Mailman 3 March 2006 - SciPy-Dev

iterative.py
by Nils Wagner March 17, 2006

March 17, 2006

Hi all, Just curious The first lines in iterative.py reads ## Automatically adapted for scipy Oct 18, 2005 by # Iterative methods using reverse-communication raw material # These methods solve # Ax = b for x # where A must have A.matvec(x,*args) defined # or be a numeric array Why is A restricted to A.matvec(x,*args) and numeric arrays ? I think a matrix input should be possible as well ;-) Nils

1 0

Re: [SciPy-dev] Computing correlations with SciPy
by John Hunter March 17, 2006

March 17, 2006

The following message is a courtesy copy of an article that has been posted to comp.lang.python as well. >>>>> "tkpmep" == tkpmep <tkpmep(a)hotmail.com> writes: tkpmep> I want to compute the correlation between two sequences X tkpmep> and Y, and tried using SciPy to do so without success.l tkpmep> Here's what I have, how can I correct it? >>>> X = [1, 2, 3, 4, 5] Y = [5, 4, 3, 2, 1] import scipy >>>> scipy.corrcoef(X,Y) tkpmep> Traceback (most recent call last): File "<interactive tkpmep> input>", line 1, in ? File tkpmep> "C:\Python24\Lib\site-packages\numpy\lib\function_base.py", tkpmep> line 671, in corrcoef d = diag(c) File tkpmep> "C:\Python24\Lib\site-packages\numpy\lib\twodim_base.py", tkpmep> line 80, in diag raise ValueError, "Input must be 1- or tkpmep> 2-d." ValueError: Input must be 1- or 2-d. >>>> Hmm, this may be a bug in scipy. matplotlib also defines a corrcoef function, which you may want to use until this problem gets sorted out In [9]: matplotlib.mlab.corrcoef(X,Y) In [10]: X = [1, 2, 3, 4, 5] In [11]: Y = [5, 4, 3, 2, 1] In [12]: matplotlib.mlab.corrcoef(X,Y) Out[12]: array([[ 1., -1.], [-1., 1.]])

2 1

linalg naming
by Jonathan Taylor March 16, 2006

March 16, 2006

in the latest svn, things like numpy.linalg.inverse , numpy.linalg.generalized_inverse, etc. have disappeared. i remember there being an email about this in the past few days. are the old names supposed to still work? i see there is an old.py module, i suppose the easy fix is just to import old as needed. only this sort of breaks things like matplotlib that tried to be compatible with Numeric and numarray.... is it too much to add from old import * to __init__.py? -- jonathan p.s. by the way, as this is a numpy issue, i don't know if this is the right list to send this to -- is there a numpy-dev? i didn't try too hard, but i couldn't find one. -- ------------------------------------------------------------------------ I'm part of the Team in Training: please support our efforts for the Leukemia and Lymphoma Society! http://www.active.com/donate/tntsvmb/tntsvmbJTaylor GO TEAM !!! ------------------------------------------------------------------------ Jonathan Taylor Tel: 650.723.9230 Dept. of Statistics Fax: 650.725.8977 Sequoia Hall, 137 www-stat.stanford.edu/~jtaylo 390 Serra Mall Stanford, CA 94305

4 5

Re: [SciPy-dev] [Fwd: Build bots?]
by bryce hendrix March 16, 2006

March 16, 2006

Eric, At Enthought we use CruiseControl (http://cruisecontrol.sf.net) for all of our internal builds as well as our open source projects, such as enthought, chaco, traits, kiva, etc. Just this week I added numpy and scipy to our builds. Right now I've only got it building in Windows, but I will have it building on Redhat 3.0 in a couple of days. If you're not familiar with CruiseControl, its an build system allows for continuous and scheduled builds with support for several SCM systems. Its a build system with a lot of features, but at the same time its kind of ill suited for building python packages. The problem with using CruiseControl is that its targetted at continuous builds and detecting build breaks, which isn't all that useful for python projects. We've taken it a bit further and use CruiseControl to do continuous builds with unit & coverage testing, plus scheduled builds which build installers, generate html and other tasks. So while CruiseControl is ill suited for bulding python packages, it is well suited for automating everything else we want to do. If anyone is interested in contributing to this effort, I welcome the help. Or if anyone would like to hear about all the nifty customizations we made, I can tell you about those too. Bryce > > Does anyone know if there are or have been any build bots for scipy > and similar pieces of software? Would there be any interest if I were > to look into them? > > It seems that a lot of the questions on the various scipy-related > lists are "I can't seem to build version foo on distro bar" or "the > latest svn check-in broke foo". > > I have an extra machine at home and some experience with xen (a linux > VMware-like virtual machine system), and so would be interested in > helping set up such a system. I was envisioning testing the following > configurations: > > projects: > numpy > scipy > pytables > ipython > matplotlib > > across: > > debian sarge > debian testing > fedora core 4 > redhat enterprise linux 4 WS > ubuntu (breezy badger) > > My experience with other linux distros (and the *bsds) (and the > sometimes-forgotten other architectures, like ppc), as well as > windows, is rather limited, but I'll happily take suggestions for > additional things to add. > > Ideally we'd produce a set of scripts that end users could eventually > run on their own machines, if they wish, which would then upload the > results to the main build test machine, or maybe http-post to a page > on the scipy wiki. > > Every night, we'd checkout the latest svn from the projects, and try a > build across the architectures; we'd output that info to a centralized > location and then stick it on the web. > > I'm sure I'm not the first person to think of this, and maybe not the > first to try, so I'll take all the advice you want to throw at > me. Eventually, it would be nice to have this set up to automatically > run the unit tests as well, but right now that seems a bit > ambitious. Maybe this summer... :) > > > ...Eric > > _______________________________________________ > Scipy-dev mailing list > Scipy-dev(a)scipy.net > http://www.scipy.net/mailman/listinfo/scipy-dev > >

5 5

bench_random errors
by Arnd Baecker March 16, 2006

March 16, 2006

Hi, with a recent svn one gets the following errors ====================================================================== ERROR: bench_random (scipy.linalg.tests.test_decomp.test_eigvals) ---------------------------------------------------------------------- Traceback (most recent call last): File "/home/abaecker/BUILDS3/BuildDir/inst_numpy/lib/python2.4/site-packages/scipy/linalg/tests/test_decomp.py", line 57, in bench_random Numeric_eigvals = linalg.eigenvalues AttributeError: 'module' object has no attribute 'eigenvalues' ====================================================================== ERROR: bench_random (scipy.linalg.tests.test_basic.test_solve) ---------------------------------------------------------------------- Traceback (most recent call last): File "/home/abaecker/BUILDS3/BuildDir/inst_numpy/lib/python2.4/site-packages/scipy/linalg/tests/test_basic.py", line 160, in bench_random basic_solve = linalg.solve_linear_equations AttributeError: 'module' object has no attribute 'solve_linear_equations' ====================================================================== FAIL: check_expon (scipy.stats.tests.test_morestats.test_anderson) ---------------------------------------------------------------------- Traceback (most recent call last): File "/home/abaecker/BUILDS3/BuildDir/inst_numpy/lib/python2.4/site-packages/scipy/stats/tests/test_morestats.py", line 57, in check_expon assert_array_less(A, crit[-2:]) File "/home/abaecker/BUILDS3/BuildDir/inst_numpy/lib/python2.4/site-packages/numpy/testing/utils.py", line 255, in assert_array_less assert cond,\ AssertionError: Arrays are not less-ordered (mismatch 50.0%): Array 1: 1.6956285208609501 Array 2: [ 1.587 1.9339999999999999] (I know, the last one is the usual statistics one which goes away when running a second time - with almost certainty ;-) The solution to the first ones seems trivial, patch attached. Best, Jan and Arnd

1 0

Re: [SciPy-dev] Build bots?
by Joe Harrington March 16, 2006

March 16, 2006

Eric, this is a great concept! Getting pre-packaged native binary installs (RPMS, .debs, and so on) for all architectures is high on the list of usability tasks. Being able to make sure that the software builds on all architectures is part of what's holding that up. I'm not up on how the packaging is currently done (other than the magic word "distutils"), but you've put your finger on a major area for improvement and I support your idea. Can your buildbot be configured to produce native binary installs, rather than just tarballs? Please put your contact info and a brief description of your project under PACKAGING on http://scipy.org/Developer_Zone, and link to any other pages you make from there. These might simply point to the trac wiki if you choose to work there, but listing under Developer Zone will get you seen by new folks who might be interested in helping, and will help us track the packaging effort. Several folks have already signed up in the Developer Zone as willing to help on packaging. Maybe one of them can help get the unit tests into your system, and another can get it to make binary installers if it doesn't already. --jh--

1 0

matplotlib and backend_qt
by Nicolas Dumoulin March 16, 2006

March 16, 2006

Hi, I'm new on this list. I've discored matplotlib that allow to plot datas, and I'm particularely interested by the QT backend, for integrating plotting unctionnality in my scripts. I've followed this tuto : http://www.scipy.org/Wiki/Cookbook/Matplotlib/Qt_with_IPython_and_Designer It is suitable for using in ipython but, I would too make standalone scripts. So I've looked at the example examples/embedding_in_qt.py in the matplotlib archive, and read at the beginning : ----------- [...] # The QApplication has to be created before backend_qt is imported, otherwise # it will create one itself. # Note: color-intensive applications may require a different color allocation # strategy. QApplication.setColorSpec(QApplication.NormalColor) app = QApplication(sys.argv) from matplotlib.backends.backend_qtagg import FigureCanvasQTAgg as FigureCanvas [...] ------------ It's annoying because I would make independant GUI compounds, and build my QApplication in the main script. To achieve this, I've hacked matplotlib/backends/backend_qt.py to comment 4 last lines that creates the (so annoying in my case) qt.QApplication, and it works fine. I suppose that this QApplication is created to simplify the use of qt backend in ipython, doesn't it ? My main questions are : * Couldn't be possible to modify this behaviour, for example by adding separated modules that enables QApplication creation for use in ipython ? * Have I missed something ? If you aggree my reflexions, I could add a page to the Wiki/Cookbook to describe how I do ... Thanks -- Nicolas Dumoulin (french) http://bobuse.webhop.net http://eucd.info : sauvons le droit d'auteur !

1 1

Build bots?
by Eric Jonas March 16, 2006

March 16, 2006

Does anyone know if there are or have been any build bots for scipy and similar pieces of software? Would there be any interest if I were to look into them? It seems that a lot of the questions on the various scipy-related lists are "I can't seem to build version foo on distro bar" or "the latest svn check-in broke foo". I have an extra machine at home and some experience with xen (a linux VMware-like virtual machine system), and so would be interested in helping set up such a system. I was envisioning testing the following configurations: projects: numpy scipy pytables ipython matplotlib across: debian sarge debian testing fedora core 4 redhat enterprise linux 4 WS ubuntu (breezy badger) My experience with other linux distros (and the *bsds) (and the sometimes-forgotten other architectures, like ppc), as well as windows, is rather limited, but I'll happily take suggestions for additional things to add. Ideally we'd produce a set of scripts that end users could eventually run on their own machines, if they wish, which would then upload the results to the main build test machine, or maybe http-post to a page on the scipy wiki. Every night, we'd checkout the latest svn from the projects, and try a build across the architectures; we'd output that info to a centralized location and then stick it on the web. I'm sure I'm not the first person to think of this, and maybe not the first to try, so I'll take all the advice you want to throw at me. Eventually, it would be nice to have this set up to automatically run the unit tests as well, but right now that seems a bit ambitious. Maybe this summer... :) ...Eric

2 2

Module level docstrings
by Tim Leslie March 16, 2006

March 16, 2006

Looking through the code for both numpy and scipy, there are lots of cases where modules have header comments of the form: # # This is the foobar module for doing stuff. <insert useful information about foobar here> # Rather than having these blocks as comments, would it be worth changing them to be docstrings so they could be parsed by tools like pydoc and doxygen? Cheers, Timl

2 1

Re: [SciPy-dev] import pylab --> ImportError: cannotimport name inverse
by LATORNELL, Doug March 15, 2006

March 15, 2006

Thanks, Robert. I'll update my matplotlib build. Doug > -----Original Message----- > From: scipy-dev-bounces(a)scipy.net > [mailto:scipy-dev-bounces@scipy.net] On Behalf Of Robert Kern > Sent: March 15, 2006 15:40 > To: SciPy Developers List > Subject: Re: [SciPy-dev] import pylab --> ImportError: > cannotimport name inverse > > LATORNELL, Doug wrote: > > I'm seeing this problem too. > > > > I re-built NumPy and SciPy from SVN this afternoon. My > matplotlib has > > been static for several weeks. > > Travis has commited a fix to matplotlib's SVN repository. > > -- > Robert Kern > robert.kern(a)gmail.com > > "I have come to believe that the whole world is an enigma, a > harmless enigma that is made terrible by our own mad attempt > to interpret it as though it had an underlying truth." > -- Umberto Eco > > _______________________________________________ > Scipy-dev mailing list > Scipy-dev(a)scipy.net > http://www.scipy.net/mailman/listinfo/scipy-dev > This email and any files transmitted with it may contain privileged or confidential information and may be read or used only by the intended recipient. If you are not the intended recipient of the email or any of its attachments, please be advised that you have received this email in error and any use, dissemination, distribution, forwarding, printing or copying of this email or any attached files is strictly prohibited. If you have received this email in error, please immediately purge it and all attachments and notify the sender by reply email or contact the sender at the number listed.

1 0