtk_helper._cache.has_key(i), True) File "/usr/lib/python2.3/unittest.py", line 302, in failUnlessEqual raise self.failureException, \ AssertionError: False != True ---------------------------------------------------------------------- Ran 21 tests in 14.962s FAILED (failures=1) test_vtk_parser.py ..... ---------------------------------------------------------------------- Ran 5 tests in 4.475s OK test_wrapper_gen.py .. ---------------------------------------------------------------------- Ran 2 tests in 0.005s OK > [...] > Christian> "/usr/lib/python2.3/site-packages/vtk_python/vtk/wx/wxVTKRenderWindowInteractor.py", > Christian> line 339, in Render > Christian> self._Iren.GetRenderWindow().SetNextWindowInfo(str(self.GetHandle())) > > If you are willing to do a little bit of surgery, just replace that > file (wxVTKRenderWindowInteractor.py), with the latest from here: > > http://public.kitware.com/cgi-bin/viewcvs.cgi/Wrapping/Python/vtk/wx/wxVTKRenderWindowInteractor.py?rev=1.17&view=log That did not work neither. He's complaining about different VTK versions. > Then it should hopefully work provided you have wxPython-2.6.x. Yes, I've wxPyhton 2.6.1. Regards, Christian From prabhu_r at users.sf.net Thu Sep 29 14:17:00 2005 From: prabhu_r at users.sf.net (Prabhu Ramachandran) Date: Thu, 29 Sep 2005 23:47:00 +0530 Subject: [SciPy-user] 3d plotting question In-Reply-To: <433C2900.80703@hoc.net> References: <433B11D5.3040404@comcast.net> <17211.29654.146184.626086@monster.linux.in> <433BC407.1050609@hoc.net> <17211.59661.320358.1243@monster.linux.in> <433BED85.7040403@hoc.net> <17211.63154.449495.387875@monster.linux.in> <433C2900.80703@hoc.net> Message-ID: <17212.12188.606470.822959@monster.linux.in> >>>>> "Christian" == Christian Kristukat writes: >> Terrific! VTK-4.4 is worth upgrading to if possible. In any >> case, I forgot that you can also run all the tests in >> tvtk/tests to make sure the core functionality works. I do the >> following: Christian> I upgraded to 4.4.2 and now simple.py is no longer Christian> working. Is that a TK issue? [...] Did you rebuild TVTK after upgrading to 4.4.2? If you did not, then remove the tvtk_classes.zip file and rebuild TVTK like so: python setup.py build_ext --inplace You should have gotten warning messages if you did not do this and imported TVTK. Also, did you build the Hybrid classes when you built 4.4.2? >> $ cd tvtk/tests $ for i in *.py; do echo $i; python $i; done Christian> Most of them work. I've got no numarray installed but I Christian> doubt that it's necessary. Yes, that is unnecessary. So that can be ignored. The ancestor caching should I think go away when you rebuild TVTK. >> If you are willing to do a little bit of surgery, just replace >> that file (wxVTKRenderWindowInteractor.py), with the latest >> from here: >> >> http://public.kitware.com/cgi-bin/viewcvs.cgi/Wrapping/Python/vtk/wx/wxVTKRenderWindowInteractor.py?rev=1.17&view=log Christian> That did not work neither. He's complaining about Christian> different VTK versions. Sorry, who is complaining about different versions? Thanks for the patient testing. cheers, prabhu From ckkart at hoc.net Thu Sep 29 14:36:15 2005 From: ckkart at hoc.net (Christian Kristukat) Date: Thu, 29 Sep 2005 20:36:15 +0200 Subject: [SciPy-user] 3d plotting question In-Reply-To: <17212.12188.606470.822959@monster.linux.in> References: <433B11D5.3040404@comcast.net> <17211.29654.146184.626086@monster.linux.in> <433BC407.1050609@hoc.net> <17211.59661.320358.1243@monster.linux.in> <433BED85.7040403@hoc.net> <17211.63154.449495.387875@monster.linux.in> <433C2900.80703@hoc.net> <17212.12188.606470.822959@monster.linux.in> Message-ID: <433C341F.3080501@hoc.net> Hi Prabhu, I'll switch to the mayavi mailing list, ok? Prabhu Ramachandran wrote: >>>>>>"Christian" == Christian Kristukat writes: > > > >> Terrific! VTK-4.4 is worth upgrading to if possible. In any > >> case, I forgot that you can also run all the tests in > >> tvtk/tests to make sure the core functionality works. I do the > >> following: > > Christian> I upgraded to 4.4.2 and now simple.py is no longer > Christian> working. Is that a TK issue? > [...] > > Did you rebuild TVTK after upgrading to 4.4.2? If you did not, then > remove the tvtk_classes.zip file and rebuild TVTK like so: Yes, I've seen the warning message and rebuilt tvtk. After that, I got the error about the missing BoxWidget. > > >> If you are willing to do a little bit of surgery, just replace > >> that file (wxVTKRenderWindowInteractor.py), with the latest > >> from here: > >> > >> http://public.kitware.com/cgi-bin/viewcvs.cgi/Wrapping/Python/vtk/wx/wxVTKRenderWindowInteractor.py?rev=1.17&view=log > > Christian> That did not work neither. He's complaining about > Christian> different VTK versions. > > Sorry, who is complaining about different versions? Well, I can't reproduce this now as I already upgraded to vtk 4.4.2 but when starting the ivtk example there was a message saying, if I remember right, that there is a version mismatch between some image lib (libpng?) of the system and that with which VTK was compiled...? Something like that. > > Thanks for the patient testing. No problem, I'm just curious about the results. Regards, Christian From stephen.walton at csun.edu Thu Sep 29 14:42:25 2005 From: stephen.walton at csun.edu (Stephen Walton) Date: Thu, 29 Sep 2005 11:42:25 -0700 Subject: [SciPy-user] Gamma distribution questions In-Reply-To: <4320B63D.1070401@ucsd.edu> References: <4320B63D.1070401@ucsd.edu> Message-ID: <433C3591.8060304@csun.edu> Robert Kern wrote, way back on September 8: >Currently, the fit() method of distributions is broken. There's a >problem with the way it passes arguments to the nnlf() method; that >problem seems to apply to all distributions. > I actually fixed this, kind of, some time ago. See http://www.scipy.net/roundup/scipy/issue230 With this fix, I can do the following: import scipy as S x=S.stats.lognorm.rvs(0.75,loc=5,scale=3,size=(500,)) (shape,ppcc)=S.stats.ppcc_plot(x,0.1,4,dist='lognorm') plot(shape,ppcc) # Notice nice local maximum near shape=0.75 (osm,osr),(scale,loc,r)=S.stats.probplot(x,sparams=0.75,dist='lognorm',fit=1) # Returns scale and loc which are tolerably close to 3 and 5, respectively. plot(osm, osr) # Notice this is fairly linear However, S.stats.ppcc_max(x,(0.1,4.),dist='lognorm') seems to return nonsense with this test data (-5 or -6), although it produces the correct value with some real data I have. >There is also a problem >with distributions like Gamma which are intrinsically positive; all of >the distribution objects take a loc parameter. For intrinsically >positive variates like Gamma, this really should be fixed to 0 all of >the time. > > I think I disagree. I have sunspot group area data which are more or less lognormally (or perhaps weibull_min) distributed but which definitely have a nonzero loc parameter. When will this parameter fitting capability be in "new scipy" and in what form? What can I do to help? Steve Walton P.S. Does anyone read the Bug Tracker entries at scipy.org? My morestats.py fix is over three months old. From oliphant at ee.byu.edu Thu Sep 29 14:56:52 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Thu, 29 Sep 2005 12:56:52 -0600 Subject: [SciPy-user] Gamma distribution questions In-Reply-To: <433C3591.8060304@csun.edu> References: <4320B63D.1070401@ucsd.edu> <433C3591.8060304@csun.edu> Message-ID: <433C38F4.9080408@ee.byu.edu> Stephen Walton wrote: > Robert Kern wrote, way back on September 8: > >> Currently, the fit() method of distributions is broken. There's a >> problem with the way it passes arguments to the nnlf() method; that >> problem seems to apply to all distributions. >> > I actually fixed this, kind of, some time ago. See > > > P.S. Does anyone read the Bug Tracker entries at scipy.org? My > morestats.py fix is over three months old. I don't read it regularly -- I read email much more. So make sure to post an announcement to this list. Also, I have been pre-occupied with the new core system. Also, the move to svn put a big delay on placing any fixes into scipy as several people lost write access. So, don't be discouraged. There have been some growing pains this summer. We also need more people who are willing to have write access to the actual source tree. Any takers? (I won't just let anybody make changes, but if you've already contributed to SciPy in some fashion, you are a likely candidate...) -Travis O. From rclewley at cam.cornell.edu Thu Sep 29 17:20:02 2005 From: rclewley at cam.cornell.edu (Robert Clewley) Date: Thu, 29 Sep 2005 17:20:02 -0400 (EDT) Subject: [SciPy-user] Problem controlling temp dir usage and verbose output for scipy_distutils vs. distutils Message-ID: Hi, I've just got scipy_distutils and SWIG to successfully build a C extension module that uses Fortran sources (after surprisingly little additional effort from the pure-C case -- nice job Pearu), but I'm having trouble with two things that I didn't have trouble with using Python's original distutils: 1) Controlling the stderr warnings and info from the compilers and 2) Controlling the temp directories that the build uses. The libraries I'm building are numerical ODE integrators which are selected in my software by the user. I want to keep the filespace tidy and not require any additional admin by the user. In the past I've been building pure C extensions and provided arguments to my (regular Python distutils) setup command like the following, which solved both of the above problems: setup(name = "integrator in pure c", author = "me", script_args = ['-q', 'build', '--build-lib=.', '--build-platlib=.', '-t'+tempdirname], ext_modules = [Extension("modC", sources=modfilelist, include_dirs=incdirs, extra_compile_args=['-w'])]) The script args placed all of the temporary files in the same place which is automatically chosen by my software according to the user's choice of integrator. The -w argument used to shut up the C compiler nicely! However, when I switched to scipy_distutils in order to access the Fortran compiler, these options no longer seem to have any effect. This is what the modules that use Fortran libraries are set up with: setup(name = "integrator with fortran lib", author = "me", script_args = ['-q', 'build', '--build-lib=.', '--build-platlib=.', '-t'+tempdirname], ext_modules = [Extension("modfort", sources=modfilelist, include_dirs=incdirs, libraries=['myfortlib'], extra_compile_args=['-w'])], libraries=[('myfortlib',{'sources': fortfilelist})]) I now end up with a deep tree of temp directories (such as I used to have in the pure C case before I discovered the --build-lib options) inside my desired compilation temporary directory, as well as a separate 'build' directory beside my temp compilation dir that holds my final .py interface file (in ./build/src/modfort_temp/). I looked at the API and I cannot see additional options that I could be passing. I can see a install_libbase attribute being inherited from Python's distutils/command/install.py module's install class into SciPy's overriding class, but I don't get why it's not being used in the same way. I'm having the same problems on both Win32 and Linux platform using Python 2.3.5 and SciPy 0.3.2. My problem is also not specific to using Fortran libraries, as my old pure-C module suffers problem (2) when I use scipy_distutils on it (although the compiler stays quiet). Any ideas please? Thanks, Rob Clewley From oliphant at ee.byu.edu Fri Sep 30 02:59:22 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Fri, 30 Sep 2005 00:59:22 -0600 Subject: [SciPy-user] Release of SciPy Core 0.4 (Beta) Message-ID: <433CE24A.6040509@ee.byu.edu> This is to announce the release of SciPy Core 0.4.X (beta) It is available at sourceforge which you can access by going to http://numeric.scipy.org Thanks to all who helped me with this release, especially Robert Kern Pearu Peterson Now, let's start getting this stable.... -Travis Oliphant From oliphant at ee.byu.edu Fri Sep 30 03:17:19 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Fri, 30 Sep 2005 01:17:19 -0600 Subject: [SciPy-user] Install of scipy core for current scipy users Message-ID: <433CE67F.3060907@ee.byu.edu> If you are a current user of scipy and want to try out the new scipy core. You can do so. However, the default install will change four __init__.py files on your system (which will probably "break" the rest of scipy). Hopefully we will get scipy as a whole to work with the new system shortly. There has also been talk of putting together a simple install to fixup the __init__ files so that old scipy and new scipy_core can live together peacefully (anyone who can help with that is welcomed). scipy\__init__.py scipy\linalg\__init__.py scipy\fftpack\__init__.py scipy\stats\__init__.py are the affected files. The ones from scipy can just be copied back, but then you won't get the functions from new scipy. -Travis From prabhu_r at users.sf.net Fri Sep 30 04:40:27 2005 From: prabhu_r at users.sf.net (Prabhu Ramachandran) Date: Fri, 30 Sep 2005 14:10:27 +0530 Subject: [SciPy-user] Release of SciPy Core 0.4 (Beta) In-Reply-To: <433CE24A.6040509@ee.byu.edu> References: <433CE24A.6040509@ee.byu.edu> Message-ID: <17212.63995.2169.219479@monster.linux.in> >>>>> "Travis" == Travis Oliphant writes: Travis> This is to announce the release of SciPy Core 0.4.X (beta) Travis> http://numeric.scipy.org Many thanks! I just read your presentation for SciPy05 and the new core looks very impressive. Thanks for all the hard work! regards, prabhu From meesters at uni-mainz.de Fri Sep 30 07:08:51 2005 From: meesters at uni-mainz.de (Christian Meesters) Date: Fri, 30 Sep 2005 13:08:51 +0200 Subject: [SciPy-user] Release of SciPy Core 0.4 (Beta) In-Reply-To: <17212.63995.2169.219479@monster.linux.in> References: <433CE24A.6040509@ee.byu.edu> <17212.63995.2169.219479@monster.linux.in> Message-ID: <08b0ad3bde9aedda81f752e937f45de3@uni-mainz.de> Prabhu is right: Great, that we can start using it now. Thanks a lot! Just two questions: - The presentation lets my PowerPoint always crash (OS X 10.3). Apparently it works elsewhere, but could you check whether it's broken, please? - And this is the reason why I'm writing to the whole list and not to Travis: I'd like to purchase the pre-version and think supporting SciPy is worth a few bucks. Just I'm sitting in Europe now and don't have a credit card (what for?). Is there an other way to purchase the book? (The sample looks really impressive.) Cheers Christian On 30 Sep 2005, at 10:40, Prabhu Ramachandran wrote: >>>>>> "Travis" == Travis Oliphant writes: > > Travis> This is to announce the release of SciPy Core 0.4.X (beta) > > Travis> http://numeric.scipy.org > > Many thanks! I just read your presentation for SciPy05 and the new > core looks very impressive. Thanks for all the hard work! > > regards, > prabhu > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user > From christophe.grimault at novagrid.com Fri Sep 30 11:48:53 2005 From: christophe.grimault at novagrid.com (christophe grimault) Date: Fri, 30 Sep 2005 15:48:53 +0000 Subject: [SciPy-user] Consequenses of SciPy Core 0.4 ? In-Reply-To: <08b0ad3bde9aedda81f752e937f45de3@uni-mainz.de> References: <433CE24A.6040509@ee.byu.edu> <17212.63995.2169.219479@monster.linux.in> <08b0ad3bde9aedda81f752e937f45de3@uni-mainz.de> Message-ID: <1128095333.5231.5.camel@pandora> Hi Travis, Does the new scipy core means that, at least on unix, we won't have to build and install Numeric anymore ? Regards, Christophe From nwagner at mecha.uni-stuttgart.de Fri Sep 30 10:27:06 2005 From: nwagner at mecha.uni-stuttgart.de (Nils Wagner) Date: Fri, 30 Sep 2005 16:27:06 +0200 Subject: [SciPy-user] Consequenses of SciPy Core 0.4 ? In-Reply-To: <1128095333.5231.5.camel@pandora> References: <433CE24A.6040509@ee.byu.edu> <17212.63995.2169.219479@monster.linux.in> <08b0ad3bde9aedda81f752e937f45de3@uni-mainz.de> <1128095333.5231.5.camel@pandora> Message-ID: <433D4B3A.7070803@mecha.uni-stuttgart.de> christophe grimault wrote: >Hi Travis, > >Does the new scipy core means that, at least on unix, we won't have to >build and install Numeric anymore ? > >Regards, >Christophe > >_______________________________________________ >SciPy-user mailing list >SciPy-user at scipy.net >http://www.scipy.net/mailman/listinfo/scipy-user > The cvs entry http://www.scipy.org/cvs/ is obsolete and should be replaced by some user friendly comments on svn. What is really needed to install scipy via svn ? Nils From oliphant at ee.byu.edu Fri Sep 30 10:32:32 2005 From: oliphant at ee.byu.edu (Travis Oliphant) Date: Fri, 30 Sep 2005 08:32:32 -0600 Subject: [SciPy-user] Consequenses of SciPy Core 0.4 ? In-Reply-To: <1128095333.5231.5.camel@pandora> References: <433CE24A.6040509@ee.byu.edu> <17212.63995.2169.219479@monster.linux.in> <08b0ad3bde9aedda81f752e937f45de3@uni-mainz.de> <1128095333.5231.5.camel@pandora> Message-ID: <433D4C80.9080003@ee.byu.edu> christophe grimault wrote: >Hi Travis, > >Does the new scipy core means that, at least on unix, we won't have to >build and install Numeric anymore ? > > > Right, scipy core replaces Numeric. Now, you still need to get scipy core built and installed. Which version of UNIX are you running? -Travis From zane-scipy at ideotrope.org Fri Sep 30 12:44:06 2005 From: zane-scipy at ideotrope.org (Zane Crawford) Date: Fri, 30 Sep 2005 09:44:06 -0700 Subject: [SciPy-user] Problems Installing SciPy on OS X 10.4.2 w/ python 2.3.5 Message-ID: <20050930164406.GA1688@ideotrope.org> With much frustration, and some patient help from Chris Fonnesbeck, I have been trying to get SciPy installed on my Mac. I am running OS X 10.4.2, and am attempting to use the stock python (2.3.5) which is installed by the Apple developer tools. I believe I have all of the underlying dependencies installed and functioning (Numeric 24.0b2, AquaTerm 1.0b2, fftw, GnuPlot 4.0.0, F2PY, and probably other things I installed along the way), and am now running into problems with the actual build of SciPy itself. I have just tried three different ways: 1) I installed a binary package which Chris created for me, built on python 2.3.5. It installs fine, and I can "import scipy" without issues, however, I can't use anything from it, because of the __cvs_version__ bug, which is referred to in the mailing list archives, but which was apparently not fixed in a persistent way. (it seems there was some clever code which allowed the software to keep track of its own version based on the CVS versions of the files, but this became broken when the source was migrated to SVN?) 2) I tried to build the last CVS version of SciPy, from July 29th, 2005. This fails almost immediately, regardless of whether I'm using GCC 3.3 or GCC 4.0, with an error from posixpath.py, line 77: i = p.rfind('/') + 1 AttributeError: 'NoneType' object has no attribute 'rfind' 3) I tried pulling the most recent version of the source from SVN. This seems to get much further than the CVS build, but fails eventually: In file included from Lib/xplt/src/play/all/hash0.c:8: build/temp.darwin-8.2.0-Power_Macintosh-2.3/config_pygist/config.h:3:3: #error destroy this config.h and rerun configure script However, there are many many warnings prior to this failure, and I don't know what the acutal problem is I've completely removed (as far as I can tell) all installed SciPy packages (from Chris' build) and think I have a clean 2.3.5 python environment ready to install SciPy. If anyone has any other sources they would recommend building from, I'm happy to try. My initial attempt to install SciPy (prior to this mess) was via Fink, which seemed to result in its own problems. -- Zane A. Crawford . + zane at ideotrope.org Amateur Human * , * https://ideotrope.org Boulder, Colorado __o work : 303.735.3729 _`\<,_ "Here we are, trapped in the cell : 303.815.6866 (*)/ (*) amber of the moment. There PGP : 0x55E0815F is no why." -- Kurt Vonnegut From managan at llnl.gov Fri Sep 30 14:21:45 2005 From: managan at llnl.gov (Rob Managan) Date: Fri, 30 Sep 2005 11:21:45 -0700 Subject: [SciPy-user] Release of SciPy Core 0.4 (Beta) In-Reply-To: <433CE24A.6040509@ee.byu.edu> References: <433CE24A.6040509@ee.byu.edu> Message-ID: OK, I am giving this a go. I seem to be not setting a flag about some math functions availability. Here are the details (aorry about the length of it but I snipped a lot!): OSX 10.3.9, Python 2.4.1 Installed g77v3.4-bin.tar.gz from http://hpc.sourceforge.net/ as described into /usr/local/... Installed fftw-3.0.1-fma.tar.gz from http://www.fftw.org/download.html into ~/Documents/local/... where I have other libraries I have installed. Got scipy_core-0.4.1.tar.gz from http://sourceforge.net/project/showfiles.php?group_id=1369&package_id=6222, wanted to try out the new replacement for Numeric tar -zxvf scipy_core-0.4.1.tar.gz cd scipy_core-0.4.1 python setup.py build Assuming default configuration (scipy/distutils/command/{setup_command,setup}.py was not found) Appending scipy.distutils.command configuration to scipy.distutils Assuming default configuration (scipy/distutils/fcompiler/{setup_fcompiler,setup}.py was not found) ... Appending scipy.base configuration to scipy blas_opt_info: FOUND: extra_link_args = ['-Wl,-framework', '-Wl,Accelerate'] define_macros = [('NO_ATLAS_INFO', 3)] extra_compile_args = ['-faltivec', '-I/System/Library/Frameworks/vecLib.framework/Headers'] lapack_opt_info: FOUND: extra_link_args = ['-Wl,-framework', '-Wl,Accelerate'] define_macros = [('NO_ATLAS_INFO', 3)] extra_compile_args = ['-faltivec'] Appending scipy.lib configuration to scipy Assuming default configuration (scipy/fftpack/{setup_fftpack,setup}.py was not found) ... scipy_core version 0.4.1 running build running config_fc running build_src building extension "scipy.base.multiarray" sources creating build creating build/src creating build/src/scipy creating build/src/scipy/base Generating build/src/scipy/base/config.h customize NAGFCompiler customize AbsoftFCompiler customize IbmFCompiler Could not locate executable f77 customize GnuFCompiler customize GnuFCompiler ... compiling C sources gcc options: '-fno-strict-aliasing -Wno-long-double -no-cpp-precomp -mno-fused-madd -fno-common -dynamic -DNDEBUG -g -O3 -Wall -Wstrict-prototypes' creating build/temp.darwin-7.9.0-Power_Macintosh-2.4/build creating build/temp.darwin-7.9.0-Power_Macintosh-2.4/build/src creating build/temp.darwin-7.9.0-Power_Macintosh-2.4/build/src/scipy creating build/temp.darwin-7.9.0-Power_Macintosh-2.4/build/src/scipy/base creating build/temp.darwin-7.9.0-Power_Macintosh-2.4/build/src/scipy/base/src compile options: '-Ibuild/src/scipy/base/src -Iscipy/base/include -Ibuild/src/scipy/base -Iscipy/base/src -I/Library/Frameworks/Python.framework/Versions/2.4/include/python2.4 -c' gcc: build/src/scipy/base/src/umathmodule.c In file included from build/src/scipy/base/src/umathmodule.c:8: scipy/base/include/scipy/arrayobject.h:84: warning: redefinition of `ushort' /usr/include/sys/types.h:82: warning: `ushort' previously declared here scipy/base/include/scipy/arrayobject.h:85: warning: redefinition of `uint' /usr/include/sys/types.h:83: warning: `uint' previously declared here build/src/scipy/base/src/umathmodule.c: In function `nc_floor_quotl': build/src/scipy/base/src/umathmodule.c:1257: warning: implicit declaration of function `floorl' build/src/scipy/base/src/umathmodule.c: In function `nc_sqrtl': build/src/scipy/base/src/umathmodule.c:1269: warning: implicit declaration of function `sqrtl' ... build/src/scipy/base/__umath_generated.c:171: error: initializer element is not constant build/src/scipy/base/__umath_generated.c:171: error: (near initialization for `arccosh_data[2]') error: Command "gcc -fno-strict-aliasing -Wno-long-double -no-cpp-precomp -mno-fused-madd -fno-common -dy namic -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -Ibuild/src/scipy/base/src -Iscipy/base/include -Ibuild/s rc/scipy/base -Iscipy/base/src -I/Library/Frameworks/Python.framework/Versions/2.4/include/python2.4 -c b uild/src/scipy/base/src/umathmodule.c -o build/temp.darwin-7.9.0-Power_Macintosh-2.4/build/src/scipy/base /src/umathmodule.o" failed with exit status 1 At 12:59 AM -0600 9/30/05, Travis Oliphant wrote: >This is to announce the release of SciPy Core 0.4.X (beta) > >It is available at sourceforge which you can access by going to > >http://numeric.scipy.org > >Thanks to all who helped me with this release, especially > >Robert Kern >Pearu Peterson > >Now, let's start getting this stable.... > >-Travis Oliphant > > > >_______________________________________________ >SciPy-user mailing list >SciPy-user at scipy.net >http://www.scipy.net/mailman/listinfo/scipy-user -- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- Rob Managan email managan at llnl.gov LLNL phone: 925-423-0903 P.O. Box 808, L-095 FAX: 925-422-3389 Livermore, CA 94551-0808 From bgranger at scu.edu Fri Sep 30 14:36:18 2005 From: bgranger at scu.edu (Brian Granger) Date: Fri, 30 Sep 2005 11:36:18 -0700 Subject: [SciPy-user] Release of SciPy Core 0.4 (Beta) Message-ID: Which version of gccc are you using? As I understand it, g77 3.4 is not compatible with gcc 4.0 (the default on OS X). Switching to gcc 3.4 might help: gcc_select 3.3 I will soon try scipy_core on Tiger with Python 2.4.1. Brian Brian Granger, Ph.D. Assistant Professor of Physics Santa Clara University bgranger at scu.edu Phone: 408-551-1891 Fax: 408-554-6965 >>> managan at llnl.gov 09/30/05 11:21 AM >>> OK, I am giving this a go. I seem to be not setting a flag about some math functions availability. Here are the details (aorry about the length of it but I snipped a lot!): OSX 10.3.9, Python 2.4.1 Installed g77v3.4-bin.tar.gz from http://hpc.sourceforge.net/ as described into /usr/local/... Installed fftw-3.0.1-fma.tar.gz from http://www.fftw.org/download.html into ~/Documents/local/... where I have other libraries I have installed. Got scipy_core-0.4.1.tar.gz from http://sourceforge.net/project/showfiles.php?group_id=1369&package_id=6222, wanted to try out the new replacement for Numeric tar -zxvf scipy_core-0.4.1.tar.gz cd scipy_core-0.4.1 python setup.py build Assuming default configuration (scipy/distutils/command/{setup_command,setup}.py was not found) Appending scipy.distutils.command configuration to scipy.distutils Assuming default configuration (scipy/distutils/fcompiler/{setup_fcompiler,setup}.py was not found) ... Appending scipy.base configuration to scipy blas_opt_info: FOUND: extra_link_args = ['-Wl,-framework', '-Wl,Accelerate'] define_macros = [('NO_ATLAS_INFO', 3)] extra_compile_args = ['-faltivec', '-I/System/Library/Frameworks/vecLib.framework/Headers'] lapack_opt_info: FOUND: extra_link_args = ['-Wl,-framework', '-Wl,Accelerate'] define_macros = [('NO_ATLAS_INFO', 3)] extra_compile_args = ['-faltivec'] Appending scipy.lib configuration to scipy Assuming default configuration (scipy/fftpack/{setup_fftpack,setup}.py was not found) ... scipy_core version 0.4.1 running build running config_fc running build_src building extension "scipy.base.multiarray" sources creating build creating build/src creating build/src/scipy creating build/src/scipy/base Generating build/src/scipy/base/config.h customize NAGFCompiler customize AbsoftFCompiler customize IbmFCompiler Could not locate executable f77 customize GnuFCompiler customize GnuFCompiler ... compiling C sources gcc options: '-fno-strict-aliasing -Wno-long-double -no-cpp-precomp -mno-fused-madd -fno-common -dynamic -DNDEBUG -g -O3 -Wall -Wstrict-prototypes' creating build/temp.darwin-7.9.0-Power_Macintosh-2.4/build creating build/temp.darwin-7.9.0-Power_Macintosh-2.4/build/src creating build/temp.darwin-7.9.0-Power_Macintosh-2.4/build/src/scipy creating build/temp.darwin-7.9.0-Power_Macintosh-2.4/build/src/scipy/base creating build/temp.darwin-7.9.0-Power_Macintosh-2.4/build/src/scipy/base/src compile options: '-Ibuild/src/scipy/base/src -Iscipy/base/include -Ibuild/src/scipy/base -Iscipy/base/src -I/Library/Frameworks/Python.framework/Versions/2.4/include/python2.4 -c' gcc: build/src/scipy/base/src/umathmodule.c In file included from build/src/scipy/base/src/umathmodule.c:8: scipy/base/include/scipy/arrayobject.h:84: warning: redefinition of `ushort' /usr/include/sys/types.h:82: warning: `ushort' previously declared here scipy/base/include/scipy/arrayobject.h:85: warning: redefinition of `uint' /usr/include/sys/types.h:83: warning: `uint' previously declared here build/src/scipy/base/src/umathmodule.c: In function `nc_floor_quotl': build/src/scipy/base/src/umathmodule.c:1257: warning: implicit declaration of function `floorl' build/src/scipy/base/src/umathmodule.c: In function `nc_sqrtl': build/src/scipy/base/src/umathmodule.c:1269: warning: implicit declaration of function `sqrtl' ... build/src/scipy/base/__umath_generated.c:171: error: initializer element is not constant build/src/scipy/base/__umath_generated.c:171: error: (near initialization for `arccosh_data[2]') error: Command "gcc -fno-strict-aliasing -Wno-long-double -no-cpp-precomp -mno-fused-madd -fno-common -dy namic -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -Ibuild/src/scipy/base/src -Iscipy/base/include -Ibuild/s rc/scipy/base -Iscipy/base/src -I/Library/Frameworks/Python.framework/Versions/2.4/include/python2.4 -c b uild/src/scipy/base/src/umathmodule.c -o build/temp.darwin-7.9.0-Power_Macintosh-2.4/build/src/scipy/base /src/umathmodule.o" failed with exit status 1 At 12:59 AM -0600 9/30/05, Travis Oliphant wrote: >This is to announce the release of SciPy Core 0.4.X (beta) > >It is available at sourceforge which you can access by going to > >http://numeric.scipy.org > >Thanks to all who helped me with this release, especially > >Robert Kern >Pearu Peterson > >Now, let's start getting this stable.... > >-Travis Oliphant > > > >_______________________________________________ >SciPy-user mailing list >SciPy-user at scipy.net >http://www.scipy.net/mailman/listinfo/scipy-user -- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- Rob Managan email managan at llnl.gov LLNL phone: 925-423-0903 P.O. Box 808, L-095 FAX: 925-422-3389 Livermore, CA 94551-0808 _______________________________________________ SciPy-user mailing list SciPy-user at scipy.net http://www.scipy.net/mailman/listinfo/scipy-user This message scanned for viruses and SPAM at SCU (MGW2)