Mailman 3 October 2004 - SciPy-Dev

Test failures
by answer＠tnoo.net Oct. 7, 2004

Oct. 7, 2004

Hi Here is the output (2 failures, 2 errors) of the tests on my box Python 2.3.3 (#1, Apr 6 2004, 01:47:39) [GCC 3.3.3 (SuSE Linux)] on linux2 gcc version 3.3.3 >>> scipy.__version__ '0.3.1_287.4340' ====================================================================== ERROR: check_defective1 (scipy.linalg.matfuncs.test_matfuncs.test_signm) ---------------------------------------------------------------------- Traceback (most recent call last): File "/usr/lib/python2.3/site-packages/scipy/linalg/tests/test_matfuncs.py", lin e 48, in check_defective1 r = signm(a) File "/usr/lib/python2.3/site-packages/scipy/linalg/matfuncs.py", line 267, in s ignm iS0 = inv(S0) File "/usr/lib/python2.3/site-packages/scipy/linalg/basic.py", line 176, in inv a1 = asarray_chkfinite(a) File "/usr/lib/python2.3/site-packages/scipy_base/function_base.py", line 33, in asarray_chkfinite raise ValueError, "Array must not contain infs or nans." ValueError: Array must not contain infs or nans. ====================================================================== ERROR: check_defective3 (scipy.linalg.matfuncs.test_matfuncs.test_signm) ---------------------------------------------------------------------- Traceback (most recent call last): File "/usr/lib/python2.3/site-packages/scipy/linalg/tests/test_matfuncs.py", lin e 69, in check_defective3 r = signm(a) File "/usr/lib/python2.3/site-packages/scipy/linalg/matfuncs.py", line 267, in s ignm iS0 = inv(S0) File "/usr/lib/python2.3/site-packages/scipy/linalg/basic.py", line 176, in inv a1 = asarray_chkfinite(a) File "/usr/lib/python2.3/site-packages/scipy_base/function_base.py", line 33, in asarray_chkfinite raise ValueError, "Array must not contain infs or nans." ValueError: Array must not contain infs or nans. ====================================================================== FAIL: check_nils_20Feb04 (scipy.linalg.basic.test_basic.test_solve) ---------------------------------------------------------------------- Traceback (most recent call last): File "/usr/lib/python2.3/site-packages/scipy/linalg/tests/test_basic.py", line 1 12, in check_nils_20Feb04 assert_array_almost_equal(X,Ainv) File "/usr/lib/python2.3/site-packages/scipy_test/testing.py", line 684, in asse rt_array_almost_equal assert cond,\ AssertionError: Arrays are not almost equal (mismatch 50.0%): Array 1: [[-43.6032104+13.1816127j 18.994452 +15.8269203j] [ 37.1899302 +2.3715189j -9.3834051-17.4805578j]] Array 2: [[-27.7762901 -5.8128393j 18.994452 +15.8269203j] [ 19.7093724+11.754924j -9.3834051-17.4805578j]] ====================================================================== FAIL: check_cdf (scipy.stats.distributions.test_distributions.test_gumbel_l) ---------------------------------------------------------------------- Traceback (most recent call last): File "<string>", line 10, in check_cdf AssertionError: D = 0.292383217812; pval = 0.00455751677656; alpha = 0.01 args = () ---------------------------------------------------------------------- Ran 972 tests in 2.078s FAILED (failures=2, errors=2) -- Martin Lüthi answer(a)tnoo.net

1 0

Matrix Structure Plots in scipy ?
by Nils Wagner Oct. 6, 2004

Oct. 6, 2004

Hi all, First of all I would like to thank Pearu for implementing Matrix Market I/O support ! The number of outputs is more than described in help (io.mmio) mmread(source) Reads the contents of the Matrix Market file 'filename' into the matrix 'A'. Inputs: source - Matrix Market filename (extension .mtx) or open file object. Outputs: a - sparse or full matrix Actually b,m,n,s = mmread('young1c') should be used. Is it possible to generate matrix structure plots in scipy ? http://math.nist.gov/MatrixMarket/structureplots.html BTW, there is another "important" matrix format namely matrices generated by NASTRAN (commercial Finite Element program).There exists already a tool that can read and write both dense and sparse matrices, either real or complex, either little or big endian, and either text (any precision) or binary. You can find the code by downloading http://savannah.nongnu.org/cgi-bin/viewcvs/tops/tops/usr/extra/op4tools/ Pearu, please can you manage the support of this format in scipy too ? I believe devoutly, that many users will appreciate this feature. Thanks in advance Nils

2 4

gui_thread and wxPython-2.5.2.8
by Prabhu Ramachandran Oct. 5, 2004

Oct. 5, 2004

Hi, I noticed that gui_thread does not work with wxPython-2.5.2.8. It worked perfectly under 2.4.x. So I spent quite a bit of time figuring out what was going on today. I finally have something that works. The only problem is that the solution requires a small patch to wx/__init__.py in the wxPython installation. Of course, I am using Pearu's excellent wxPython_thread code to handle the threading issues. AFAICT the major problem with 2.5.2.8 and gui_thread is that a lot is done when one does `import wx`. This is unlike 2.4.x where importing wxPython alone merely imports wxc. It also turns out that a function called `_core_._wxPySetDictionary(vars())` in wx/_core.py is the major pain in the neck. Boy did that take a while to figure out! What do you folks recommend that I do about the issue. Does someone have the time to look at my changes to gui_thread/wxPython_thread.py and the patch to wx/__init__.py and help? Should I also ask Robin Dunn if I could work out a cleaner solution by perhaps changing the way _wxPySetDictionary works? I am not sure. Any advice would be appreciated. In any case the result of my hacking is that with 2 small patches one can use gui_thread fully from within IPython for both new and old (not tested but should work) versions of wxPython. Also, if 0.3.2 is round the corner is this a good thing to push into scipy? I am not sure how many folks use gui_thread, but it would certainly be cool if this could make it into 0.3.2. Comments/suggestions? cheers, prabhu

3 20

Re: [SciPy-user] Matrix square root
by Travis Oliphant Oct. 5, 2004

Oct. 5, 2004

Pearu Peterson wrote: > > > On Mon, 4 Oct 2004, Nils Wagner wrote: > >> AFAIK there is no direct support for sqrtm but you may use funm. > > > I just discovered that there might be a bug in funm. I have added > a test to scipy.linalg.test_matfuncs that demonstrates the bug. > So, be aware to check the results of funm(..,sqrt). As far as I undertand, the idea of a general purpose matrix function has precision problems. The algorithm in scipy is based on a published algorithm using the Schur decomposition. That is why it is usually a good idea to write separate algorithms for special matrix functions if possible. Is this really a bug, or just a problem with using the generic funm algorithm when a more accurate algorithm exists. The fact that you can return a reasonable square root for non-special matrices makes me believe that it is not a bug in funm, but a demonstration that the generic funm algorithm does not always produce the best results for specific functions of matrices. -Travis O.

2 2

Hessenberg form
by Nils Wagner Oct. 4, 2004

Oct. 4, 2004

Hi all, I'd like to have a function called hess in linalg which reduces a general matrix A to upper Hessenberg form before the next release. Any comment ? Nils

5 6

gui_thread: works with 2.4.x and 2.5.x!
by Prabhu Ramachandran Oct. 2, 2004

Oct. 2, 2004

Hi everyone, Finally, some good news! I have gui_thread working nicely using both wxPython 2.4.x and 2.5.x on Linux. I've tested this on Debian Sarge with wxPython-2.5.2.8u (built from source) and wxPython-2.4.2.6 installed via apt-get install libwxgtk2.4-python. With the new code I am able to run the respective wxPython demo's interactively from within IPython and script it! I've checked in this into CVS. Please test it out and let me know if there are problems. cheers, prabhu

3 2

gui_thread: please test under 2.5.x
by Prabhu Ramachandran Oct. 2, 2004

Oct. 2, 2004

Hi, I've checked in a version that works nicely for me from the interpreter with pretty complex apps like wxPythonDemo. For example: import gui_thread gui_thread.start() cd <path_to_wxPython_demo> import Main d = Main.wxPythonDemo(None, 'a') d.Show(1) d.SetTitle('wxPythonDemo from IPython!') This well enough for me under wxPython-2.5.2.8u and I'm happy. I'll try and work on a version similar to this that works for earlier wxPython versions. This version is not guaranteed to work with the demo for wxPython-2.4.x and is untested for that version. In the meantime *please* test gui_thread and see if it works for you on your platform/machine using wxPython-2.5.x. Thanks! cheers, prabhu

1 0

Pre-release fixes
by Robert Kern Oct. 1, 2004

Oct. 1, 2004

I've just checked in a fix for fastumath's logical functions. They used to give wrong answers on OSX and caused some tests to fail. I believe this to be an endianness issue, but I don't have another big-endian machine to test on. All tests still pass on Athlon/Linux. The changes made largely track what Numeric did, so I'm not sure if these were conscious deviations to speed up fastumath or just a failure to keep up with the bugfixes in Numeric. I have only implemented fastumath_unsigned.inc as yet. I will make similar changes in fastumath_nounsigned.inc as I accumulate tuits. Travis, could you check my changes to make sure I didn't revert differences you deliberately made? If they were deliberate, then we need to figure a way to revert to the correct, if slower, Numeric implementations on the Mac. Also, if anyone is using SciPy on a Sun or other big-endian machine, I would appreciate the output of scipy.test() with and without the patch. Thank you, everyone. -- Robert Kern rkern(a)ucsd.edu "In the fields of hell where the grass grows high Are the graves of dreams allowed to die." -- Richard Harter

2 6

Re: [Scipy-cvs] world/scipy_core/scipy_distutils from_template.py, 1.6, 1.7
by Pearu Peterson Oct. 1, 2004

Oct. 1, 2004

Travis, What is the reason that process_str() lowers its input? There may be other macros in addition to MIN, MAX that .pyf files may contain and it would be nice to if process_str would not lower the input. Pearu On Fri, 1 Oct 2004 travo(a)scipy.org wrote: > Update of /home/cvsroot/world/scipy_core/scipy_distutils > In directory scipy.org:/tmp/cvs-serv19408/scipy_core/scipy_distutils > > Modified Files: > from_template.py > Log Message: > Added wrappers to iterative methods > > > Index: from_template.py > =================================================================== > RCS file: /home/cvsroot/world/scipy_core/scipy_distutils/from_template.py,v > retrieving revision 1.6 > retrieving revision 1.7 > diff -C2 -d -r1.6 -r1.7 > *** from_template.py 31 Aug 2004 01:00:05 -0000 1.6 > --- from_template.py 1 Oct 2004 08:30:06 -0000 1.7 > *************** > *** 173,176 **** > --- 173,187 ---- > char = str[ind] > return ''.join(extra) > + > + maxre = re.compile(r"max[(](.+),(.+)[)]") > + minre = re.compile(r"min[(](.+),(.+)[)]") > + > + def fix_capitals(astr): > + # Need to convert max(x,x) to MAX(x,x) > + # and min(x,x) to MIN(x,x) > + astr = maxre.sub(r"MAX(\g<1>,\g<2>)",astr) > + astr = minre.sub(r"MIN(\g<1>,\g<2>)",astr) > + return astr > + > > def process_str(allstr): > *************** > *** 184,193 **** > oldend = 0 > for sub in struct: > ! writestr += newstr[oldend:sub[0]] > expanded = expand_sub(newstr[sub[0]:sub[1]],get_line_header(newstr,sub[0])) > ! writestr += expanded > oldend = sub[1] > > ! writestr += newstr[oldend:] > return writestr > > --- 195,205 ---- > oldend = 0 > for sub in struct: > ! writestr += fix_capitals(newstr[oldend:sub[0]]) > expanded = expand_sub(newstr[sub[0]:sub[1]],get_line_header(newstr,sub[0])) > ! writestr += fix_capitals(expanded) > oldend = sub[1] > > ! > ! writestr += fix_capitals(newstr[oldend:]) > return writestr > > > _______________________________________________ > Scipy-cvs mailing list > Scipy-cvs(a)scipy.org > http://scipy.net/mailman/listinfo/scipy-cvs >

3 3

MCMC, Kalman Filtering, AI for SciPy?
by chris＠fisher.forestry.uga.edu Oct. 1, 2004

Oct. 1, 2004

For the past couple of years, I have been maintaining a package called PyMC (http://pymc.sourceforge.net) which provides a general Markov chain Monte Carlo sampler (Metropolis-Hastings), linear and non-linear Kalman filtering algorithms and several reinforcement learning algorithms for AI applications. To date, it has been used mostly for ecological modeling applications, but is far more general than that. This package currently relies on SciPy, particularly for such tasks as plotting, and uses f2py extensively. I am wondering if there is interest in integrating such functionality into SciPy. If so, I would be happy to contribute, integrate and maintain the code under the auspices of the SciPy project. If not, I am equally content to continue maintaining PyMC separately. I encourage those interested to check out PyMC and let me know if it would be worthwhile to add any of this to SciPy. Cheers, Chris Fonnesbeck chris (at) fisher.forestry.uga.edu

12 62