Mailman 3 April 2006 - SciPy-User

install - floating point exception (scipy-tests)
by python＠axtom.com 27 Apr '06

27 Apr '06

Hi all, I have installed scipy, and have some fp exceptions when running the tests. Does anyone else experiment the same thing? Thanks -- Jean Pierre configuration ----------------- Debian GNU/Linux 3.1 kernel 2.6 compilers: gcc-3.3.5 and g77-3.3.5 (c++ not installed) python-2.4.2 blas http://www.netlib.org/blas/blas.tgz -- installation from src lapack http://www.netlib.org/lapack/lapack.tgz -- installation from src fftw http://www.fftw.org/fftw-2.1.5.tar.gz -- installation from src numpy-0.9.6 scipy-0.4.8 (without cluster and weave packages) output of the tests (by package and by module when there is an exception) ------------------------------------------------------------------ >>> from scipy import integrate;NumpyTest(integrate).test(level=10) Found 10 tests for scipy.integrate.quadpack Found 1 tests for scipy.integrate Found 0 tests for __main__ ..Floating point exception >>> from scipy import fftpack;NumpyTest(fftpack).test(level=10) Found 23 tests for scipy.fftpack.basic Found 24 tests for scipy.fftpack.pseudo_diffs Found 4 tests for scipy.fftpack.helper Found 0 tests for __main__ Fast Fourier Transform ================================================= | real input | complex input ------------------------------------------------- size | scipy | Numeric | scipy | Numeric ------------------------------------------------- 100 | 0.10 | N/A | 0.10 | N/A (secs for 7000 calls) 1000 | 0.08 | N/A | 0.12 | N/A (secs for 2000 calls) 256 | 0.17 | N/A | 0.19 | N/A (secs for 10000 calls) 512 | 0.23 | N/A | 0.31 | N/A (secs for 10000 calls) 1024 | 0.04 | N/A | 0.05 | N/A (secs for 1000 calls) 2048 | 0.08 | N/A | 0.11 | N/A (secs for 1000 calls) 4096 | 0.07 | N/A | 0.13 | N/A (secs for 500 calls) 8192 | 0.19 | N/A | 0.57 | N/A (secs for 500 calls) ..Warning: Skipping check_djbfft (failed to import FFT) .. Multi-dimensional Fast Fourier Transform =================================================== | real input | complex input --------------------------------------------------- size | scipy | Numeric | scipy | Numeric --------------------------------------------------- 100x100 | 0.09 | N/A | 0.07 | N/A (secs for 100 calls) 1000x100 | 0.09 | N/A | 0.08 | N/A (secs for 7 calls) 256x256 | 0.11 | N/A | 0.12 | N/A (secs for 10 calls) 512x512 | 0.30 | N/A | 0.30 | N/A (secs for 3 calls) ..... Inverse Fast Fourier Transform =============================================== | real input | complex input ----------------------------------------------- size | scipy | Numeric | scipy | Numeric ----------------------------------------------- 100 | 0.09 | N/A | 0.13 | N/A (secs for 7000 calls) 1000 | 0.09 | N/A | 0.18 | N/A (secs for 2000 calls) 256 | 0.17 | N/A | 0.22 | N/A (secs for 10000 calls) 512 | 0.26 | N/A | 0.35 | N/A (secs for 10000 calls) 1024 | 0.04 | N/A | 0.07 | N/A (secs for 1000 calls) 2048 | 0.09 | N/A | 0.12 | N/A (secs for 1000 calls) 4096 | 0.09 | N/A | 0.15 | N/A (secs for 500 calls) 8192 | 0.21 | N/A | 0.59 | N/A (secs for 500 calls) ....... Inverse Fast Fourier Transform (real data) ================================== size | scipy | Numeric ---------------------------------- 100 | 0.12 | N/A (secs for 7000 calls) 1000 | 0.10 | N/A (secs for 2000 calls) 256 | 0.19 | N/A (secs for 10000 calls) 512 | 0.26 | N/A (secs for 10000 calls) 1024 | 0.04 | N/A (secs for 1000 calls) 2048 | 0.08 | N/A (secs for 1000 calls) 4096 | 0.09 | N/A (secs for 500 calls) 8192 | 0.19 | N/A (secs for 500 calls) ..Warning: Skipping check_djbfft (failed to import FFT) .. Fast Fourier Transform (real data) ================================== size | scipy | Numeric ---------------------------------- 100 | 0.10 | N/A (secs for 7000 calls) 1000 | 0.08 | N/A (secs for 2000 calls) 256 | 0.18 | N/A (secs for 10000 calls) 512 | 0.24 | N/A (secs for 10000 calls) 1024 | 0.03 | N/A (secs for 1000 calls) 2048 | 0.07 | N/A (secs for 1000 calls) 4096 | 0.07 | N/A (secs for 500 calls) 8192 | 0.17 | N/A (secs for 500 calls) ..Warning: Skipping check_djbfft (failed to import FFT: No module named FFT) . Differentiation of periodic functions ===================================== size | convolve | naive ------------------------------------- 100 | 0.03 | 0.23 (secs for 1500 calls) 1000 | 0.03 | 0.21 (secs for 300 calls) 256 | 0.04 | 0.34 (secs for 1500 calls) 512 | 0.04 | 0.38 (secs for 1000 calls) 1024 | 0.02 | 0.34 (secs for 500 calls) 2048 | 0.03 | 0.27 (secs for 200 calls) 4096 | 0.02 | 0.28 (secs for 100 calls) 8192 | 0.04 | 0.33 (secs for 50 calls) .......... Hilbert transform of periodic functions ========================================= size | optimized | naive ----------------------------------------- 100 | 0.03 | 0.17 (secs for 1500 calls) 1000 | 0.03 | 0.14 (secs for 300 calls) 256 | 0.04 | 0.24 (secs for 1500 calls) 512 | 0.03 | 0.26 (secs for 1000 calls) 1024 | 0.03 | 0.23 (secs for 500 calls) 2048 | 0.02 | 0.16 (secs for 200 calls) 4096 | 0.03 | 0.19 (secs for 100 calls) 8192 | 0.03 | 0.25 (secs for 50 calls) ........ Shifting periodic functions ============================== size | optimized | naive ------------------------------ 100 | 0.03 | 0.23 (secs for 1500 calls) 1000 | 0.01 | 0.25 (secs for 300 calls) 256 | 0.03 | 0.38 (secs for 1500 calls) 512 | 0.03 | 0.43 (secs for 1000 calls) 1024 | 0.02 | 0.39 (secs for 500 calls) 2048 | 0.01 | 0.30 (secs for 200 calls) 4096 | 0.03 | 0.31 (secs for 100 calls) 8192 | 0.03 | 0.32 (secs for 50 calls) .. Tilbert transform of periodic functions ========================================= size | optimized | naive ----------------------------------------- 100 | 0.02 | 0.23 (secs for 1500 calls) 1000 | 0.02 | 0.17 (secs for 300 calls) 256 | 0.04 | 0.33 (secs for 1500 calls) 512 | 0.04 | 0.32 (secs for 1000 calls) 1024 | 0.03 | 0.28 (secs for 500 calls) 2048 | 0.03 | 0.20 (secs for 200 calls) 4096 | 0.03 | 0.21 (secs for 100 calls) 8192 | 0.04 | 0.25 (secs for 50 calls) ........ ---------------------------------------------------------------------- Ran 51 tests in 26.269s OK <unittest.TextTestRunner object at 0x406432ac> >>> from scipy import interpolate;NumpyTest(interpolate).test(level=10) Found 5 tests for scipy.interpolate.fitpack Found 0 tests for __main__ /opt/scipy/lib/scipy/interpolate/fitpack2.py:410: UserWarning: The coefficients of the spline returned have been computed as the minimal norm least-squares solution of a (numerically) rank deficient system (deficiency=7). If deficiency is large, the results may be inaccurate. Deficiency may strongly depend on the value of eps. warnings.warn(message) ..... ---------------------------------------------------------------------- Ran 5 tests in 0.012s OK <unittest.TextTestRunner object at 0x40643c4c> >>> from scipy import io;NumpyTest(io).test(level=10) Found 4 tests for scipy.io.array_import Found 12 tests for scipy.io.mmio Found 0 tests for __main__ Don't worry about a warning regarding the number of bytes read. Warning: 1000000 bytes requested, 20 bytes read. ................ ---------------------------------------------------------------------- Ran 16 tests in 0.137s OK <unittest.TextTestRunner object at 0x406ce66c> >>> from scipy import lib;NumpyTest(lib).test(level=10); **************************************************************** WARNING: clapack module is empty ----------- See scipy/INSTALL.txt for troubleshooting. Notes: * If atlas library is not found by numpy/distutils/system_info.py, then scipy uses flapack instead of clapack. **************************************************************** Found 42 tests for scipy.lib.lapack Found 0 tests for __main__ .......................................... ---------------------------------------------------------------------- Ran 42 tests in 0.059s OK <unittest.TextTestRunner object at 0x410bb9ec> >>> from scipy import linalg;NumpyTest(linalg).test(level=10) Found 128 tests for scipy.linalg.fblas Found 37 tests for scipy.linalg.decomp Found 4 tests for scipy.linalg.lapack Found 44 tests for scipy.linalg.basic Found 7 tests for scipy.linalg.matfuncs Found 14 tests for scipy.linalg.blas Found 0 tests for __main__ ...caxpy:n=4 ..caxpy:n=3 ....ccopy:n=4 ..ccopy:n=3 .............cscal:n=4 ....cswap:n=4 ..cswap:n=3 .....daxpy:n=4 ..daxpy:n=3 ....dcopy:n=4 ..dcopy:n=3 .............dscal:n=4 ....dswap:n=4 ..dswap:n=3 .....saxpy:n=4 ..saxpy:n=3 ....scopy:n=4 ..scopy:n=3 .............sscal:n=4 ....sswap:n=4 ..sswap:n=3 .....zaxpy:n=4 ..zaxpy:n=3 ....zcopy:n=4 ..zcopy:n=3 .............zscal:n=4 ....zswap:n=4 ..zswap:n=3 ......... Finding matrix eigenvalues ================================== | contiguous ---------------------------------------------- size | scipy 20 | 0.08 (secs for 150 calls) 100 | 0.17 (secs for 7 calls) 200 | 0.36 (secs for 2 calls) .................Floating point exception >>> from scipy import maxentropy;NumpyTest(maxentropy).test(level=10) Found 2 tests for scipy.maxentropy Found 0 tests for __main__ .. ---------------------------------------------------------------------- Ran 2 tests in 0.003s OK <unittest.TextTestRunner object at 0x406e550c> >>> from scipy import ndimage;NumpyTest(ndimage).test(level=10) Found 397 tests for scipy.ndimage Found 0 tests for __main__ ............................................................................................................................................................................................................................................................................................................................................................................................................. ---------------------------------------------------------------------- Ran 397 tests in 1.174s OK <unittest.TextTestRunner object at 0x406edcac> >>> from scipy import optimize;NumpyTest(optimize).test(level=10) Found 6 tests for scipy.optimize.optimize Found 2 tests for scipy.optimize.zeros Found 1 tests for scipy.optimize.cobyla Found 0 tests for __main__ ......%s f2 is a symmetric parabola, x**2 - 1 f3 is a quartic polynomial with large hump in interval f4 is step function with a discontinuity at 1 f5 is a hyperbola with vertical asymptote at 1 f6 has random values positive to left of 1 , negative to right of course these are not real problems. They just test how the 'good' solvers behave in bad circumstances where bisection is really the best. A good solver should not be much worse than bisection in such circumstance, while being faster for smooth monotone sorts of functions. TESTING SPEED times in seconds for 2000 iterations function f2 cc.bisect : 0.080 cc.ridder : 0.040 cc.brenth : 0.030 cc.brentq : 0.030 function f3 cc.bisect : 0.110 cc.ridder : 0.030 cc.brenth : 0.040 cc.brentq : 0.030 function f4 cc.bisect : 0.090 cc.ridder : 0.110 cc.brenth : 0.100 cc.brentq : 0.110 function f5 cc.bisect : 0.090 cc.ridder : 0.120 cc.brenth : 0.110 cc.brentq : 0.110 function f6 cc.bisect : 0.100 cc.ridder : 0.120 cc.brenth : 0.110 cc.brentq : 0.130 .TESTING CONVERGENCE zero should be 1 function f2 cc.bisect : 1.0000000000001952 cc.ridder : 1.0000000000004658 cc.brenth : 0.9999999999999997 cc.brentq : 0.9999999999999577 function f3 cc.bisect : 1.0000000000001952 cc.ridder : 1.0000000000000000 cc.brenth : 1.0000000000000009 cc.brentq : 1.0000000000000011 function f4 cc.bisect : 1.0000000000001952 cc.ridder : 1.0000000000001452 cc.brenth : 0.9999999999993339 cc.brentq : 0.9999999999993339 function f5 cc.bisect : 1.0000000000001952 cc.ridder : 1.0000000000004574 cc.brenth : 0.9999999999991442 cc.brentq : 0.9999999999991442 function f6 cc.bisect : 1.0000000000001952 cc.ridder : 0.9999999999995509 cc.brenth : 1.0000000000004117 cc.brentq : 0.9999999999988777 .Result: [ 4.957975 0.64690335] (exact result = 4.955356249106168, 0.666666666666666) . ---------------------------------------------------------------------- Ran 9 tests in 2.344s OK <unittest.TextTestRunner object at 0x40ece94c> >>> from scipy import signal;NumpyTest(signal).test(level=10) Found 4 tests for scipy.signal.signaltools Found 0 tests for __main__ .... ---------------------------------------------------------------------- Ran 4 tests in 0.005s OK <unittest.TextTestRunner object at 0x410b00cc> >>> from scipy import sparse;NumpyTest(sparse).test(level=10) Found 89 tests for scipy.sparse.sparse Found 0 tests for __main__ 2 3 1 2 2 3 2 1 3 3 3 3 1 3 3 . 3 3 1 3 3 ...........Use minimum degree ordering on A'+A. .....................Use minimum degree ordering on A'+A. .....................Use minimum degree ordering on A'+A. .......................Use minimum degree ordering on A'+A. ............ ---------------------------------------------------------------------- Ran 89 tests in 0.487s OK <unittest.TextTestRunner object at 0x410b022c> >>> from scipy import special;NumpyTest(special).test(level=10) Found 341 tests for scipy.special.basic Found 0 tests for __main__ .........Floating point exception >>> from scipy import stats;NumpyTest(stats).test(level=10) Found 95 tests for scipy.stats.stats Found 70 tests for scipy.stats.distributions Found 10 tests for scipy.stats.morestats Found 0 tests for __main__ ..................................................................................................................................................Floating point exception ***************BY MODULE****************** python ./integrate/tests/test_integrate.py -l 10 Found 1 tests for __main__ Residual: 1.05006950433e-07 . ---------------------------------------------------------------------- Ran 1 test in 0.010s OK python ./integrate/tests/test_quadpack.py -l 10 Found 10 tests for __main__ ..Floating point exception python ./linalg/tests/test_atlas_version.py -l 10 NO ATLAS INFO AVAILABLE python ./linalg/tests/test_basic.py -l 10 Found 44 tests for __main__ Finding matrix determinant ================================== | contiguous | non-contiguous ---------------------------------------------- size | scipy | basic | scipy | basic 20 | 0.21 | 0.22 | 0.26 | 0.28 (secs for 2000 calls) 100 | 0.52 | 0.47 | 0.66 | 0.68 (secs for 300 calls) 500 | 0.68 | 0.68 | 0.72 | 0.73 (secs for 4 calls) ...... Finding matrix inverse ================================== | contiguous | non-contiguous ---------------------------------------------- size | scipy | basic | scipy | basic 20 | 0.34 | 0.31 | 0.36 | 0.36 (secs for 2000 calls) 100 | 1.21 | 1.51 | 1.22 | 1.68 (secs for 300 calls) 500 | 2.42 | 2.73 | 2.40 | 2.68 (secs for 4 calls) ....../opt/numpy/lib/numpy/core/oldnumeric.py:573: DeprecationWarning: integer argument expected, got float result = a.round(decimals) .......Floating point exception python ./linalg/tests/test_blas.py -l 10 Found 14 tests for __main__ **************************************************************** WARNING: cblas module is empty ----------- See scipy/INSTALL.txt for troubleshooting. Notes: * If atlas library is not found by numpy/distutils/system_info.py, then scipy uses fblas instead of cblas. **************************************************************** .............. ---------------------------------------------------------------------- Ran 14 tests in 0.013s OK python ./linalg/tests/test_decomp.py -l 10 Found 37 tests for __main__ ....... Finding matrix eigenvalues ================================== | contiguous ---------------------------------------------- size | scipy 20 | 0.08 (secs for 150 calls) 100 | 0.16 (secs for 7 calls) 200 | 0.35 (secs for 2 calls) .................Floating point exception python ./linalg/tests/test_fblas.py -l 10 Found 128 tests for __main__ ...caxpy:n=4 ..caxpy:n=3 ....ccopy:n=4 ..ccopy:n=3 .............cscal:n=4 ....cswap:n=4 ..cswap:n=3 .....daxpy:n=4 ..daxpy:n=3 ....dcopy:n=4 ..dcopy:n=3 .............dscal:n=4 ....dswap:n=4 ..dswap:n=3 .....saxpy:n=4 ..saxpy:n=3 ....scopy:n=4 ..scopy:n=3 .............sscal:n=4 ....sswap:n=4 ..sswap:n=3 .....zaxpy:n=4 ..zaxpy:n=3 ....zcopy:n=4 ..zcopy:n=3 .............zscal:n=4 ....zswap:n=4 ..zswap:n=3 .. ---------------------------------------------------------------------- Ran 128 tests in 0.074s OK python ./linalg/tests/test_lapack.py -l 10 Found 4 tests for __main__ .. **************************************************************** WARNING: clapack module is empty ----------- See scipy/INSTALL.txt for troubleshooting. Notes: * If atlas library is not found by numpy/distutils/system_info.py, then scipy uses flapack instead of clapack. **************************************************************** .. ---------------------------------------------------------------------- Ran 4 tests in 0.004s OK python ./linalg/tests/test_matfuncs.py -l 10 Found 7 tests for __main__ .Floating point exception python ./stats/tests/test_distributions.py -l 10 Found 70 tests for stats.distributions Found 0 tests for __main__ ...................................................Floating point exception python ./stats/tests/test_morestats.py -l 10 Found 10 tests for __main__ ..Ties preclude use of exact statistic. ..Ties preclude use of exact statistic. .....Floating point exception python ./stats/tests/test_stats.py -l 10 Found 95 tests for __main__ ............................................................................................... ---------------------------------------------------------------------- Ran 95 tests in 0.050s OK python ./special/tests/test_basic.py -l 10 Found 341 tests for __main__ .........Floating point exception

2 1

repmat of matrix returns array
by Keith Goodman 27 Apr '06

27 Apr '06

Do most operations on matrices return matrices? In porting Octave code, I noticed that the repmat of a matrix is an array. I decided to use matrices simply because that's what they are called in Octave. What do you recommend for new users who have a background in Octave, matrix or array? Do people generally pick one and not use the other?

3 3

import stats -> failed
by Keith Goodman 27 Apr '06

27 Apr '06

I installed scipy from SVN source less than an hour ago. I can't import stats. (If it's not something I did wrong, would a bug report like this be useful information?) >> from scipy import * --------------------------------------------------------------------------- exceptions.ValueError Traceback (most recent call last) /home/me/<ipython console> /usr/local/scipy/lib/python2.4/site-packages/scipy/stats/__init__.py 5 from info import __doc__ 6 ----> 7 from stats import * 8 from distributions import * 9 from rv import * /usr/local/scipy/lib/python2.4/site-packages/scipy/stats/stats.py 1693 1694 import scipy.stats -> 1695 import distributions 1696 def kstest(rvs, cdf, args=(), N=20): 1697 """Return the D-value and the p-value for a /usr/local/scipy/lib/python2.4/site-packages/scipy/stats/distributions.py 3812 lvals = where(vals==0,0.0,log(vals)) 3813 return -sum(vals*lvals) -> 3814 binom = binom_gen(name='binom',shapes="n,pr",extradoc=""" 3815 3816 Binomial distribution /usr/local/scipy/lib/python2.4/site-packages/scipy/stats/distributions.py in __init__(self, a, b, name, badvalue, moment_tol, values, inc, longname, shapes, extradoc) 3373 self.numargs=0 3374 else: -> 3375 self._vecppf = new.instancemethod(sgf(_drv2_ppfsingle,otypes='d'), 3376 self, rv_discrete) 3377 self.generic_moment = new.instancemethod(sgf(_drv2_moment, /usr/local/scipy/lib/python2.4/site-packages/numpy/lib/function_base.py in __init__(self, pyfunc, otypes, doc) 594 """ 595 def __init__(self, pyfunc, otypes='', doc=None): --> 596 nin, ndefault = _get_nargs(pyfunc) 597 self.thefunc = pyfunc 598 self.ufunc = None /usr/local/scipy/lib/python2.4/site-packages/numpy/lib/function_base.py in _get_nargs(obj) 559 nargs -= 1 560 return nargs, ndefaults --> 561 raise ValueError, 'failed to determine the number of arguments for %s' % (obj) 562 563 ValueError: failed to determine the number of arguments for <function _drv2_ppfsingle at 0x2aaab2995320>

2 1

Chang*ed* the Trac authentication
by Robert Kern 26 Apr '06

26 Apr '06

Trying not to embarass myself again, I made the changes without telling you. :-) In order to create or modify Wiki pages or tickets on the NumPy and SciPy Tracs, you will have to be logged in. You can register yourself by clicking the "Register" link in the upper right-hand corner of the page. Developers who previously had accounts have the same username/password as before. You can now change your password if you like. Only developers have the ability to close tickets, delete Wiki pages entirely, or create new ticket reports (and possibly a couple of other things). Developers, please enter your name and email by clicking on the "Settings" link up at top once logged in. Thank you for your patience. If there are any problems, please email me, and I will try to correct them quickly. -- 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

2 1

example dict_sort.py compile error
by Chuckles McGregor 26 Apr '06

26 Apr '06

good day, I'm trying to run the dict_sort.py script from the weave/examples dir and at this piece of the script: def c_sort2(adict): assert(type(adict) is dict) code = """ #line 44 "dict_sort.py" py::list keys = adict.keys(); py::list items(keys.len()); keys.sort(); int N = keys.length(); for(int i = 0; i < N;i++) items[i] = adict[keys[i]]; return_val = items; """ return inline_tools.inline(code,['adict'],verbose=1) I'm getting this error from the compiler: dict_sort.py(49) : error C2593: 'operator [' is ambiguous C:\Python24\Lib\site-packages\scipy\weave\scxx/dict.h(120): could be 'py ::object::keyed_ref py::dict::operator [](const std::string &)' C:\Python24\Lib\site-packages\scipy\weave\scxx/dict.h(113): or 'py::obje ct::keyed_ref py::dict::operator [](const char *)' C:\Python24\Lib\site-packages\scipy\weave\scxx/dict.h(105): or 'py::obje ct::keyed_ref py::dict::operator [](const std::complex<double> &)' C:\Python24\Lib\site-packages\scipy\weave\scxx/dict.h(101): or 'py::obje ct::keyed_ref py::dict::operator [](double)' C:\Python24\Lib\site-packages\scipy\weave\scxx/dict.h(97): or 'py::objec t::keyed_ref py::dict::operator [](int)' while trying to match the argument list '(py::dict, py::indexed_ref)' Traceback (most recent call last): File "dict_sort.py", line 119, in ? sort_compare(a,n) File "dict_sort.py", line 96, in sort_compare b=c_sort2(a) File "dict_sort.py", line 52, in c_sort2 return inline_tools.inline(code,['adict'],verbose=1) File "C:\Python24\Lib\site-packages\scipy\weave\inline_tools.py", line 334, in inline auto_downcast = auto_downcast, File "C:\Python24\Lib\site-packages\scipy\weave\inline_tools.py", line 442, in compile_function verbose=verbose, **kw) File "C:\Python24\Lib\site-packages\scipy\weave\ext_tools.py", line 353, in co mpile verbose = verbose, **kw) File "C:\Python24\Lib\site-packages\scipy\weave\build_tools.py", line 274, in build_extension setup(name = module_name, ext_modules = [ext],verbose=verb) File "C:\Python24\Lib\site-packages\numpy\distutils\core.py", line 85, in setu p return old_setup(**new_attr) File "C:\Python24\lib\distutils\core.py", line 166, in setup raise SystemExit, "error: " + str(msg) distutils.errors.CompileError: error: Command "cl.exe /c /nologo /Ox /MD /W3 /EHsc /DNDEBUG -IC:\Python24\Lib\site-packages\scipy\weave -IC:\Python24\Lib\site-p ackages\scipy\weave\scxx -IC:\python24\lib\site-packages\numpy\core\include -IC: \python24\include -IC:\python24\PC /Tpc:\docume~1\mcgreg~1\locals~1\temp\mcgrego r1\python24_compiled\sc_9fef7eb8a0d63221b946305b186f457b1.cpp /Foc:\docume~1\mcg reg~1\locals~1\temp\mcgregor1\python24_intermediate\compiler_d41d8cd98f00b204e98 00998ecf8427e\Release\docume~1\mcgreg~1\locals~1\temp\mcgregor1\python24_compile d\sc_9fef7eb8a0d63221b946305b186f457b1.obj /Zm1000" failed with exit status 2 I'm using python 2.4.3, scipy 0.4.8, and the visual C++ express edition (ver 8) compiler. any suggestions? chuckles

1 0

singular matrix linalg.basic.LinAlgError in optimize.leastsq
by Taro Sato 26 Apr '06

26 Apr '06

I frequently use optimize.leastsq, and there is this annoying tendency for the routine to throw an exception when the resulting covariant matrix ends up singular. The error message follows. Basically, I need full_output=1 to just obtain a stopping condition in the output parameter ier, but the way the code prepare the full output, it also tries to compute covariant matrix. The problem is when it fails, it throws an exception which is not handled within minpack.py: ----------------------------------------------------------------- Traceback (most recent call last): File "./test.py", line 27, in ? if __name__ == '__main__': main() File "./test.py", line 23, in main p = O.leastsq(residual, [10.,.1,.1], args=(x,y),full_output=1) File "/usr/lib/python2.3/site-packages/scipy/optimize/minpack.py", line 271, in leastsq cov_x = sl.inv(dot(transpose(R),R)) File "/usr/lib/python2.3/site-packages/scipy/linalg/basic.py", line 221, in inv if info>0: raise LinAlgError, "singular matrix" scipy.linalg.basic.LinAlgError: singular matrix ---------------------------------------------------------------- I'm not sure when exactly a covariant matrix cannot be computed this way. My test code (attached below and used to produce the error above) seems to indicate when the initial guess parameters are way off from the best parameters, the above error can result. The existing behavior of leastsq is very annoying, since to obtain a stopping condition that can be handled easily (i.e., in ier as an integer), I must use full_output=1, but when the computation of a covarant matrix is not well behaving, it simply explodes; there's no way to get all the other info that are relevant for knowing why the fitting failed. Is there any better way to handle the situation (which might include suggesting a modification to minpack.py)? Also, it appears to make more sense (at least to me) for leastsq to return ier when full_output=0, rather than mesg, as it makes error handling easier...a string message is human friendly but not really friendly to coders.... :) By the way, in the mailing list archive, I noticed there was a similar request about a year ago for adding an exception handling in this case. Thank you for your time, Taro ---CODE BEGINS (test.py)------------------------------------------- #!/usr/bin/env python import numpy as N import scipy.optimize as O def fgauss(lamb, params): lambc, a, b = params return 1.+a*N.exp(-0.5*((lamb-lambc)/b)**2) def residual(p, x, y): return y - fgauss(x, p) def main(): p0 = [100., 50., 5.] x = N.arange(200).astype(float) y0 = fgauss(x, p0) y = N.zeros(y0.shape).astype(float) for i in xrange(y.size): y[i] = N.random.poisson(y0[i]) p = O.leastsq(residual, [10.,.1,.1], args=(x,y),full_output=1) print p[0],p[3:] if __name__ == '__main__': main() ---CODE ENDS-------------------------------------------

2 2

Re: [SciPy-user] OSX Issue -- Symbol not found: _fprintf$LDBLStub
by Bill Northcott 25 Apr '06

25 Apr '06

On 25/04/2006, at 12:38 PM, Greg wrote: > Hey! I'm attempting to get SciPy to run on OSX 10.4 (Tiger), with > the latest version of Xcode (2.2.1, I believe) installed. When I run > python and import scipy.fftpack, I get the following error: > > Python 2.4.3 (#1, Mar 30 2006, 11:02:15) > [GCC 4.0.1 (Apple Computer, Inc. build 5250)] on darwin > Type "help", "copyright", "credits" or "license" for more information. >>>> import scipy.fftpack > Traceback (most recent call last): > File "<stdin>", line 1, in ? > File "/Library/Frameworks/Python.framework/Versions/2.4/lib/ > python2.4/site-packages/scipy/fftpack/__init__.py", line 10, in ? > from basic import * > File "/Library/Frameworks/Python.framework/Versions/2.4/lib/ > python2.4/site-packages/scipy/fftpack/basic.py", line 13, in ? > import _fftpack as fftpack > ImportError: Failure linking new module: /Library/Frameworks/ > Python.framework/Versions/2.4/lib/python2.4/site-packages/scipy/ > fftpack/_fftpack.so: Symbol not found: _fprintf$LDBLStub > .......... > > Does anyone have thoughts on what could be going wrong? This is caused by trying to link object code files and or static libraries produced by different versions of gcc. It will go away if you ensure that all objects are compiled either with gcc-3.x/g77 or gcc-4.x/gfortran. It does not matter what compiler is used for linked dynamic libraries, which is why they are preferred. Bill Northcott

1 0

Re: [SciPy-user] Unable to build : Unknown distribution option: 'configuration'
by Etienne Ringuet 25 Apr '06

25 Apr '06

Hello Robert, I do have the latest version of numpy, compiled from yesterday svn. I am using CentOS 4, python 2.3, I build BLAS and LAPACK from source, I tried removing blas and lapack librairies from the CentOS install. Thanks, Etienne -----Message d'origine----- De : Robert Kern [mailto:robert.kern@gmail.com] Envoyé : 25 avril 2006 12:51 À : ringueet(a)ete.inrs.ca;SciPy Users List Objet : Re: [SciPy-user] Unable to build : Unknown distribution option: 'configuration' Etienne Ringuet wrote: > This is my first post to the list so hello everyone. > > I am trying to build today's svn source of scipy on a dual opteron server. > > I am using this guide as a reference: > http://pong.tamu.edu/tiki/tiki-view_blog_post.php?blogId=6&postId=97 > > Here is the error: > > [root@marie scipy]# python setup.py build > /usr/lib64/python2.3/distutils/dist.py:227: UserWarning: Unknown > distribution option: 'configuration' Please also get the most recent SVN of numpy. -- 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

1 0

Unable to build : Unknown distribution option: 'configuration'
by Etienne Ringuet 25 Apr '06

25 Apr '06

This is my first post to the list so hello everyone. I am trying to build today's svn source of scipy on a dual opteron server. I am using this guide as a reference: http://pong.tamu.edu/tiki/tiki-view_blog_post.php?blogId=6&postId=97 Here is the error: [root@marie scipy]# python setup.py build /usr/lib64/python2.3/distutils/dist.py:227: UserWarning: Unknown distribution option: 'configuration' warnings.warn(msg) running build running config_fc [root@marie scipy]# uname -a Linux marie.ad.inrs.ca 2.6.9-5.0.5.ELsmp #1 SMP Tue Apr 19 17:06:07 CDT 2005 x86_64 x86_64 x86_64 GNU/Linux I don't know much about python, I am just adding requested packages to our cluster. Thanks in advance for your help, Etienne Ringuet

2 1

ndimage and 64 bit
by Gennan Chen 25 Apr '06

25 Apr '06

Hi! It looks like some of functions I want to port from Matlab are already in scipy.ndimage. However, I remember i saw a post a while ago about it is not working for 64 bit. Is it still true? or It is just my illusion. Gen-Nan Chen, PhD Chief Scientist Research and Development Group CorTechs Labs Inc (www.cortechs.net) 1020 Prospect St., #304, La Jolla, CA, 92037 Tel: 1-858-459-9700 ext 16 Fax: 1-858-459-9705 Email: gnchen(a)cortechs.net

4 6