Mailman 3 January 2008 - NumPy-Discussion

Do not use -fomit-frame-pointer with mingw !
by David Cournapeau 21 Jan '08

21 Jan '08

Hi, I don't know if this is known, but since I wasted half a day on it, I thought I could avoid the pain for someone else: do not build numpy with -fomit-frame-pointer with mingw (windows), it will crash. I don't know if this just cannot be used, if this is a mingw bug, or something on our side. But you do not want to look for a bug caused by using this option. cheers, David

1 0

Optimizing speed for large-array inter-element algorithms (specifically, color space conversion)
by James A. Bednar 21 Jan '08

21 Jan '08

| Date: Mon, 21 Jan 2008 09:39:21 -0500 | From: "theodore test" <tteststudent(a)gmail.com> | | Hello all, | | I'm scratching my head over how to make this image color space | conversion from "RGB" to "HSV" quicker. It requires input from all | three bands of a given pixel at each pixel, and thus can't be | easily flattened. I've already implemented psyco and removed the | extra step of calling colorsys's rgb_to_hsv function by adapting | the code into my primary for loop. | | Right now it takes around 9 seconds for a single 1600x1200 RGB image, a | conversion that I've seen implemented more or less instantly in, say, | ImageJ. What can I do to make this conversion more efficient? I would love to see a decent implementation of that! We have versions quite similar to yours, instead. We've had a to-do list item to reimplement rgb_to_hsv (and its converse hsv_to_rgb) in C using weave, but haven't ever gotten around to it. I believe the best solution would be for PIL to be modified to accept an image of type HSV; it already accepts four others, including the non-obvious CMYK: m.mode => string Image mode. This is a string specifying the pixel format used by the image. Typical values are "1", "L", "RGB", or "CMYK." Adding CSV to this list would seem quite reasonable, and would allow simple conversion between the various color spaces. But I don't know anything about the internals of the PIL code, to know whether that would indeed make sense. Jim

1 0

numpy.distutils does not output compilation warning on win32 ?
by David Cournapeau 21 Jan '08

21 Jan '08

Hi, I noticed a strange behaviour with distutils: when compiling C code on windows (using mingw), the compilation warning are not output on the console. For example, umathmodule.c compilation emits the following warnings: numpy\core\src\umathmodule.c.src:128: warning: static declaration of 'acoshf' follows non-static declaration numpy\core\src\umathmodule.c.src:136: warning: static declaration of 'asinhf' follows non-static declaration I think this is coming from distutils because when I execute the exact same command on the shell, I get those warnings. If there is an error (by inserting #error), then all the warnings appear also when using distutils, which would suggest that distutils checks the return status of gcc to decide when to output should be sent on the shell ? Anyway, if we can do something about it, I think it would be better to always output warnings (it took me quite a while to understand why I got warnings with scons and not with distutils...). cheers, David

2 2

Bus Error with object arrays on big endian system
by Christopher Hanley 21 Jan '08

21 Jan '08

Hi, The following will cause a bus error on a big endian machine (Solaris 10 Sun in this case): > Python 2.5.1 (r251:54863, Jun 29 2007, 15:29:55) [C] on sunos5 > Type "help", "copyright", "credits" or "license" for more information. >>>> import numpy >>>> o = numpy.ndarray(shape=3,dtype=[('SEGMENT', '|S4'), ('SPEC_FOUND', '|i1')]) >>>> o1 = o.getfield(numpy.dtype('|S4'),0) >>>> print o1[0] > UXÐ >>>> print o1[1] > 4 >>>> print o1[2] > NT >>>> print o1 > Bus error (core dumped) There are no issues on Linux or Mac OS X Intel based systems. This example was done on the latest svn version of numpy (r1.0.5.dev47360). Does anyone have an idea of what may have broken? Thank you for your time and help, Chris -- Christopher Hanley Systems Software Engineer Space Telescope Science Institute 3700 San Martin Drive Baltimore MD, 21218 (410) 338-4338

1 0

Scipy page
by Charles R Harris 21 Jan '08

21 Jan '08

Hi All, Just two comments on the current scipy top page: 1) I really miss the recent changes link, it was the only way I could keep current on the Wiki contents. It was also a good way to track spam. 2) The addition of the bug reporting link is good, but it is hard to notice. How about putting it up top using one of the neat python logos with a bug addition? Chuck

5 10

creating rec arrays
by Nicholas 21 Jan '08

21 Jan '08

I am trying to package a set of vector arrays into a single rec array with column titles. c = numpy.rec.fromarrays([a,b],names='a,b') the problem is there is no 'copy' argument in .fromarrays which I can set to False. i.e. is there a way to do this without the overhead of copying? Thanks Nicholas

2 1

r4730 broke numpy in maskedarray branch
by Eric Firing 20 Jan '08

20 Jan '08

Stefan, The renaming of concatenator to AxisConcatenator in r4730 in the maskedaray branch needs to be propagated into ma/extras.py; I don't know whether there are other places that similar changes are needed. As it stands, after 4730, numpy cannot be imported. Eric

3 3

maskedarray branch
by David Huard 20 Jan '08

20 Jan '08

Hi all, I am trying to install the maskedarray numpy branch for use with the timeseries package and I get the following error when importing numpy: /usr/local/lib64/python2.5/site-packages/numpy/__init__.py in <module>() 45 import random 46 import ctypeslib ---> 47 import ma 48 49 # Make these accessible from numpy name-space /usr/local/lib64/python2.5/site-packages/numpy/ma/__init__.py in <module>() 12 __date__ = '$Date: 2007-10-29 17:18:13 +0200 (Mon, 29 Oct 2007) $' 13 ---> 14 import core 15 from core import * 16 /usr/local/lib64/python2.5/site-packages/numpy/ma/core.py in <module>() 3073 return numpy.inner(fa, fb).view(MaskedArray) 3074 inner.__doc__ = numpy.inner.__doc__ -> 3075 inner.__doc__ += "\n*Notes*:\n Masked values are replaced by 0." 3076 innerproduct = inner 3077 TypeError: unsupported operand type(s) for +=: 'NoneType' and 'str' Everything being fresh from svn, I tried installing the maskedarray living in the sandbox, but I had no success either with the installation of timeseries. Cheers, David

2 6

Problems installing on OS X
by Sam Watson 20 Jan '08

20 Jan '08

Dear all, I am running OS X 10.4.11, and when I run gcc -v, I get: Using built-in specs. Target: powerpc-apple-darwin8 Configured with: /private/var/tmp/gcc/gcc-5026.obj~19/src/configure -- disable-checking --prefix=/usr --mandir=/share/man --enable- languages=c,objc,c++,obj-c++ --program-transform-name=/^[cg][^+.-]*$/ s/$/-4.0/ --with-gxx-include-dir=/include/gcc/darwin/4.0/c++ -- build=powerpc-apple-darwin8 --host=powerpc-apple-darwin8 -- target=powerpc-apple-darwin8 Thread model: posix gcc version 4.0.0 (Apple Computer, Inc. build 5026) However, when I move into the numpy-1.0.4 directory and run python setup.py install, I get hundreds of error messages which appear to relate to my C compiler. Here is the output with lots of similar C messages in the middle omitted: non-existing path in 'numpy/distutils': 'site.cfg' F2PY Version 2_4422 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'] running install running build running config_cc unifing config_cc, config, build_clib, build_ext, build commands -- compiler options running config_fc unifing config_fc, config, build_clib, build_ext, build commands -- fcompiler options running build_src building py_modules sources building extension "numpy.core.multiarray" sources Generating build/src.macosx-10.3-fat-2.5/numpy/core/config.h customize NAGFCompiler Could not locate executable f95 customize AbsoftFCompiler Could not locate executable f90 Found executable /sw/bin/f77 absoft: no Fortran 90 compiler found absoft: no Fortran 90 compiler found absoft: no Fortran 90 compiler found absoft: no Fortran 90 compiler found absoft: no Fortran 90 compiler found absoft: no Fortran 90 compiler found customize IBMFCompiler Could not locate executable xlf90 Could not locate executable xlf customize IntelFCompiler Could not locate executable ifort Could not locate executable ifc customize GnuFCompiler Found executable /sw/bin/g77 gnu: no Fortran 90 compiler found gnu: no Fortran 90 compiler found customize GnuFCompiler gnu: no Fortran 90 compiler found gnu: no Fortran 90 compiler found customize GnuFCompiler using config C compiler: gcc -arch ppc -arch i386 -isysroot /Developer/SDKs/ MacOSX10.4u.sdk -fno-strict-aliasing -Wno-long-double -no-cpp-precomp -mno-fused-madd -fno-common -dynamic -DNDEBUG -g -O3 compile options: '-I/Library/Frameworks/Python.framework/Versions/2.5/ include/python2.5 -Inumpy/core/src -Inumpy/core/include -I/Library/ Frameworks/Python.framework/Versions/2.5/include/python2.5 -c' gcc: _configtest.c In file included from _configtest.c:2: /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ Python.h:18:20: error: limits.h: No such file or directory /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ Python.h:21:2: error: #error "Something's broken. UCHAR_MAX should be defined in limits.h." /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ Python.h:25:2: error: #error "Python's source code assumes C's unsigned char is an 8-bit type." /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ Python.h:32:19: error: stdio.h: No such file or directory /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ Python.h:34:5: error: #error "Python.h requires that stdio.h define NULL." /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ Python.h:37:20: error: string.h: No such file or directory /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ Python.h:39:19: error: errno.h: No such file or directory /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ Python.h:41:20: error: stdlib.h: No such file or directory ... /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ pystrtod.h:11: error: parse error before 'size_t' In file included from _configtest.c:2: /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ Python.h:134: error: parse error before '*' token /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ Python.h:134: error: parse error before '*' token /Library/Frameworks/Python.framework/Versions/2.5/include/python2.5/ Python.h:134: warning: data definition has no type or storage class _configtest.c: In function 'main': _configtest.c:9: error: 'FILE' undeclared (first use in this function) _configtest.c:9: error: (Each undeclared identifier is reported only once _configtest.c:9: error: for each function it appears in.) _configtest.c:9: error: 'fp' undeclared (first use in this function) _configtest.c:17: warning: incompatible implicit declaration of built- in function 'fprintf' lipo: can't figure out the architecture type of: /var/tmp//ccPjVTJJ.out failure. removing: _configtest.c _configtest.o I know that my C compiler works, because I do my research in C on a regular basis. Does anybody have any ideas? Thanks, Sam Watson

2 1

numpy c++ integration
by Neal Becker 19 Jan '08

19 Jan '08

If anyone is interested in this subject, I have various tests I've been working on, available via this mercurial repo: https://nbecker.dyndns.org/hg/numpy/

1 0