Mailman 3 June 2007 - SciPy-User

bug in lexsort with two different dtypes?
by Tom Denniston June 26, 2007

June 26, 2007

In [1]: intArr1 = numpy.array([ 0, 1, 2,-2,-1, 5,-5,-5]) In [2]: intArr2 = numpy.array([1,1,1,2,2,2,3,4]) In [3]: charArr = numpy.array(['a','a','a','b','b','b','c','d']) Here I sort two int arrays. As expected intArr2 dominates intArr1 but the items with the same intArr2 values are sorted forwards according to intArr1 In [6]: numpy.lexsort((intArr1, intArr2)) Out[6]: array([0, 1, 2, 3, 4, 5, 6, 7]) This, however, looks like a bug to me. Here I sort an int array and a str array. As expected charArray dominates intArr1 but the items with the same charArray values are sorted *backwards* according to intArr1 In [5]: numpy.lexsort((intArr1, charArr)) Out[5]: array([2, 1, 0, 5, 4, 3, 6, 7]) Is this a bug or am I missing something? --Tom

1 0

floating point pedagogical problem
by Ryan Krauss June 26, 2007

June 26, 2007

I have a floating point issue that is forcing me to teach my students more python/computer science than I want to: In [36]: temp Out[36]: NumPy array, format: long [-1. -0.83045267 -0.64773663 -0.45185185 -0.24279835 -0.02057613 0.21481481 0.46337449 0.72510288 1. ] In [37]: arccos(temp) Out[37]: NumPy array, format: long [ nan 2.55071609 2.27540616 2.03963643 1.81604581 1.59137391 1.35429411 1.08899693 0.75961255 nan] We are doing a robotics problem involving inverse kinematics where they need to take the arccos and arcsin of some vectors. The problem is that In [38]: eps=1e-15 In [39]: -1-eps < temp[0] < -1+eps Out[39]: True So, my current solution is to check for theta +/- 1 +/- eps problems like this: tempout = [] for item in temp: if -1-eps<item<-1+eps: tempout.append(-1.0) elif 1-eps<item<1+eps: tempout.append(1.0) else: tempout.append(item) tempout = array(tempout) Is there a better way? Is making arccos and arcsin check for +/-1 +/- eps reasonable? Or should I give a lecture on floating point evils? Ryan

2 2

Is this a bug in stats.geom.pmf?
by Fernando Perez June 25, 2007

June 25, 2007

Hi all, just curious. The stats.geom docstring says: Geometric distribution geom.pmf(k,p) = (1-p)**(k-1)*p for k >= 1 But I see this: In [10]: k,p = 2.0,0.5 In [11]: (1-p)**(k-1)*p Out[11]: 0.25 In [12]: stats.geom.pmf(k,p) Out[12]: array(0.125) However: In [13]: stats.geom.pmf(k-1,p) Out[13]: array(0.25) Is this an off-by-one bug, or am I misreading something here? Cheers, f

3 7

matrix mult operator
by dmitrey June 25, 2007

June 25, 2007

hi all, don't you think that numpy way of handling matrix multiplication is too complicated: suppose I have F = A . (B . C)^T . D + C . ((D . B)^T * A) (last is elementwise multiplication, '.' are matrix mult) then I must do def evalF(A,B,C,D) from numpy import mdot, dot # Also, I'm not fond of writing such lines each time return mdot(A, dot(B, C).T, D) + dot(C, dot(D, B).T * A) it's very long to read, especially if I have much longer lines what about an operator, something like !* or *! def evalF(A,B,C, D): return A !* (B !* C).T !* D + C !* (D !* B).T * A it's much more readable. Of course, you can wait till Python will support unicode, but as for me I would prefere not to wait so long. I know that there had been already some discussions about the matrix multiplication, but I decided to rise the question once again, because mdot() turned out to not solve all problems. Also, I think the names 'dot'/ 'mdot' are inappropriate ones, because it is sound like 'dotwise' multiplication. D.

5 9

removing NaN...
by fred June 25, 2007

June 25, 2007

Hi, I work on data arrays with many NaN. How can I remove them and find min & max values ? TIA. Cheers, -- http://scipy.org/FredericPetit

2 2

[cookbook] matplotlib & plotting tuto...
by fred June 25, 2007

June 25, 2007

Hi, Matplotlib cookbok & Plotting Tutorial should not be merged ? Cheers, -- http://scipy.org/FredericPetit

2 2

can't build on OS X from SVN
by Zachary Pincus June 24, 2007

June 24, 2007

Hi all, I tried to build scipy from today's svn (r3105) on OS X (PPC), using GCC 4.0.1 and gfortran, (as 'python setup.py config_fc -- fcompiler=gnu95 build') and I got the following error: /usr/bin/ld: flag: -undefined dynamic_lookup can't be used with MACOSX_DEPLOYMENT_TARGET environment variable set to: 10.1 Which is strange since MACOSX_DEPLOYMENT_TARGET is set to 10.4. I nuked the build tree and the current install, made sure that the MACOSX_DEPLOYMENT_TARGET value was correct, tried again, and again got the same error. Any thoughts? (The context within which this error was raised is below). The only other irregularity during the build was this error text that was printed earlier: library 'mach' defined more than once, overwriting build_info {'sources': ['Lib/integrate/mach/d1mach.f', 'Lib/integrate/mach/ i1mach.f', 'Lib/integrate/mach/r1mach.f', 'Lib/integrate/mach/ xerror.f'], 'config_fc': {'noopt': ('Lib/integrate/setup.pyc', 1)}, 'source_languages': ['f77']} with {'sources': ['Lib/special/mach/ d1mach.f', 'Lib/special/mach/i1mach.f', 'Lib/special/mach/r1mach.f', 'Lib/special/mach/xerror.f'], 'config_fc': {'noopt': ('Lib/special/ setup.pyc', 1)}, 'source_languages': ['f77']}. Zach Pincus Program in Biomedical Informatics and Department of Biochemistry Stanford University School of Medicine /usr/bin/ld: flag: -undefined dynamic_lookup can't be used with MACOSX_DEPLOYMENT_TARGET environment variable set to: 10.1 collect2: ld returned 1 exit status /usr/bin/ld: flag: -undefined dynamic_lookup can't be used with MACOSX_DEPLOYMENT_TARGET environment variable set to: 10.1 collect2: ld returned 1 exit status error: Command "/usr/local/bin/gfortran -Wall -undefined dynamic_lookup -bundle build/temp.darwin-8.9.0-Power_Macintosh-2.4/ build/src.darwin-8.9.0-Power_Macintosh-2.4/Lib/fftpack/ _fftpackmodule.o build/temp.darwin-8.9.0-Power_Macintosh-2.4/Lib/ fftpack/src/zfft.o build/temp.darwin-8.9.0-Power_Macintosh-2.4/Lib/ fftpack/src/drfft.o build/temp.darwin-8.9.0-Power_Macintosh-2.4/Lib/ fftpack/src/zrfft.o build/temp.darwin-8.9.0-Power_Macintosh-2.4/Lib/ fftpack/src/zfftnd.o build/temp.darwin-8.9.0-Power_Macintosh-2.4/ build/src.darwin-8.9.0-Power_Macintosh-2.4/fortranobject.o -L/usr/ local/lib/gcc/powerpc-apple-darwin8.9.0/4.3.0 -Lbuild/ temp.darwin-8.9.0-Power_Macintosh-2.4 -ldfftpack -lgfortran -o build/ lib.darwin-8.9.0-Power_Macintosh-2.4/scipy/fftpack/_fftpack.so" failed with exit status 1

3 7

[ANN]New numpy, scipy and atlas rpms for FC 5, 6 and 7 and openSUSE (with 64 bits arch support)
by David Cournapeau June 24, 2007

June 24, 2007

Hi there, After quite some pain, I finally managed to build a LAPACK + ATLAS rpm useful for numpy and scipy. Read the following if you use Fedora Core or OpenSuse and are tired to install unsuccessfully numpy, scipy, BLAS, LAPACK or ATLAS. Instructions are given there: http://www.scipy.org/Installing_SciPy/Linux (ashigabou repository) Basically: - Fedora Core 5, 6 and 7 and openSUSE 10.2 are supported (x86, and x86_64 for FC 7 and openSuse). - binary rpms for numpy, scipy and blas/lapack dependencies. - source rpm for atlas, for a really easy, 3 commands build of ATLAS (should work for both x86 and x86_64). numpy and scipy are the last releases, including some backported changes to make it work on 64 bits. Atlas is the last developement version, with a trivial patch to build shared blas and lapack which can be used as drop in replacements for netlib blas and lapack. I would like to hear people complains. If people want other distributions supported by the opensuse build system (such as mandriva), I would like to hear it too. cheers, David

1 0

weave test failures
by Ryan Krauss June 23, 2007

June 23, 2007

Thanks to David, my seg fault problem with cluster is fixed. I have two remaining test failures and they are both with weave: ====================================================================== FAIL: check_1d_3 (scipy.weave.tests.test_size_check.test_dummy_array_indexing) ---------------------------------------------------------------------- Traceback (most recent call last): File "/usr/lib/python2.5/site-packages/scipy/weave/tests/test_size_check.py", line 168, in check_1d_3 self.generic_1d('a[-11:]') File "/usr/lib/python2.5/site-packages/scipy/weave/tests/test_size_check.py", line 135, in generic_1d self.generic_wrap(a,expr) File "/usr/lib/python2.5/site-packages/scipy/weave/tests/test_size_check.py", line 127, in generic_wrap self.generic_test(a,expr,desired) File "/usr/lib/python2.5/site-packages/scipy/weave/tests/test_size_check.py", line 123, in generic_test assert_array_equal(actual,desired, expr) File "/usr/lib/python2.5/site-packages/numpy/testing/utils.py", line 223, in assert_array_equal verbose=verbose, header='Arrays are not equal') File "/usr/lib/python2.5/site-packages/numpy/testing/utils.py", line 215, in assert_array_compare assert cond, msg AssertionError: Arrays are not equal a[-11:] (mismatch 100.0%) x: array([1]) y: array([10]) ====================================================================== FAIL: check_1d_6 (scipy.weave.tests.test_size_check.test_dummy_array_indexing) ---------------------------------------------------------------------- Traceback (most recent call last): File "/usr/lib/python2.5/site-packages/scipy/weave/tests/test_size_check.py", line 174, in check_1d_6 self.generic_1d('a[:-11]') File "/usr/lib/python2.5/site-packages/scipy/weave/tests/test_size_check.py", line 135, in generic_1d self.generic_wrap(a,expr) File "/usr/lib/python2.5/site-packages/scipy/weave/tests/test_size_check.py", line 127, in generic_wrap self.generic_test(a,expr,desired) File "/usr/lib/python2.5/site-packages/scipy/weave/tests/test_size_check.py", line 123, in generic_test assert_array_equal(actual,desired, expr) File "/usr/lib/python2.5/site-packages/numpy/testing/utils.py", line 223, in assert_array_equal verbose=verbose, header='Arrays are not equal') File "/usr/lib/python2.5/site-packages/numpy/testing/utils.py", line 215, in assert_array_compare assert cond, msg AssertionError: Arrays are not equal a[:-11] (mismatch 100.0%) x: array([9]) y: array([0]) ---------------------------------------------------------------------- Ran 1771 tests in 4.081s FAILED (failures=2) Out[3]: <unittest.TextTestRunner object at 0x963416c> How do I fix these? Thanks, Ryan

1 0

Ubuntu Feisty upgrade problems
by Ryan Krauss June 23, 2007

June 23, 2007

Sorry, I feel like I whine about installation problems all the time lately. I just upgraded my home desktop to Ubuntu Feisty and have had some issues getting Scipy working. I think part of the problem is that this computer was running python 2.4 and there were some problems with side-by-side installation. I removed all references to the 2.4 directory from my path and PYTHONPATH, but still couldn't make the ubuntu python-scipy packages work. Anyways, I installed from source and things seemed to go o.k., but scipy.test() seg faulted. I ran this script that Robert Kern helped me come up with when I was testing my AMD executable for processors without SSE2: from numpy import NumpyTest packages = """ scipy.cluster scipy.fftpack scipy.integrate scipy.interpolate scipy.io scipy.lib scipy.linalg scipy.linsolve scipy.maxentropy scipy.misc scipy.odr scipy.optimize scipy.signal scipy.sparse scipy.special scipy.stats scipy.stsci scipy.weave """.strip().split() #packages.pop(0) #packages.pop() for subpkg in packages: print subpkg t = NumpyTest(subpkg) t.test(1, 2) And I get this: In [3]: run scipy_test.py scipy.cluster Warning: No test file found in /usr/lib/python2.5/site-packages/scipy/cluster/tests for module <module 'scipy.cluster' from '...ges/scipy/cluster/__init__.pyc'> Warning: No test file found in /usr/lib/python2.5/site-packages/scipy/cluster/tests for module <module 'scipy.cluster.info' from '...ackages/scipy/cluster/info.pyc'> Found 9 tests for scipy.cluster.vq Found 0 tests for __main__ Testing that kmeans2 init methods work.Segmentation fault (core dumped) meaning the seg fault is from scipy.cluster. If I uncomment out the to pop lines (not testing scipy.cluster or scipy.weave) I get no errors. I don't use cluster that I can think of, but I would prefer a completely working scipy installation. Can someone please help me with this? Thanks, Ryan

3 7