Mailman 3 September 2013 - NumPy-Discussion

openblas?
by Neal Becker 23 Sep '13

23 Sep '13

Does numpy/scipy support building with openblas for blas,lapack instead of atlas?

2 1

Runtime Warning by numpy.divide(0.0, 0.0) can't be ignored
by Lorenzo Di Gregorio 23 Sep '13

23 Sep '13

Hello, numpy.divide(0.0,0.0) raises a warning which doesn't appear to get caught by numpy.seterr(divide='ignore') This example: import numpy numpy.divide(0.0,0.0) numpy.divide(1.0,0.0) numpy.seterr(divide='ignore') numpy.divide(1.0,0.0) numpy.divide(0.0,0.0) produces this output: Warning (from warnings module): File "C:\Users\ldigrego\Desktop\tst.py", line 2 numpy.divide(0.0,0.0) RuntimeWarning: invalid value encountered in divide Warning (from warnings module): File "C:\Users\ldigrego\Desktop\tst.py", line 3 numpy.divide(1.0,0.0) RuntimeWarning: divide by zero encountered in divide Warning (from warnings module): File "C:\Users\ldigrego\Desktop\tst.py", line 6 numpy.divide(0.0,0.0) RuntimeWarning: invalid value encountered in divide Strangely, on Ipython the numpy.seterr(divide='ignore') seems to catch the warning: import numpy numpy.divide(0.0,0.0) Out[1]: nan C:\eclipse\plugins\org.python.pydev_2.8.1.2013072611\pysrc\pydevconsole.py:1: RuntimeWarning: invalid value encountered in divide from __future__ import nested_scopes #Jython 2.1 support numpy.divide(1.0,0.0) Out[1]: inf C:\eclipse\plugins\org.python.pydev_2.8.1.2013072611\pysrc\pydevconsole.py:1: RuntimeWarning: divide by zero encountered in divide from __future__ import nested_scopes #Jython 2.1 support numpy.seterr(divide='ignore') Out[1]: {'divide': 'warn', 'invalid': 'warn', 'over': 'warn', 'under': 'ignore'} numpy.divide(1.0,0.0) Out[1]: inf numpy.divide(0.0,0.0) Out[1]: nan I could not find information on Google: is this a known problem? Is there a way to suppress this warning? I'm working on a 64b Win7 machine employing numpy-MKL-1.7.0.win-amd64-py2.7.exe. Best Regards, Lorenzo

2 1

Generating a (uniformly distributed) random bit list of length N
by David Goldsmith 23 Sep '13

23 Sep '13

Is np.random.randint(2, size=N) the fastest way to do this? Thanks! DG

2 1

Min Error
by David Reed 22 Sep '13

22 Sep '13

Hi, I am getting a strange error when finding the minimum of a matrix. The weird thing is I get this while running within iPython shell, and if I do %debug and go to the line where this fails and run the command `a = np.min(D, axis=0)`, I get no error. Here is the trace: draw_lines/main.pyc in haus_distance(A, B) 28 D = D.reshape((len(A), None)) 29 ---> 30 a = np.min(D, axis=0) 31 32 # vp = np.max(np.min(D, axis=0)) anaconda/lib/python2.7/site-packages/numpy/core/fromnumeric.pyc in amin(a, axis, out, keepdims) 1985 else: 1986 return _methods._amin(a, axis=axis, -> 1987 out=out, keepdims=keepdims) 1988 1989 def alen(a): anaconda/lib/python2.7/site-packages/numpy/core/_methods.pyc in _amin(a, axis, out, keepdims) 12 def _amin(a, axis=None, out=None, keepdims=False): 13 return um.minimum.reduce(a, axis=axis, ---> 14 out=out, keepdims=keepdims) 15 16 def _sum(a, axis=None, dtype=None, out=None, keepdims=False): TypeError: an integer is required

2 4

multiprocessing, numpy and 32-64 bit cohabitation
by Antony Lee 21 Sep '13

21 Sep '13

Hi all, While I realize that this is certainly tweaking multiprocessing beyond its specifications, I would like to use it on Windows to start a 32-bit Python process from a 64-bit Python process (use case: I need to interface with a 64-bit DLL and use an extension (pyFFTW) for which I can only find a 32-bit compiled version (yes, I could try to install MSVC and compile it myself but I'm trying to avoid that...)) In fact, this is "easy" to do by using multiprocessing.set_executable (...while that may not be its original role): import multiprocessing as mp import imp, site, sys if "32" in sys.executable: # checking for my 32-bit Python install del sys.path[1:] # recompute sys.path print(sys.path) site.main() print(sys.path) # now points to the 32bit site-packages import numpy if __name__ == '__main__': mp.set_executable(sys.executable.replace("33", "33-32")) # path of my 32-bit Python install mp.Process(target=lambda: None).start() The sys.path modifications are needed as otherwise the child process inherits the parent's sys.path and importing numpy (from the 64-bit path) fails as it is "not a valid Win32 application", complains Python (rightly). However, even after the sys.path modifications, the numpy import fails with the error message (that I had never seen before): <sorry, I can't copy paste from the Windows command prompt...> from . import multiarray # <- numpy/core/__init__.py, line 5 SystemError: initialization of multiarray raised an unreported exception Any hints as to how this could be fixed would be most welcome. Thanks in advance, Antony

6 15

ValueError: matrices are not aligned!!!
by Happyman 18 Sep '13

18 Sep '13

Hello, I am trying to solve linear Ax=b problem, but some error occurs in the process like: ---------------------------------------------- Traceback (most recent call last): File "C:\Python27\Conjugate_Gradient.py", line 66, in <module> x, iter_number = conjGrad(Av,A, x, b) File "C:\Python27\Conjugate_Gradient.py", line 51, in conjGrad u = Av(A,s) File "C:\Python27\Conjugate_Gradient.py", line 41, in Av return np.dot(A,v) ValueError: matrices are not aligned ----------------------------------------------- I put the code below to check it. import numpy as np import math def Av(A,v): return np.dot(A,v) def conjGrad(Av,A, x, b, tol=1.0e-9): n = len(x) r = b - Av(A,x) s = r.copy() for i in range(n): u = Av(A,s) alpha = np.dot(s,r)/np.dot(s,u) x = x + aplha*s r = b - Av(A,x) if (math.sqrt(np.dot(r,r))) < tol: break else: beta = - np.dot(r,u)/np.dot(s,u) s = r + beta * s return x,i if __name__ == '__main__': # Input values A = np.arange(32, dtype=float).reshape((4,8)) x = np.zeros(8) b = np.array([2.5, 4.5, 6.5, 8.0]) x, iter_number = conjGrad(Av,A, x, b) I would appreciate any solution to this problem... Thanks in advance -- happy man

2 2

Using setuptools develop mode with numpy distutils
by Chris Barker 18 Sep '13

18 Sep '13

Hi, I need to use numpy's distutils Extension to build a fortran / f2py extension. But I also really like setuptools' "develop" mode. However, if I do: from numpy.distutils.core import Extension and use that to define my Extension object, but use: from setuptools import setup And use that for the main setup, I get: error: each element of 'ext_modules' option must be an Extension instance or 2-tuple So apparently setuptools doesn't recognize numpy's Extension object as an Extension object. Is there a way around this? -Chris -- Christopher Barker, Ph.D. Oceanographer Emergency Response Division NOAA/NOS/OR&R (206) 526-6959 voice 7600 Sand Point Way NE (206) 526-6329 fax Seattle, WA 98115 (206) 526-6317 main reception Chris.Barker(a)noaa.gov

2 2

ANN: SfePy 2013.3
by Robert Cimrman 18 Sep '13

18 Sep '13

I am pleased to announce release 2013.3 of SfePy. Description ----------- SfePy (simple finite elements in Python) is a software for solving systems of coupled partial differential equations by the finite element method. The code is based on NumPy and SciPy packages. It is distributed under the new BSD license. Home page: http://sfepy.org Downloads, mailing list, wiki: http://code.google.com/p/sfepy/ Git (source) repository, issue tracker: http://github.com/sfepy Highlights of this release -------------------------- - implementation of Mesh topology data structures in C - implementation of regions based on C Mesh (*) - MultiProblem solver for conjugate solution of subproblems - new advanced examples (vibro-acoustics, Stokes flow with slip conditions) (*) Warning: region selection syntax has been changed in a principal way, see [1]. Besides the simple renaming, all regions meant for boundary conditions or boundary/surface integrals need to have their kind set explicitly to 'facet' (or 'edge' in 2D, 'face' in 3D). [1] http://sfepy.org/doc-devel/users_guide.html#regions For full release notes see http://docs.sfepy.org/doc/release_notes.html#id1 (rather long and technical). Best regards, Robert Cimrman and Contributors (*) (*) Contributors to this release (alphabetical order): Vladimír Lukeš

2 2

Handling warnings when solving nearly-singular matrices
by Aron Ahmadia 17 Sep '13

17 Sep '13

Hey folks, The mathematicians over at Walking Randomly have expressed some dissatisfaction with the lack of warnings when solving nearly-singular matrices: http://www.walkingrandomly.com/?p=5092 An issue has been raised here: https://github.com/numpy/numpy/issues/3755 The Julia developers are also discussing this here: https://github.com/JuliaLang/julia/issues/4286 I don't have the cycles to submit a patch immediately, but I can come up with something this weekend if somebody doesn't want to work on this sooner. Cheers, Aron

1 0

Unexpected casting result
by Warren Weckesser 16 Sep '13

16 Sep '13

An unexpected casting result was just reported on stackoverflow: http://stackoverflow.com/questions/18833639/attributeerror-in-python-numpy-… The following show the essence of the issue: In [1]: np.__version__ Out[1]: '1.9.0.dev-6ce65d8' In [2]: type(np.array(1.) * (2**64-1)) Out[2]: numpy.float64 In [3]: type(np.array(1.) * (2**64)) Out[3]: float Note that the result of `np.array(1.0) * 2**64` is a Python float, not a numpy float64. Is this intentional? (As pointed out in the stackoverflow question, the issue https://github.com/numpy/numpy/issues/3409 is at least tangentially related.) Warren

1 1

NumPy-Discussion September 2013