Mailman 3 March 2008 - SciPy-User

[ann] aureservoir0.1 - library for analog reservoir computing neural networks (Echo State Networks)
by Georg Holzmann 13 Mar '08

13 Mar '08

Hallo! This is a release of aureservoir, an open-source C++ library for analog reservoir computing neural networks (mainly Echo State Networks) with bindings to python/numpy. The goal of this library is low memory usage, be efficient, extensively tested and easy to extend with new algorithms. Networks can be used with double or singe precision floating points and various different activation functions, simulation, training and adaptation algorithms are implemented and can be exchanged at runtime. For an introduction into reservoir computing neural networks, a feature overview, installation instructions, examples and documentation see: http://aureservoir.sourceforge.net/ Any feedback is welcome and I am happy to help out with installation/usage problems ! LG Georg

1 0

"in" vs. looped "=="
by R. Padraic Springuel 12 Mar '08

12 Mar '08

I have some code that uses a "1. in x" statement early on, where x is a rank-1 ndarray with dtype=float. This statement returns False and the program continues accordingly. However, later on in the same code I loop over the elements of x and the statement "x[i] == 1" is evaluated and comes back as True for some elements of x. To my mind, that shouldn't be happening. Is there a difference in how "1. in x" (i.e. the __contains__ method of ndarray) and "x[i] == 1." (i.e. the __eq__ property of float) behave that I'm not aware of? -- R. Padraic Springuel Research Assistant Department of Physics and Astronomy University of Maine Bennett 309 Office Hours: By appointment only

2 1

scipy installation without root access
by Alastair Basden 12 Mar '08

12 Mar '08

Hi, can anyone give me some help/tips installing a working scipy on a Suse 10.0 machine (86_64) without root access? I can't use RPMs, as some of the dependencies are non-relocatable... The main problem that I'm having (as I see it) is getting the gfortran/g77 mix right, such that the functions all work when installed... Thanks...

3 5

Re: [SciPy-user] FFTW use
by Charrett, Thomas 12 Mar '08

12 Mar '08

Thanks for the prompt reply, I've worked it out now, the problem was not with the FFT (although the numbers did differ slightly from Matlab - rounding errors?) but in my processing where I cut out a window from the FFT and place into a new array, however I forgot to set the type of the array to complex128, hence the strange results when computing the inverse FFT. I'll probably go back to using the precompiled version of scipy for now as it seems to work fine. Tom Charrett

1 0

RHEL 5 and CENTOS 5 rpms for blas/lapack/numpy/scipy available on ashigabou
by David Cournapeau 11 Mar '08

11 Mar '08

Hi, Since some people had problems with RHEL/CENTOS, and since the opensuse build system does provide facilities to build rpms for RHEL and CENTOS for some time, I quickly updated the ashigabou repository to handle those distributions. I also added opensuse 10.3 and FC 8, but those did not require any changes: http://download.opensuse.org/repositories/home:/ashigabou/ (note that it may take time for the rpms to appear there from the time they successfully build on the compiler farm, which they just did). cheers, David

1 0

Re: [SciPy-user] Problem with scipy.optimize.leastsq: Improper input parameters
by Adam Ginsburg 11 Mar '08

11 Mar '08

Thanks Dave, Mark. I eventually figured out the input parameters and got my minimization to work out; the biggest problem was that you can't simply output a 2D array if you're fitting image data, you have to do something like the process outlined http://www.scipy.org/Cookbook/FittingData#head-11870c917b101bb0d4b34900a0da… (which I had a surprisingly difficult time finding - maybe links to the cookbooks could be put in the documentation?). Basically, my issue wasn't putting in the errors vs. the function, but that I didn't ravel() my data. I have come across a new problem in my leastsq fitting when trying to pass in my own analytic derivative/gradient function, which I finally discovered after a few days of hunting is because a patch has not been applied to the latest version of scipy, specifically this one: http://article.gmane.org/gmane.comp.python.scientific.devel/5848 . I'm in the process of applying the patch and recompiling all of scipy with it. Is there any reason the patch wasn't applied / should I be asking the scipy-dev list this question? Thanks for the help, Adam

1 0

FFTW use
by Charrett, Thomas 11 Mar '08

11 Mar '08

Hello, I'm fairly new to using python, having converted from matlab and I have some questions concerning the FFT functions in scipy.fftpack. I have some matlab code which calculates the fft of an image (2d array) before carrying out some processing and then doing the inverse transform. However the same code converted to Python gives a different result. I've been trying to track down the problem and it appears that the problem occurs when calculate the fft, with matlab calculating slightly different values than scipy. I've have tried to build scipy to use the FFTW library that Matlab uses and think I have succeeded (scipy.show_config() says fftw3 is available). Does fft2 and fftn automatically use FFTW when installed? If not how would I set this? Are the functions in scipy.fftpack the correct ones? - I have seen post suggesting scipy.fftw which doesn't appear to exist. Also I say I've tried to build scipy with FFTW as the instructions for windows are unclear and it took my a while to get it working - could someone post some instructions on the correct way to compile scipy with FFTW. I used MinGW and MSYS to compile FFTW using the configure and make files, copied libfftw3.a and fftw3.h to a new folder. Downloaded the prebuilt atlas version from the scipy webpage and used the following in my site.cfg file: [atlas] library_dirs = C:\build\atlas3.6.0_WinNT_P3 atlas_libs = lapack, f77blas, cblas, atlas [fftw3] library_dirs = [path to fftw files] include_dirs = [path to fftw libs] libraries = fftw3 I then ran and installed the result using python setup.py config --compiler=mingw32 build --compiler=mingw32 bdist_wininst When I import scipy, scipy.test() crashed but the fftpack.test() works fine, However my fft results look exactly the same am I actually using FFTW? Any help greatly appreciated, I'd really like to ditch matlab completely. Tom

2 1

illegal instruction error in scipy.linalg.decomp.svd
by Lev Givon 11 Mar '08

11 Mar '08

On a Pentium 4 Linux box running python 2.5.1, scipy 0.6.0, numpy 1.0.4, and lapack 3.0, I recently noticed that scipy.linalg.decomp.svd() fails (and causes python to crash) with an "illegal instruction" error. A bit of debugging revealed that the error occurs in the line lwork = calc_lwork.gesdd(gesdd.prefix,m,n,compute_uv)[1] in scipy/linalg/decomp.py. Interestingly, scipy 0.5.2.1 on the same box (with all of the other software packages unchanged) does not exhibit this problem. Moreover, when I install scipy 0.6.0 along with all of the other packages on a Linux machine containing a Pentium D CPU, I do not observe the problem either. Being that I am running Mandriva Linux, closed scipy bug 540 caught my eye. I'm not sure how it could be related to the above problem, though (and I also do not know what the lapack patch mentioned in the ticket could have been - even though I have been maintaining Mandriva's lapack lately :-). Thoughts? L.G.

1 2

citing numpy, scipy openopt
by Emanuele Olivetti 11 Mar '08

11 Mar '08

Hi Numpy, Scipy and OpenOpt people, I'd like to cite these three projects in a scientific paper. According from what I've read in the lists these are the main references for Numpy: 1) Travis E. Oliphant, "Python for Scientific Computing," Computing in Science & Engineering, vol. 9, no. 3, May/June 2007, pp. 10-20. 2) David Ascher, Paul F. Dubois, Konrad Hinsen, Jim Hugunin, Travis Oliphant, "Numerical Python", tech. report UCRL-MA-128569, Lawrence Livermore National Laboratory, 2001; http://numpy.scipy.org. 3) Travis E. Oliphant (2006) Guide to NumPy, Trelgol Publishing, USA I guess the most appropriate reference would be (1) since it introduces the reader to both Python and Numpy. What do you suggest for Scipy? About OpenOpt: I'm using it for non-linear problems (NLP), mainly with "scipy_cg" and "ralg" algorithms. Then I believe I should cite both OpenOpt and Naum Z. Shor's work (ralg). Dmitrey, what do you suggest? Best Regards, Emanuele

4 7

Problem with plot
by Natali Melgarejo Diaz 11 Mar '08

11 Mar '08

Hello everyone, How can i plot an sparse matrix with complex numbers? It seems to be a problem for sparse matrix to plot the values with a simple 'plot'. Thanks in advance.

4 3