Mailman 3 May 2012 - SciPy-User

How to assemble large sparse matrices effectively
by cp3028 09 May '12

09 May '12

Hello everyone, I am working on an FEM project using Scipy. Now my problem is, that the assembly of the sparse matrices is to slow. I compute the contribution of every element in dense small matrices (one for each element). For the assembly of the global matrices I loop over all small dense matrices and set the matrice entries the following way: ... [i,j] = someList[k][l] Mglobal[i,j] = Mglobal[i,j] + Mlocal[k,l] ... Mglobal is a lil_matrice of appropriate size, someList maps the indexing variables. Of course this is rather slow and consumes most of the matrice assembly time. Is there a better way to assemble a large sparse matrix from many small dense matrices? I tried scipy.weave but it doesn't seem to work with sparse matrices Yours, cm

1 0

Linux compile questions for SciPy
by Crooks, Cynthia J 09 May '12

09 May '12

I am trying to compile SciPy for our Linux systems. I have been able to compile NumPy and ATLAS. These have been installed in user specified directories, /hpc/soft/NumPy and /hpc/soft/atlas, using the -prefix option. When I try to build SciPy it seems to be picking up an older version of numpy installed in the system directories. How do I point the SciPy build to look for the version I have compiled? OS SLES SP 1 uname -a Linux hpcp8002 2.6.32.12-0.7-default #1 SMP 2010-05-20 11:14:20 +0200 x86_64 x86_64 x86_64 GNU/Linux numpy version 1.6.1 **The version I compiled numpy version 1.3.0 **Version install in system directories scipy version 0.10.1 atlas version 3.9.74 python version 2.6 I have attached the screen output from the scipy build attempt. Any help would be appreciated. Thank you. Cindy <<scipy.build.log>>

2 1

scipy.integrate.dblquad with variable boundary
by Guy Moïse Dongho Nguimdo 08 May '12

08 May '12

I am trying to use dblquad to do double integration, but one of the boundaries of the integrale is also a variable. This is the integral: $\int^b_a \int^y_c f(x,y) dydx$ May some one help? Thanks -- GM DONGHO NGUIMDO --------------------------------------------------------------------------------------------------------------------------- "When it is obvious that the goals cannot be reached, don't adjust the goals, adjust the action steps." *-Confucius*

1 0

scipy.sparse.csgraph merged
by Ralf Gommers 07 May '12

07 May '12

Hi, I finally merged https://github.com/scipy/scipy/pull/119, the sparse graph algorithms module written by Jake Vanderplas. This will certainly by one of the highlights of the 0.11 release (which shouldn't be too far off). Thanks again to Jake for all the work he put in. Ralf

1 0

scipy.stats.distributions.erlang - "boring" consensus building in a ticket
by josef.pktd＠gmail.com 07 May '12

07 May '12

Should we restrict the shape parameter to be an integer instead of a float? http://projects.scipy.org/scipy/ticket/1647 Let's ask the users: Does anyone want an exception if the shape parameter is not an integer? Is there a demand or use case for estimating the shape parameter as an integer instead of a float? right now erlang and gamma are essentially the same, as far as I can see Josef

3 6

Python 2.7 / 64-bit
by Chris Weisiger 06 May '12

06 May '12

I'm in the unfortunate situation of needing to make a OSX 64-bit version of a program that uses both Scipy/Numpy and wxWidgets. The former only has unofficial 64-bit builds and only for Python 2.6; the latter only has 64-bit builds for 2.7 -- before that, they were using Carbon for UI calls, which is 32-bit only. So there's a problem there. Hand-building my own version of either seems to be a pretty gnarly problem that I'd rather avoid if possible. Any ideas how far we are from having a 64-bit, Python2.7 build of scipy and numpy for OSX? All else being equal, I'd rather be on 2.7 than 2.6. Thanks. -Chris

6 8

ndimage/morphology - binary dilation and erosion?
by klo uo 06 May '12

06 May '12

While preparing some images for OCR, I usually discard those with low DPI, but as this happens often I thought to try some image processing and on suggestion (morphological operations) I tried ndimage.morph with idea to play around binary_dilation Images were G4 TIFFs which PIL/MPL can't decode, so I convert to 1bit PNG which I normalized after to 0 and 1. On sample img I applied: ndi.morphology.binary_dilation(img).astype(img.dtype) and ndi.morphology.binary_erosion(img).astype(img.dtype) I attached result images, and wanted to ask two question: 1. Is this result correct? From what I read today seems like what dilation does is erosion and vice versa, but I probably overlooked something 2. Does someone maybe know of better approach for enhancing original sample for OCR (except thresholding, for which I'm aware)? TIA [image: Inline image 1]

3 12

... numpy/linalg/lapack_lite.so: undefined symbol: zungqr_
by Kliment 04 May '12

04 May '12

Hi, I compiled lapack, atlas, umfpack, fftw in local folder, in similar way as described here: http://www.scipy.org/Installing_SciPy/Linux on 32bit Ubuntu Precise In ~/.local/lib I have: ======================================== libamd.2.2.3.a libamd.a -> libamd.2.2.3.a libatlas.a libcblas.a libf77blas.a libfftw3.a libfftw3.la* liblapack.a librefblas.a libsatlas.so* libtmglib.a libumfpack.5.5.2.a libumfpack.a -> libumfpack.5.5.2.a ======================================== In ~/.local/include I have: ======================================== amd.h atlas/ cblas.h clapack.h fftw3.f fftw3.f03 fftw3.h fftw3l.f03 fftw3q.f03 UFconfig.h umfpack.h <various umfpack_[...].h files> ======================================== My site.cfg looks like this: ======================================== [DEFAULT] library_dirs = $HOME/.local/lib include_dirs = $HOME/.local/include [atlas] atlas_libs = lapack, f77blas, cblas, atlas [amd] amd_libs = amd [umfpack] umfpack_libs = umfpack, gfortran [fftw] libraries = fftw3 ======================================== I extracted numpy and run: python setup.py build --fcompiler=gnu95 python setup.py install --prefix=$HOME/.local I then run python interpreter and try to import numpy, when I receive import error: ImportError: /home/vlad/.local/lib/python2.7/site- packages/numpy/linalg/lapack_lite.so: undefined symbol: zungqr_ What did I do wrong?

2 3

scipy.integrate with a function vector
by Jeffrey 04 May '12

04 May '12

Hi, all. My problem came with an integration problem. Below is a brief sample: >>> e1=numpy.arange(10) >>> d1=e1/10 >>> def e2(w): >>> return w*e1 >>> def d2(w): >>> return w*d1 >>> >>> scipy.integrate.quad(lambda x: e1(x)-d1(x), 0 , 1) Since e1,d1 are array with length 10, d2(w) and e2(w) will also each return an array of length 10. However scipy.integrate.quad only supports a scale function. Is there any integrating function in scipy, which could solve this problem? Or my code writes wrongly? Any alternative way to solve this problem? Thanks a lot. -- Jeffrey

1 0

Is there any quantile function in scipy.stats module?
by 袁振飞 04 May '12

04 May '12

Hi, all. My work involved in some probability calculation, which needs the quantile function of distributions like standard normal and t. However, I looked into scipy.stats.t, there's only pdf and cdf functions, and didn't see quantile function. What can I do with this problem? I want your help, thanks a lot ! Best wishes: ) -- Jeffrey

2 2