Mailman 3 July 2008 - NumPy-Discussion

Instructions on building from source
by Fernando Perez July 24, 2008

July 24, 2008

Howdy, I was just trying to explain to a new user how to build numpy from source on ubuntu and I realized that there's not much info on this front in the source tree. Scipy has a nice INSTALL.txt that even lists the names of the debian/ubuntu packages needed for the build (which can be a useful guide on other distros). Should we have a stripped-down copy of this doc somewhere in the top-level directory of numpy? It's also disconcerting for a new user that there's no visible documentation directory at the top, perhaps listing the existence of numpy/doc in the readme could help. I know there's plenty of work being done right now on the docs, so perhaps sprucing up the 'out of the tarball' experience for those building from source could be done with a very reasonable amount of effort. Cheers, f

5 10

Schedule for 1.1.1
by Charles R Harris July 24, 2008

July 24, 2008

I think we should try to get a quick bug fix version out by the end of the month. What do others think? Chuck

6 12

f2py - a recap
by Fernando Perez July 24, 2008

July 24, 2008

Howdy, today's exercise with f2py left some lessons learned, mostly thanks to Robert's excellent help, for which I'm grateful. I'd like to recap here what we have, mostly to decide what changes (if any) should go into numpy to make the experience less "interesting" for future users: - Remove the f2py_options flag from numpy.distutils.extension.Extension? If so, do options like '--debug_capi' get correctly passed via setup.cfg? This flag is potentially very confusing, because only *some* f2py options get actually set this way, while others need to be set in calls to config_fc. - How to properly set the compiler options in a setup.py file? Robert suggested the setup.cfg file, but this doesn't get picked up unless config_fc is explicitly called by the user: ./setup.py config_fc etc... This is perhaps a distutils problem, but I don't know if we can solve it more cleanly. It seems to me that it should be possible to provide a setup.py file that can be used simply as ./setup.py install (with the necessary setup.cfg file sitting next to it). I'm thinking here of what we need to do when showing how 'easy' these tools are for scientists migrating from matlab, for example. Obscure, special purpose incantations tend to tarnish our message of ease :) - Should the 'instead' word be removed from the f2py docs regarding the use of .pyf sources? It appears to be a mistake, which threw at least me for a loop for a while. - Why does f2py in the source tree have *both* a doc/ and a docs/ directory? It's really confusing to see this. f2py happens to be a very important tool, not just because scipy couldn't build without it, but also to position python as a credible integration language for scientific work. So I hope we can make using it as easy and robust as is technically feasible. Cheers, f

4 5

Re: [Numpy-discussion] Reference guide updated
by Neil Crighton July 23, 2008

July 23, 2008

Ok, thanks. I meant the amount of vertical space between lines of text - for example, the gaps between parameter values and their description, or the large spacing between both lines of text and and the text boxes in the examples section. If other people agree it's a problem, I thought the spacing could be tweaked. It's not a problem that there's more than one function per page. I have been helping out with the docs (where I feel I'm qualified enough - I'm no numpy expert!). I think it will make numpy much easier to learn to have easily-accessible, comprehensive docstrings, and the documentation editor makes it very easy to contribute. I'm also learning a lot reading other people's docstrings :) > > It there a reason why there's so much vertical space between all of the > > text sections? I find the docstrings much easier to read in the editor: > > Roughly: > * In the editor, you have one function per page, vs several function per page > in the reference > * in the editor, the blocks 'Parameters', 'Returns'... are considered as > sections, while in the reference, they are Field lists (roughly). > > In the end, it's only a matter of taste, of course. But you raise an > interesting point, we should provide some kind of options to choose between > behaviors. > > Please note as well that the reference guide is a work in progress: you're > more than welcome to join and work with us.

2 1

Re: [Numpy-discussion] Schedule for 1.2.0
by Joe Harrington July 23, 2008

July 23, 2008

Hi Jarrod, I'm just catching up on my numpy lists and I caught this; sorry for the late reply! > Another issue that we should address is whether it is OK to postpone > the planned API changes to histogram and median. A couple of people > have mentioned to me that they would like to delay the API changes to > 1.3, which seems reasonable to me. If anyone would prefer that we > make the planned API changes for histogram and median in 1.2, please > speak now. I *strongly* want both these changes for 1.2, as I am sure do the many people teaching courses using numpy for the fall. It is hard to get students to understand why there are inconsistencies and irrationalities in software, and it's even worse when it's open-source, since somehow it's the lecturer's fault that he picked a package that isn't right in some major way. Worse, we're changing these behaviors like 6 months from now, so students will have to learn it wrong and code it wrong, and then their code may break on top of it. On behalf of this year's new students and their instructors, I ask you to keep these changes in the release as planned. Thanks, --jh-- Prof. Joseph Harrington Department of Physics MAP 414 4000 Central Florida Blvd. University of Central Florida Orlando, FL 32816-2385

3 2

integer array creation oddity
by Suchindra Sandhu July 23, 2008

July 23, 2008

Hi, Can someone please explain to me this oddity? In [1]: import numpy as n In [8]: a = n.array((1,2,3), 'i') In [9]: type(a[0]) Out[9]: <type 'numpy.int32'> In [10]: type(a[0]) == n.int32 Out[10]: False When I create an array with 'int', 'int32' etc it works fine What is the type of 'i' and what is n.int0? Thanks, Suchindra

3 5

Documenting chararrays
by Stéfan van der Walt July 23, 2008

July 23, 2008

Hi all, Should we document character arrays? Does anybody still use them? I think their behaviour can largely be duplicated by object arrays. They also seem to be broken: >>> x = np.array(['1', '2', '3', '4']).view(np.chararray) >>> x*3 chararray(['111', '222', '333', '444'], dtype='|S4') All good, but: >>> x = np.array(['12', '34', '56', '78']).view(np.chararray) >>> x * 3 chararray(['1212', '3434', '5656', '7878'], dtype='|S4') Whereas with object arrays: >>> np.array(['12', '34', '56', '78'], dtype=object) * 3 array([121212, 343434, 565656, 787878], dtype=object) Similaryly: >>> x.rjust(3) chararray([' a', ' b', ' c', ' d'], dtype='|S3') >>> x = np.array(['a','b','c','d']).view(np.chararray) >>> x.rjust(3) chararray([' a', ' b', ' c', ' d'], dtype='|S3') But then >>> x = np.array([['ab','cd'], ['ef','gh']]).view(np.chararray) >>> x.rjust(5) # BOOM Cheers Stéfan

3 2

Benchmarking code
by Alan McIntyre July 23, 2008

July 23, 2008

There's a function (_test_unique1d_speed) in numpy/lib/arraysetops.py that looks to me like it should be in a benchmark suite instead of in the library module. Would anyone mind if I moved it to numpy/lib/benchmarks?

2 2

Representation of array subclasses
by Stéfan van der Walt July 23, 2008

July 23, 2008

Hi all, I noticed that subclasses are not represented correctly: In [8]: np.chararray([1, 2, 3]) Out[8]: chararray([[['\x03', '', ''], ['\xc0', '\x03', '']]], dtype='|S1') Notice how the indentation of the second row is completely wrong. I tried to fix this in array_repr_builtin (arrayobject.c:4318), but the fix had no impact. `array_repr_builtin` is not called unless preceded by np.set_string_function(None, repr=1) Does anybody know where the right place is to fix the representation? Thanks Stéfan

1 0

Reference guide updated
by Stéfan van der Walt July 22, 2008

July 22, 2008

Hi all, A new copy of the reference guide is now available at http://mentat.za.net/numpy/refguide/ and http://mentat.za.net/numpy/refguide/NumPy.pdf I'd like to thank Pauli Virtanen, who put in a lot of effort to improve Sphinx interaction and document layout. The guide is not yet complete, but we are working on ironing out remaining problems. As always, we could benefit from more content, so please continue writing and reviewing! Regards Stéfan

3 2