Mailman 3 October 2010 - NumPy-Discussion

Atlas build issues
by Gael Varoquaux Oct. 20, 2010

Oct. 20, 2010

I am really sorry to be landing on the mailing list with Atlas build issues. I usually manage be myself to build Atlas, but this time I have been fighting for a couple of days with little success. The reason I need to build Atlas is that our work computers are stuck on Mandriva 2008.0, in which the version of Atlas packaged by the system is not usable. Anyhow, I lost quite a while with Atlas 3.9.26, for which I was never able to build libaries that could be used in a '.so' (seemed that the -fPIC was not working, but really I don't understand what was going on in the maze of makefiles). After that I switched to 3.9;25, which I have already have gotten working on another system (Mandriva 2008, but 32 bits). Know everything builds, but I get a missing symbol at the import: from numpy.linalg import lapack_lite ImportError: /volatile/varoquau/dev/numpy/numpy/linalg/lapack_lite.so: undefined symbol: zgesdd_ I don't have g77 installed on the system at all, to avoid gfortran/g77 mixups. I am attaching the stout of 'python setup.py build_ext -i', in case it contains useful information. Would anybody have suggestions with regards to what I am doing wrong? Cheers, Ga�l

4 12

dtype comparison and hashing - bug?
by Matthew Brett Oct. 20, 2010

Oct. 20, 2010

Hi, I have just run into this oddness: In [28]: dt1 = np.dtype('f4') In [29]: dt1.str Out[29]: '<f4' In [30]: dt2 = dt1.newbyteorder('<') In [31]: dt2.str Out[31]: '<f4' In [32]: dt1 == dt2 Out[32]: True In [33]: hash(dt1) == hash(dt2) Out[33]: False This is the same as: http://www.mail-archive.com/numpy-discussion@scipy.org/msg13299.html My question was - does the team still agree this is a bug? Can anyone offer a pointer as to how it should be fixed? Best, Matthew

3 3

how to ask numpy what bit version it is running (32/64)
by Vincent Davis Oct. 20, 2010

Oct. 20, 2010

What is the best/good way to know what version of numpy is running (32 or 64 bit). Showing my ignorance maybe but does it always match the python version that is running, which can be ask/tested using sys.maxint? -- Thanks Vincent Davis

4 5

patch for structured array comparison bug
by Mark Wiebe Oct. 19, 2010

Oct. 19, 2010

I found a bug in structured array comparison when fields have multi-dimensional types. I created a ticket here: http://projects.scipy.org/numpy/ticket/1640 and a patch here: http://github.com/m-paradox/numpy/compare/master...fix_structured_compare Could someone review it for me? Thanks, Mark

3 5

in1d for structure arrays
by Jeremy Lewi Oct. 19, 2010

Oct. 19, 2010

Hi, I noticed that the recent numpy (1.5) implements the function in1d in set operations. Unfortunately this doesn't seem to work for structure arrays because "mergesort" isn't implemented for structure arrays. I was able to implement a quick work-around simply by adding a structure field to keep track of which list the data came from and then making this field the lowest priority in the sort. Is this worth creating a ticket for? (I don't have trac privileges to do so) Jeremy Jeremy Lewi Engineering Scientist The Intellisis Corporation jlewi(a)intellisis.com

1 0

stride tricks for cellular automata
by Robert Cimrman Oct. 19, 2010

Oct. 19, 2010

FYI: I have added a new Cookbook entry, see [1], hoping that it might be useful to someone. The stride tricks are awesome! r. [1] http://scipy.org/Cookbook/GameOfLifeStrides

2 1

numpy installation in ubuntu
by Alex Ter-Sarkissov Oct. 19, 2010

Oct. 19, 2010

hi everyone, i'm very new to ubuntu, now trying to install numpy and wxpython. Having unpacked numpy and running the standard commnad in idle from numpy import * i get the message ImportError: Error importing numpy: you should not try to import numpy from its source directory; please exit the numpy source tree, and relaunch your python intepreter from there. I quit the folder with numpy, relaunched python, but the module wasn't even identified. thanks

5 7

Another merge at github
by Charles R Harris Oct. 18, 2010

Oct. 18, 2010

Here<http://github.com/numpy/numpy/commit/43b05fdd1c0ba1c6686297d16d257985dd986d…>. This looks harmless but it makes the history really ugly. We need to get the word out *not* to do things this way. Chuck

9 15

DataArray fixes
by Lluís Oct. 18, 2010

Oct. 18, 2010

I started writing some fixes for DataArray, but some others need discussion on whether they should be implemented as I describe. To this extent, please see issues at http://github.com/fperez/datarray/issues and comment on them, so that I'll know if I should implement these. Thanks, Lluis -- "And it's much the same thing with knowledge, for whenever you learn something new, the whole world becomes that much richer." -- The Princess of Pure Reason, as told by Norton Juster in The Phantom Tollbooth

4 8

Re: [Numpy-discussion] Need help for the migration from Numeric to numpy
by Frank Thommen Oct. 18, 2010

Oct. 18, 2010

Bruce Southey wrote: > [...] >>> It took me a while to track it down, but there used to be a module/file >>> called 'convertcode.py' for conversion. It was mentioned as tip in >>> Travis's 'Guide to Numpy' page 31 in Jan 6 2005 version. But the >>> alter_code path is in Dec 7 2006 (linked at http://docs.scipy.org/doc/). >>> It is advisable to read it as it address differences between Numeric and >>> old numpy ie section 2.6! >> Thanks for the hint. I'm currently reading through this document. I'm >> sure this will help a lot. >> >> convertcode.py doesn't seem to provided with current numpy releases. >> >> >>> Travis said in the thread: >>> http://mail.scipy.org/pipermail/numpy-discussion/2006-August/010085.html >>> "The transition approach is to use the compatibility layer first by >>> running oldnumeric.alter_code1.py and then running alter_code2.py which >>> will take you from the compatibility layer to NumPy (but alter_code2 is >>> not completed yet)." >> My experience is, that these scripts don't work at all (any more). >> >> >> frank >> > Hi, > Yeah I never had much success with using such codes. > > Anyhow, I have attached the last version of convertcode.py I have (came > with numpy 0.9.6). If you want that numpy version as does not seem to be > on sourceforge please let me know (the gzipped tar file is about 930K). > > Bruce > Hi, thanks a lot for digging this out, but `python convertcode.py ./test.py` only adds a line '## Automatically adapted for numpy Oct 18, 2010 by convertcode.py' on top of the file. That's all. All other Numeric code in the script (including impoerts) are unchanged. I think it's the most timesaving and nerve-conserving way to let the user change his scripts manually (with the help of Travis's 'Guide to Numpy') and providing Numeric in the meantime ;-) Thanks to all respondents for the hints and links frank

1 0