Mailman 3 July 2008 - NumPy-Discussion

svd
by Charles Doutriaux 16 Jul '08

16 Jul '08

Hello, I'm using 1.1.0 and I have a bizarre thing happening it seems as if: doing: import numpy SVD = numpy.linalg.svd if different as doing import numpy.oldnumeric.linear_algebra SVD = numpy.oldnumeric.linear_algebra.singular_value_decomposition In the first case passing an array (204,1484) retuns array of shape: svd: (204, 204) (204,) (1484, 1484) in the second case I get (what i expected actually): svd: (204, 204) (204,) (204, 1484) But looking at the code, it seems like numpy.oldnumeric.linear_algebra.singular_value_decomposition is basicalyy numpy.linalg.svd Any idea on what's happening here? Thx, C.

2 2

Ticket review #850: leak in _strings_richcompare
by Michael Abbott 16 Jul '08

16 Jul '08

This one is easy, ought to go in. Fixes a (not particularly likely) memory leak.

2 1

kinds
by Charles Doutriaux 16 Jul '08

16 Jul '08

Hello, A long long time ago, there used to be this module named "kinds" It's totally outdated nowdays but it had one nice functionality and i was wondering if you knew how to reproduce that it was: maxexp=kinds.default_float_kind.MAX_10_EXP minexp=kinds.default_float_kind.MIN_10_EXP and a bunch of similar flags that would basically tell you the limits on the machine you're running (or at least compiled on) Any idea on how to reproduce this? While we're at it, does anybody know of way in python to know how memory is available on your system (similar to the free call unde rLinux) I'm looking for something that works accross platforms (well we can forget windows for now I could live w/o it) Thanks, C.

2 2

Ticket review #849: reference to deallocated object?
by Michael Abbott 16 Jul '08

16 Jul '08

Tenuous but easy fix, and conformant to style elsewhere.

2 1

Second revised list of backports for 1.1.1.
by Charles R Harris 16 Jul '08

16 Jul '08

After the second pass only a few remain. Fernando, if you don't get to these I'll do them tomorrow. fperez r5298 r5301 r5303 Thanks to everyone for your prompt response. Chuck

2 3

Histogram bin definition
by Stéfan van der Walt 16 Jul '08

16 Jul '08

Hi all, I am busy documenting `histogram`, and the definition of a "bin" eludes me. Here is the behaviour that troubles me: >>> np.histogram([1,2,1], bins=[0, 1, 2, 3], new=True) (array([0, 2, 1]), array([0, 1, 2, 3])) >From this result, it seems as if a bin is defined as the half-open interval [right_edge, left_edge). Now, looks what happens in the following case: >>> np.histogram([1,2,3], bins=[0,1,2,3], new=True) (array([0, 1, 2]), array([0, 1, 2, 3])) Here, the last bin is defined by the closed interval [right_edge, left_edge]! Is this a bug, or a design consideration? Regards Stéfan

2 1

Recommendations for using numpy ma?
by Russell E. Owen 16 Jul '08

16 Jul '08

I have some code that does this: # an extra array cast is used because "compressed" returns what *looks* like an array # but is actually something else (I'm not sure exactly what) unmaskedArr = numpy.array( numpy.core.ma.array( dataArr, mask = mask & self.stretchExcludeBits, dtype = float, ).compressed()) That was working fine in numpy 1.0.4 but I've just gotten a report that it fails in 1.1. So...is there a notation that is safer: compatible with the widest possible range of versions? If I replace "numpy.core.ma" with "numpy.ma" this weems to work in 1.0.4 (I'm not sure about 1.1). But I fear it might not work with older versions of numpy. This software is used by a wide range of users with a wide range of versions of numpy. -- Russell

2 3

Infinity definitions
by Bruce Southey 16 Jul '08

16 Jul '08

Hi, Since we are discussing namespace and standardization, I am curious in why there are multiple definitions for defining infinity in numpy when perhaps there should be two (one for positive infinity and one for negative infinity). I really do understand that other people have use of these definitions and that it is easier to leave them in than take them out. Also, it is minor reduction in namespace because I do know that much of the namespace is either defining variables (like different floats and complex numbers) or mathematical functions (like logs and trig functions). Currently we have: numpy.Inf numpy.Infinity numpy.inf numpy.infty numpy.NINF numpy.PINF Most of these are defined in numeric.py: 'Inf = inf = infty = Infinity = PINF' In the f2py/tests subdirectories, the files return_real.py and return_complex.py uses both 'inf','Infinity'. The only occurrence of NINF and PINF are in core/src/umathmodule.c but I don't see any other usage. There does not seem to be any use of 'infty'. A similar case could be made for 'Not a Number' (nan = NaN = NAN). Regards Bruce

7 9

It appears that f2py fails to pass --compiler information to distutils on Windows XP
by Ryan Gutenkunst 15 Jul '08

15 Jul '08

Hi all, A project of mine uses f2py internally to dynamically build extension modules, and we've recently had problems with it on Windows XP using Python 2.5.2. I think I've isolated the problem to f2py passing the -- compiler option to distutils. It appears not to be doing so. If I run "python c:\python25\Scripts\f2py.py -c --compiler=mingw32" I get the message: running build running scons No module named msvccompiler in numpy.distutils; trying from distutils error: Python was built with Visual Studio 2003; extensions must be built with a compiler that can generate compatible binaries. Visual Studio 2003 was not found on this system. If you have Cygwin installed, you can try compiling with MingW32, by passing "-c mingw32" to setup.py. It appears that the mingw32 information isn't getting passed to distutils. I do indeed have a good installation of mingw32, as I can successfully build a related project that uses both C and Fortran extension modules via "python setup.py build -c mingw32". If I just run "python c:\python25\Scripts\f2py.py", the help screen ends with: Version: 2_5239 numpy Version: 1.1.0 Our Windows testing is very sporadic. I'm quite sure this worked a year ago, but I'm not sure when it broke. Does anyone have any suggestions? Is this an actual bug, or am I using something incorrectly? Thanks, Ryan -- Ryan Gutenkunst Biological Statistics and Computational Biology Cornell University http://pages.physics.cornell.edu/~rgutenkunst/

1 0

Calculating roots with negative numbers
by Matthias Hillenbrand 15 Jul '08

15 Jul '08

Hello, I want to calculate the root of a numpy array with negative numbers. Here is an example: x = linspace(-10,10,100) h = zeros(100) h[where(abs(x) < 2)] = sqrt( abs(x) -2 ) h[where(2 <= abs(x))] = 1j * sqrt( 2 - abs(x) ) Unfortunately I get the following error: Warning: invalid value encountered in sqrt and h contains some NaN-values How can I make a correct case differentiation in combination with roots that might contain negative values? Thank you very much, Matthias

2 2