Mailman 3 May 2011 - NumPy-Discussion

f2py and lapack DLARUV
by Stephane Raynaud 19 May '11

19 May '11

Hi, trying to build a fortran extension linked to lapack (f77), I faced a problem that I reduced to the following example. ------------ testfile.f90 ------------ subroutine testfunc integer,parameter :: n=2 integer :: iseed(4) real(kind=8) :: x(n) iseed = (/0,1,3,4/) call DLARUV(iseed, n, x) end subroutine testfunc -------------------------------------- Then : f2py -c -m testmod -llapack -lblas -L/home/raynaud/soft/lapack_gfortran/lib testfile.f90 It rises the following error (using gfortran) : /usr/bin/ld: /home/raynaud/soft/lapack_gfortran/lib/liblapack.a(dlaruv.o): relocation R_X86_64_32S against `.data' can not be used when making a shared object; recompile with -fPIC /home/raynaud/soft/lapack_gfortran/lib/liblapack.a: could not read symbols: Bad value BLAS and LAPACK where compilated with gfortran and options '-fPIC -O3'. I tried to use the DATA statement directly in a subroutine that I converted to a python module, and it worked. Any idea ? Thanks. -- Stephane Raynaud

1 0

Importing a list into a structured numpy array, what is my error?
by Robert Love 19 May '11

19 May '11

my_list = [[1960, 'W0'],[2001, 'D5']] my_arr = np.array(my_list, dtype=('uint', '|S2')) gives me an error I don't understand: ValueError: mismatch in size of old and new data-descriptor What are the old and new? I get the same problem if try np.asarray() If I make the dtype ('|S4', '|S4') I can read in as strings without errors How do I convert a list of mixed types to a structured array? Thanks for any pointers.

2 1

[ANN] Courses in Colorado: "Introduction to Python and Python for Scientists and Engineers" - one day left for early bird rate
by Mike Müller 18 May '11

18 May '11

Python Course in Golden, CO, USA ================================ **There is only one day left to take advantage of the early bird rate.** Introduction to Python and Python for Scientists and Engineers -------------------------------------------------------------- June 3 - 4, 2011 Introduction to Python June 5, 2011 Python for Scientists and Engineers Both courses can be booked individually or together. Venue: Colorado School of Mines, Golden, CO (20 minutes west of Denver) Trainer: Mike Müller Target Audience --------------- The introductory course is designed for people with basic programming background. Since it is a general introduction to Python it is suitable for everybody interested in Python. The scientist's course assumes a working knowledge of Python. You will be fine if you take the two-day introduction before hand. The topics are of general interest for scientists and engineers. Even though some examples come from the groundwater modeling domain, they are easy to understand for people without prior knowledge in this field. About the Trainer ----------------- Mike Müller, has been teaching Python since 2004. He is the founder of Python Academy and regularly gives open and in-house Python courses as well as tutorials at PyCon US, OSCON, EuroSciPy and PyCon Asia-Pacific. More Information and Course Registration ---------------------------------------- http://igwmc.mines.edu/short-course/intro_python.html -- Mike mmueller(a)python-academy.de

1 0

Windows dev-build Numpy 1.6.1
by Wieland Brendel 18 May '11

18 May '11

Hello, I am glad Mark was able to resolve the bug in einsum very fast since I am heavily relying on this function. I am working on a Windows 7 64 bit system with latest Python 2.6. I tried to build numpy from SVN following along http://www.scipy.org/Build_for_Windows After installing numpy I received the error RuntimeError: module compiled against ABI version 1000009 but this version of numpy is 2000000 This seems to be related to my Scipy installation, so I tried to build Scipy as well but couldn't get around installing ATLAS (I am used to using the superpacks...). Could somebody help me out? Maybe even with a 1.6.1 branch dev build? Thanks very much in advance!! Wieland

4 6

Re: [Numpy-discussion] strange dimension-dependent behaviour of einsum
by Wieland Brendel 18 May '11

18 May '11

3 3

cPickle.load says 'invalid signature'
by Neal Becker 18 May '11

18 May '11

The file is pickle saved on i386 and loaded on x86_64. It contains a numpy array (amoungst other things). On load it says: RuntimeError: invalid signature Is binary format not portable?

3 2

strange dimension-dependent behaviour of einsum
by Wieland Brendel 18 May '11

18 May '11

Hello, I am encountering a very strange behaviour of einsum on my machine. I tracked the problem down to the following test code: from numpy import * T = random.random((3,10,10)) W = random.random((3,10,7,275)) print all(einsum('ij...,j...->i...',T[0],W[0]) + einsum('ij...,j...->i...',T[1],W[1]) + einsum('ij...,j...->i...',T[2],W[2]) - einsum('pij,pjts->its',T,W)< 1e-10) print sum(abs(einsum('ij...,j...->i...',T[0],W[0]) + einsum('ij...,j...->i...',T[1],W[1]) + einsum('ij...,j...->i...',T[2],W[2]) - einsum('pij,pjts->its',T,W))) On my machine the equality is not fulfilled. However, this depends, strange enough, on the dimensions of W: if the last dimension e.g. is 500 instead of 275, things work again and the equality is fulfilled. Are you encountering similar problems or is this just my machine/installation? Since I am heavily relying on einsum but can't trust my code any more, I am happily waiting for you answers. Thanks a lot! Wieland

2 6

Linear assignment problem: Jonker-Volgenant algorithm
by Gael Varoquaux 17 May '11

17 May '11

Following a suggestion by Joseph, I am trying to implement the Jonker-Volgenant algorithm for best linear assignment in Python, using numpy. Unsuprisingly, it is proving time-costly. I cannot afford to spend too much time on this, as it not to solve a problem of mine, but for the scikits.learn. Thus I was wondering if someone had a BSD-licensed Python version of the algorithm that he would be willing to share. Cheers, Gael

5 13

fitting legendre polynomial
by Wolfgang Kerzendorf 17 May '11

17 May '11

Hello, The science package I'm using fits legendre polynomials to data. I heard it is more stable than the normal polynomials for a fit. Is there a polyfit for legendre polynomials? How do I do that with the new legendre polynomials module? Thanks Wolfgang

2 1

Trying to turn warning to errors
by Ondrej Marsalek 17 May '11

17 May '11

Dear all, I am trying to turn Python warnings to errors, but it does not work as expected with numpy. The following two piece of code work: $ python -W error -c 'import warnings; import numpy; warnings.warn("test", numpy.core.numeric.ComplexWarning)' Traceback (most recent call last): File "<string>", line 1, in <module> numpy.core.numeric.ComplexWarning: test $ python -W error -c 'import warnings; import numpy; raise numpy.core.numeric.ComplexWarning' Traceback (most recent call last): File "<string>", line 1, in <module> numpy.core.numeric.ComplexWarning while this does not (i.e. still produces just a warning): $ python -W error -c 'import numpy; x=numpy.ones(2); x+=1j' numpy.core.numeric.ComplexWarning: Casting complex values to real discards the imaginary part This is with Python 2.6.6 and numpy 1.5.1, but also confirmed with Python 2.7. Thanks for any suggested solutions, Ondrej Marsalek

2 3