Mailman 3 December 2009 - NumPy-Discussion

Help to port numpy to python3?
by Xavier Gnata Dec. 7, 2009

Dec. 7, 2009

Hi, Is there a way to help to port numpy to python3? I don't thing I have time to rewrite some code but I can test whatever has to be tested. Is there an official web page showing the status of this port? Same question from scipy? It is already nice to see that the last numpy version is compatible with python2.6 :) Xavier

3 2

SIGFPE handling in ufunc
by David Cournapeau Dec. 7, 2009

Dec. 7, 2009

Hi, While investigating why np.array([0]) ** -1/-1 crashes, I wondered about how signal handling is supposed to work in ufuncs when a SIGFPE is raised. In particular, the following is puzzling: import numpy as np x = np.array([1, 2, 3, 4]) / 0 # x is now array([0, 0, 0, 0]) Note that no warning is raised - this seems to be a regression, since I came across e.g. this ticket: http://projects.scipy.org/numpy/ticket/541. cheers, David

1 1

non-standard standard deviation
by Dr. Phillip M. Feldman Dec. 7, 2009

Dec. 7, 2009

All of the statistical packages that I am currently using and have used in the past (Matlab, Minitab, R, S-plus) calculate standard deviation using the sqrt(1/(n-1)) normalization, which gives a result that is unbiased when sampling from a normally-distributed population. NumPy uses the sqrt(1/n) normalization. I'm currently using the following code to calculate standard deviations, but would much prefer if this could be fixed in NumPy itself: def mystd(x=numpy.array([]), axis=None): """This function calculates the standard deviation of the input using the definition of standard deviation that gives an unbiased result for samples from a normally-distributed population.""" xd= x - x.mean(axis=axis) return sqrt( (xd*xd).sum(axis=axis) / (numpy.size(x,axis=axis)-1.0) ) -- View this message in context: http://old.nabble.com/non-standard-standard-deviation-tp26566843p26566843.h… Sent from the Numpy-discussion mailing list archive at Nabble.com.

14 22

another numpy/ATLAS problem
by David Warde-Farley Dec. 7, 2009

Dec. 7, 2009

It seems like I've done this enough times on enough machines but I'm still running into problems... After installing what I'm pretty sure is a complete ATLAS+LAPACK and building numpy against it (appropriately editing site.cfg, etc.) I'm faced with this when building numpy-1.3.0: dwf@mirage:~/tmp$ python Python 2.6.4 (r264:75706, Nov 2 2009, 14:44:17) [GCC 4.4.1] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import numpy Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/home/dwf/virtualenv/stable/lib/python2.6/site-packages/numpy/ __init__.py", line 130, in <module> import add_newdocs File "/home/dwf/virtualenv/stable/lib/python2.6/site-packages/numpy/ add_newdocs.py", line 9, in <module> from lib import add_newdoc File "/home/dwf/virtualenv/stable/lib/python2.6/site-packages/numpy/ lib/__init__.py", line 13, in <module> from polynomial import * File "/home/dwf/virtualenv/stable/lib/python2.6/site-packages/numpy/ lib/polynomial.py", line 18, in <module> from numpy.linalg import eigvals, lstsq File "/home/dwf/virtualenv/stable/lib/python2.6/site-packages/numpy/ linalg/__init__.py", line 47, in <module> from linalg import * File "/home/dwf/virtualenv/stable/lib/python2.6/site-packages/numpy/ linalg/linalg.py", line 22, in <module> from numpy.linalg import lapack_lite ImportError: /home/dwf/virtualenv/stable/lib/python2.6/site-packages/ numpy/linalg/lapack_lite.so: undefined symbol: dlamc3_ FWIW these are the files that ATLAS has installed: dwf@mirage:~/Downloads$ ls /opt/sw/ATLAS/lib/ -al total 23472 drwxr-xr-x 2 root root 4096 2009-11-26 20:36 . drwxr-xr-x 4 root root 4096 2009-11-26 20:36 .. -rw-r--r-- 1 root root 12963138 2009-11-26 20:36 libatlas.a -rw-r--r-- 1 root root 466244 2009-11-26 20:36 libcblas.a -rw-r--r-- 1 root root 572812 2009-11-26 20:36 libf77blas.a -rw-r--r-- 1 root root 8979016 2009-11-26 20:36 liblapack.a -rw-r--r-- 1 root root 466678 2009-11-26 20:36 libptcblas.a -rw-r--r-- 1 root root 573196 2009-11-26 20:36 libptf77blas.a liblapack.a looks big enough for it to be a complete LAPACK. ATLAS (3.9.17) was configured with the following: ../configure -Ss kern /usr/bin/gcc-4.2 --prefix=/opt/sw/ATLAS --with- netlib-lapack-tarfile=/home/dwf/tmp/lapack-lite-3.1.1.tgz -Fa alg - fPIC -D c -DPentiumCPS=2668 -t 4 -b 64 I tried a few different things, including building dylibs, in which case I get an undefined symbol error for ATL__zttrsm. Any help would be appreciated. To be honest I'm not sure why lapack_lite is getting involved at all; I thought that was what got used in case of no detected LAPACK, but config definitely does detect ATLAS/LAPACK. David

4 11

histogram for discrete data
by Ernest Adrogué Dec. 6, 2009

Dec. 6, 2009

Hi, A few weeks ago there was a discussion about a histogram_discrete() function --sorry for starting a new thread but I have lost the mails. Somebody pointed out that bincount() already can be used to histogram discrete data (except that it doesn't work with negative values). I have just discovered a function in scipy.stats called itemfreq() that does handle negative values. In [17]: scipy.stats.itemfreq([-1,-1,0,5]) Out[17]: array([[-1., 2.], [ 0., 1.], [ 5., 1.]]) Bye.

2 1

How to solve homogeneous linear equations with NumPy?
by Peter Cai Dec. 6, 2009

Dec. 6, 2009

How to solve homogeneous linear equations with NumPy? If I have homogeneous linear equations like this array([[-0.75, 0.25, 0.25, 0.25], [ 1. , -1. , 0. , 0. ], [ 1. , 0. , -1. , 0. ], [ 1. , 0. , 0. , -1. ]]) And I want to get a non-zero solution for it. How can it be done with NumPy? linalg.solve only works on A * x = b where b does not contains only 0. -- look to the things around you,the immediate world around you, if you are alive,it will mean something to you ——Paul Strand

5 8

Re: [Numpy-discussion] Numpy/Scipy for EC2
by Dan Yamins Dec. 5, 2009

Dec. 5, 2009

On Sat, Dec 5, 2009 at 1:32 AM, David Goldsmith <d.l.goldsmith(a)gmail.com>wrote: > Dan- > > I almost hate to ask - after what you've already provided, which is > substantial: did you ever try out what alestic had to offer? If not, I may > be that guinea pig. ;-) > > David -- I'm posting my response to your question to the group as well ... I never tried out the alestic AMIs *directly* ... that's because at least as far as I could tell (I might be wrong about this), they didn't have blas, lapack, scipy, etc prebuilt on them. But the whole StarCluster thing *is* basically Alestic fundamentally, since the AMIs that StarCluster provides were themselves built directly from Alestic's 32-bit and 64-bit Ubuntu jaunty AMIs ... with comparatively small but crucial modifications (building optimized blas, scipy, the Sun Grid Engine, NFS &c). From a "systems" point of view should basically have the same stability properties. The real strength of starcluster is that, in addition to those small but crucial additions to the Alestic AMIs, Starcluster also provides a simple command for making startup/admin/termination of a multi-node cluster really easy. With one invocation, it will interface with EC2 commands to start up the relevant machine instances, mount a shared NFS drive from an amazon EBS volume, start up the grid engine, etc... so that all the headaches of setting up and breaking down the pieces of a multi-node shared-memory cluster are completely handled for you. Is there a particular reason you had in mind for using Alestic directly? I'd be interested to hear your use case ... Dan Btw -- I got an email from Justin Riley, the developer of starcluster -- he seems to be hard at work on the next release (which will have hadoop, couchdb, and some important improvements to the management of multiple clusters. still no word on dynamic load balancing tho)

2 1

ANN: upfirdn 0.1.0
by Tom K. Dec. 5, 2009

Dec. 5, 2009

(also posted on scipy-user) ANNOUNCEMENT I am pleased to announce the initial release of "upfirdn." This package provides an efficient polyphase FIR resampler object (SWIG-ed C++) and some python wrappers. https://opensource.motorola.com/sf/projects/upfirdn MOTIVATION As a long time user of MATLAB and the Signal Processing Toolbox, I have missed an "upfirdn" analogue in numpy / scipy since I switched over to python a couple years ago. I've been looking for a way to contribute to the wider numpy / scipy community since I love these tools and appreciate all the efforts of those who have developed them. Since we have the polyphase resampling functionality within Motorola and I use it for my work, I thought it best to go the "official route" and get approval from my company to publish it under a BSD compatible license. NOTES TO USERS AND REGARDING INSTALL It is my hope that others find this functionality useful. Suggestions for improvements or bug reports are welcome. As installation is an area that I am very green, I recommend only installing it locally for now, e.g. python setup.py build_ext -i or even just make I'd like to spend some time learning more about distutils and release a version with a more user-friendly install. -- View this message in context: http://old.nabble.com/ANN%3A-upfirdn-0.1.0-tp26656025p26656025.html Sent from the Numpy-discussion mailing list archive at Nabble.com.

1 0

Import numpy fails on cygwin python
by Olivia Cheronet Dec. 4, 2009

Dec. 4, 2009

Hello! I have built numpy (updated to the trunk) for my cygwin (1.5.25) Python (2.5.2). However, testing fails when I try to import numpy in python (see output below). I have been searching around for a solution, but everything has failed so far... I would be grateful for any advice. Thank you, Olivia $ python Python 2.5.2 (r252:60911, Dec 2 2008, 09:26:14) [GCC 3.4.4 (cygming special, gdc 0.12, using dmd 0.125)] on cygwin Type "help", "copyright", "credits" or "license" for more information. >>> import numpy Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/usr/lib/python2.5/site-packages/numpy/__init__.py", line 132, in <modul e> import add_newdocs File "/usr/lib/python2.5/site-packages/numpy/add_newdocs.py", line 9, in <modu le> from lib import add_newdoc File "/usr/lib/python2.5/site-packages/numpy/lib/__init__.py", line 13, in <mo dule> from polynomial import * File "/usr/lib/python2.5/site-packages/numpy/lib/polynomial.py", line 17, in < module> from numpy.linalg import eigvals, lstsq File "/usr/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 47, in <module> from linalg import * File "/usr/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 22, in <m odule> from numpy.linalg import lapack_lite ImportError: No such file or directory

3 5

Three bugs fixed
by Travis Oliphant Dec. 4, 2009

Dec. 4, 2009

Thanks to the reporters of tickets #1108, #1197 (similar to #1279), and #1222. Pointing out these problems allowed me to find and squash two subtle memory leaks and one just plain stupid bug lurking in reduce-at (when using the buffered internal loop). I think I fixed the problems. I added a test for #1108, but I don't have good tests for the memory leak fixes. I'm not sure how to write a good test for those. Even though it may take several months to get to the tickets with my limited time, please do continue to report problems. I'm hoping to have more time for NumPy next year. Best regards, -Travis

3 9