Mailman 3 April 2009 - NumPy-Discussion

numpy.int32 -> double?
by Bruce Sherwood 28 Apr '09

28 Apr '09

I'm not sure this is the right forum for my question; it's quite possible that this is a Boost question. VPython (vpython.org) is the name of Python plus the 3D Visual module, which is mostly written in C++ and connected to Python using the Boost libraries. Visual imports numpy. When I build Visual for Python 2.6 and numpy 1.3.0, I get a failure on something that works on Python 2.5 with numpy 1.2.1. There is a vector(double,double,double) class in Visual which now fails if handed numpy.int32 arguments (works fine with numpy.float64 arguments). I was using Boost 1.35 with Python 2.5 and am using Boost 1.38 with Python 2.6. Any ideas or suggestions? Thanks. Bruce Sherwood

1 0

patching docs
by william ratcliff 28 Apr '09

28 Apr '09

Hi! I'd like to suggest a patch for: numpy <http://docs.scipy.org/numpy/docs/numpy/>.core<http://docs.scipy.org/numpy/docs/numpy.core/> .fromnumeric <http://docs.scipy.org/numpy/docs/numpy.core.fromnumeric/>.put The docstring contains: for i, val in zip(ind,v): x.flat[i]=val It should be: for i, val in zip(ind,v): a.flat[i]=val This would be more consistent with the rest of the docstring. For the future, is there an easy way for us to patch docstrings? Thanks, William

5 10

Some minor issues with numpy and fractions
by Michael S. Gilbert 27 Apr '09

27 Apr '09

I had mentioned recently some interest in using fractions in the numpy polynomial class. Suprisingly, it actually works for the most part out of the box, which is great. However, there are some minor issues. For example: >>>> numpy.poly1d( [ fractions.Fraction(1,2) , fractions.Fraction(1,8) ] )/fractions.Fraction(3,4) fails, so I have to move the division into the inner list, which leads to a lot of duplicate typing and general messiness: >>>> numpy.poly1d( [ fractions.Fraction(1,2)/fractions.Fraction(3,4) , fractions.Fraction(1,8)/fractions.Fraction(3,4) ] ) Another item of general interest for numpy is that there is no simple way to allocate memory for fraction array. >>>> def zeros( nsamples ): >>>> list = [] >>>> for n in range( nsamples ): >>>> list.append( fractions.Fraction( 0 ) ) >>>> return numpy.array( list ) is a solution, but it would be nice to have this built-in, e.g.: >>>> numpy.zeros( n , dtype=numpy.fraction ) Perhaps developing a fraction dtype would solve the above poly1d issue as well? Anyway, just some thoughts to ponder. Best wishes, Mike

3 3

help with vectorization
by Mathew Yeates 27 Apr '09

27 Apr '09

I know this must be trivial but I can't seem to get it right I have N 2x2 arrays which perform a rotation. I also have N xy pairs to transpose. What is the simplest way to perform the transformation without looping? Thanks from someone about to punch their screen.

2 2

after building from source, how to register numpy with synaptic?
by Chris Colbert 25 Apr '09

25 Apr '09

Like the subject says, is there a way to register numpy with synaptic after building numpy from source? I would like to snag matplotlib from the ubuntu repos, but it won't let me without also getting numpy and its dependency, which would ruin all the work I did building numpy on my machine. This is probably a linux n00b question, but well, that's what I am :) Chris

6 5

Idea: fractional polynomial class
by Michael S. Gilbert 25 Apr '09

25 Apr '09

Hello, I've been working with numpy's poly1d class recently, and it would be very useful to me if the class worked with fractions instead of floats (since I'm encountering quantities that often mostly cancel out, which lead to instabilities in my algorithms; hence it would be useful to use fractions that don't lose precision). Since python 2.6 introduced a "fractions" class, it should be fairly straightforward to implement/reimplement the polynomial class with support for fractions. I'm curious as to whether there would be any interest from the project on this and whether you all would be open to patches/code contributions. Best Regards, Mike

6 8

Re: [Numpy-discussion] patching docs
by Joe Harrington 24 Apr '09

24 Apr '09

> 2009/4/24 william ratcliff <william.ratcliff(a)gmail.com>: > > Actually, if I look here: > > http://docs.scipy.org/numpy/docs/numpy.core.fromnumeric.put/ > > > > The text that appears in by browser is: > > > > put(a, ind, v, mode='raise') > > > > Changes specific elements of one array by replacing from another array. > This is pretty cryptic. I've changed it to > """ > put(a, ind, v, mode='raise') > > Replaces specified elements of an array with given values. > """ > Your error report was correct. I've changed the x to an a. Ok, I thought you were complaining that the a and x in the examples were not consistently using the same variable (which is not super-important but Ralf fixed it anyway). Now I see that you were saying that the variable name in parameters was not the same as that in the example. Good catch. I've now made the examples all use a. Also, the example in the extended summary was very unpythonic, so I removed the loop. I think that's ok but someone tell me if it's not, or just go fix it. I did test it on a case like one of the examples and it worked. Thanks, --jh--

1 0

Re: [Numpy-discussion] patching docs
by Joe Harrington 24 Apr '09

24 Apr '09

william ratcliff <william.ratcliff(a)gmail.com> writes: > Hi! I'd like to suggest a patch for: > numpy <http://docs.scipy.org/numpy/docs/numpy/>.core<http://docs.scipy.org/numpy/docs/numpy.core/> > .fromnumeric <http://docs.scipy.org/numpy/docs/numpy.core.fromnumeric/>.put > The docstring contains: > for i, val in zip(ind,v): > x.flat[i]=val > It should be: > for i, val in zip(ind,v): > a.flat[i]=val Thanks! It appears you are basing your patch on a numpy release rather than the wiki you have actually linked to, as this problem was fixed on 2009-03-16 by Ralf Gommers. > This would be more consistent with the rest of the docstring. For the > future, is there an easy way for us to patch docstrings? Sure! Follow the instructions on docs.scipy.org/numpy/ and make yourself a doc wiki account. Post the account name to the scipy-dev mailing list and ask to be made an editor. Then please edit! But, please be careful that you develop any changes based on the current docstrings and not something from a release. Because the releases are snapshots of a fast-moving project, they are often out of date (In fact, I tell my students to use the doc wiki as their best source of docs, and to ignore help()). Thanks again, --jh--

3 3

Keeping track of backports in Trac
by Pauli Virtanen 24 Apr '09

24 Apr '09

Hi, Is it possible to assign a bug to multiple versions in Trac? Now that we have active stable branches eg. 0.7.x in Scipy, it would be useful if it was easy to track what to backport. For example, having a Trac custom field "Pending backports", either a boolean or a string listing the branches, would help this. -- Pauli Virtanen

1 0

numpy 1.3.0 test() failures on ubuntu 9.04
by Chris Colbert 24 Apr '09

24 Apr '09

Hi all, I'm building numpy 1.3.0 from source with libatlas-sse2-dev from the jaunty repos. I'm running into 16 failures when running the nose test. This is a fresh install of 9.04 and i've repod the following packages: build-essential swig gfortran python-dev libatlas-sse2-dev libatlas-base-dev (the atlas packages include lapack and blas libraries) The pastebin links to site.cfg build.log and test.log are at the end of this email. If anyone could help me out here as to what the problems may be, i would appreciate it! Thanks! Chris site.cfg http://ubuntu.pastebin.com/m671e9454 build.log http://ubuntu.pastebin.com/m581eee89 test.log http://ubuntu.pastebin.com/m7b63f6a0

4 4