Jarrod Millman

Since you liked zip, why not to like reduce (altho...

2010-10-13T17:44:12.117-07:00

Since you liked zip, why not to like reduce (although iirc Guido doesn't like it ;-)), so for the actual interleave just use both functions you would like:

*In [1]: reduce(tuple.__add__, zip([1,2],[11,21]))
Out[2]: (1, 11, 2, 21)

?

I would use itertools.chain: from itertools impor...

2010-10-11T13:25:09.604-07:00

I would use itertools.chain:

from itertools import chain

a = list('abcde')
b = list('12345')

list(chain(*zip(a,b)))

['a', '1', 'b', '2', 'c', '3', 'd', '4', 'e', '5']

It can be extended to merge more than two lists:

c = list('XYZZY')
list(chain(*zip(a,b,c)))

If the lists are long, or are iterables, or you want to merge the lists lazily, use iterttools.izip:

from itertools import chain, count, izip

for value in chain(*izip(count(),a+b+c)):
print value

another variation of solution 7: reduce(lambda ou...

2010-10-11T03:52:40.521-07:00

another variation of solution 7:

reduce(lambda out, p: out + [p, p+'64'], paths, list())

Or using a bit from the previous comment (and rela...

2010-10-11T01:41:32.844-07:00

Or using a bit from the previous comment (and relating it to your example):

out = []
map(out.extend,([p, p + '64'] for p in paths))

Here's another simple one (note that you are t...

2010-10-11T01:36:17.672-07:00

Here's another simple one (note that you are throwing away the result of the map statement).

1> A = [1,2,3,4,5]
2> B = [6,7,8,9,10]
3> out = []
4> map(out.extend,zip(A,B))
<4> [None, None, None, None, None]
5> out
<5> [1, 6, 2, 7, 3, 8, 4, 9, 5, 10]

Good thing there's only one way to do it with ...

2010-10-10T15:14:58.410-07:00

Good thing there's only one way to do it with Python... :-)

A variation of solution 6: [item for items in (...

2010-10-10T14:35:55.494-07:00
A variation of solution 6:
[item for items in ([p, p + '64'] for p in paths) for item in items]. No need for the zip just use a generator expression.

I also like solutions 5 as a general solution, tho...

2010-10-10T06:51:49.180-07:00
I also like solutions 5 as a general solution, though stylistically, I prefer "map(out.extend, ([p, p + '64'] for p in paths))" to the for loop. I think yours is the more pythonic solution (and might even be more efficient). I like to use map if all a for loop does is call a function on each element of an iterable, since I think it's a bit clearer and it saves a line of code.

Another solution (a modification of solution 7) is "return sum(([p, p + '64'] for p in paths), [])", which saves an import. I'd probably go with that since it's shorter, and efficiency only matters if the lists are extremely long or this function gets called a lot, neither of which is true in your application.

As far as I understand it from the NumPy mailing l...

2010-07-19T06:40:45.708-07:00
As far as I understand it from the NumPy mailing list, we are not too far away from a release which will deliver support for bothy Python 2.7 and 3.1, see also: http://mail.scipy.org/pipermail/numpy-discussion/2010-July/051436.html

I believe the current development version of numpy...

2010-06-24T10:56:32.927-07:00
I believe the current development version of numpy is compatible with Python 3.

Any word on the status of the 3.0 port?

2010-06-03T12:52:53.803-07:00
Any word on the status of the 3.0 port?

Please take a look at the FOSSEE website. You can...

2009-12-15T23:35:11.961-08:00
Please take a look at the FOSSEE website. You can find videos and slides on their download page:
http://fossee.in/download

As soon as the FOSSEE team gets time, they will add links to slides and video uploads from the conference there as well. I will also make a blog post when this material is available.

Starting at the end of May 2009, Prabhu very qu...

2009-12-15T04:30:57.187-08:00
Starting at the end of May 2009, Prabhu very quickly gathered together an amazing team that immediately created a significant amount of documentation and training materials including tutorials, audio/video demonstrations, written material, and lectures. They've created a great two-day hands-on introductory tutorial to scientific programming with Python and have all ready conducted several of these tutorials all across India. Now they are working on creating a couple of semester long college courses and will be offering the first one next semester at IIT Bombay.

Great! Is this material available online?

I can't find any mention of this on LinkedIn--...

2009-12-06T12:27:42.798-08:00
I can't find any mention of this on LinkedIn-- is it just too soon, just overlooked, ??? I would be glad to add it on your behalf if need be.

FYI, I am one of 5 council reps for the maemo.org linux developers/enthusiasts, and we count numerous Python developers amongst our membership. I am beginning Texas outreach toward Python developers and would love to represent Maemo at any events, particularly Dallas and Austin areas.

Randall Arnold
maemo.org community council 2009-2010
http://talk.maemo.org/index.php
http://tabulacrypticum.wordpress.com/

Cool! I blogged about the event myself: http://ne...

2009-11-05T11:38:52.552-08:00
Cool! I blogged about the event myself:

http://neopythonic.blogspot.com/2009/11/python-in-scientific-world.html

s this greatest effort by him will be fruitful to ...

2009-09-18T02:29:25.316-07:00
s this greatest effort by him will be fruitful to all linux users...........

Hi, I'm used to high level JAVA programming bu...

2009-06-08T08:10:43.393-07:00
Hi, I'm used to high level JAVA programming but I'm going to migrate to Python largely because of numpy/scipy. But I'm having quite a run around trying to just get it installed. I'm running Vista Home Basic. I'm trying to get MinGW and MSYS installed ( under cygwin . . . although I don't quite understand the relationship between cygwin and MinGW ( right now I have them installed in separate directories )and between MinGW and MSYS) and then build numpy and scipy on NetBeans. I've got Cygwin installed sucessfully as far as I can tell. But I'm having trouble installing MSYS.
from the neuroimaging.scipy.org web site I'm told I need:
* bzip2-1.0.3-MSYS-1.0.11-snapshot.tar.bz2
* coreutils-5.97-MSYS-1.0.11-snapshot.tar.bz2
* diffutils-2.8.7-MSYS-1.0.11-snapshot.tar.bz2
* gawk-3.1.5-MSYS-1.0.11-snapshot.tar.bz2
* make-3.81-MSYS-1.0.11-snapshot.tar.bz2
* msysCORE-1.0.11-2007.01.19-1.tar.bz2
* binutils-2.17.50-20070129-1.tar.gz

I'm having trouble finding gawk and diffutils.
I found these programs on the gnu ftp site but the file names don't contain MSYS. I'm unsure about if these are really the same files . . . if MSYS is added to the file name just for convenience when they are packaged together for MSYS or if they actually are completely different files that are made to interact with MSYS in some way thats way over my head?
Also I ran into an installer on SourcForge but I'm not sure if using that is equivalent to using the list of tar files above, and it didn't seem to say if it installed all of the components in the tar file list above that are required for scipy. I'm trying to make sure I don't make a mistake that I have no way to diagnose later. When I get through this process I'm going to be very happy to use scipy in a high level way, but right now I'm pretty traumatized :(
any help is welcome
digdug2

As one of the thousands who admire SciPy from afar...

2009-04-22T17:11:00.000-07:00
As one of the thousands who admire SciPy from afar and will probably be unable to attend, I would love to be able to see the conference in absentia through online video and links to presentation materials.

PyCon 2009 is the reference standard for making a conference live on after it ends and making conference reach far broader than the people who were able to attend in person. All the PyCon 2009 presentations, and most of the tutorials, as well as the many excellent Lightning Talks were captured for posterity and are now visible at PyCon.blip.tv. You can hover over the conference schedule at us.pycon.org/2009/conference/schedule/ to get links to slides, papers, code, etc. for each of the scheduled talks.

If SciPy could do something similar, many of us unable to attend in person would benefit greatly! The PyCon organizers can probably help with the specifics.

Thanks, I've joined the mailing list now and am ca...

2009-03-26T10:58:00.000-07:00
Thanks, I've joined the mailing list now and am catching up on the Python 3 discussions there.

Hello Barry,

Thanks for the interest in wor...

2009-03-25T14:09:00.000-07:00
Hello Barry,

Thanks for the interest in working on the Python 3 port as part of the 2009 Google Summer of Code. The best place to find additional information is to ask on the NumPy and SciPy developer's list. There is all ready at least one other student who has started looking into this. I am not sure that porting to Python 3 is a particularly good project, but there are a number of related sub-projects that could be of great benefit. In particular, increasing test coverage in preparation for the port would be a very reasonable place to focus.

I'm a student planning on applying for the Summer ...

2009-03-20T07:45:00.000-07:00
I'm a student planning on applying for the Summer of Code this year. I think a project porting NumPy/SciPy to Python 3 sounds interesting. Is there somewhere I could find more information on this project? I've looked at the SummerofCodeIdeas wiki page, but there isn't much information there. Is there somewhere I could find more information (potential mentors, experience/skills required, what would be a realistic goal for a summer project, etc.)?

NumPy is quite far from being a dinosaur. It is a...

2009-02-05T23:39:00.000-08:00
NumPy is quite far from being a dinosaur. It is an active, vibrant community project that has undergone at least two major rewrites in its 10+ year history.

My post wasn't intended to suggest in any way that the project is dying--far from it. I agree that for a pure Python project and even some extension code porting to Python 3 should be relatively easy and may happen quickly. However, for the reasons that I listed I don't believe NumPy will support Python 3 until at least 2010.

Of course we would welcome additional developers. And as Andrew suggested, it is possible that converting some or most of the C extension code in NumPy to Cython may be a useful approach. If anyone is interested in looking into this, it would be very useful (perhaps a Google Summer of Code project). Last summer, Dag Sverre Seljebotn completed a related GSoC project to provide enhanced NumPy support in Cython, which is all ready being used in SciPy (for instance see the kd-tree class for efficient nearest-neighbor queries).

I have the feeling that is something wrong with Nu...

2009-02-05T20:53:00.000-08:00
I have the feeling that is something wrong with NumPy, Python goal is to make the life and the program simple and easy, if developpers of NumPy created a dinosaur that cannot be maintained this is a big problem, I can see two solutions (because wait more that one year that one package to move for me is the dead of the package):
1. ask Google and all the Python community for help (don't be scary to avoid that you need external help, you have actions inside Matlab?)
2. decide inside the Python community that we need to build alternative package Python 3 oriented and be able to support more quickly the Python version changes.

hi,

with pygame, we have made our 'pgreload...

2009-02-05T13:18:00.000-08:00
hi,

with pygame, we have made our 'pgreloaded' branch work with py3k and python 2.x. Well Marcus did it all himself in a day or so part time.

So it is possible to do.

Also Lenard managed to partially compile numpy for python 2.6 on windows... so if you need any help from him with this please ask him - he's a windows ninja.

http://www3.telus.net/len_l/pygame/

cu.

As I understand it, Cython is already working to g...

2009-02-05T09:41:00.000-08:00
As I understand it, Cython is already working to generate Python 3 compatible .c files from .pyx and also supports the buffer interface -- we might be able to leverage their work to get something working sooner than anticipated, and perhaps even to have the same source (.pyx and .c) work for both 2.x and 3.

Thanks for the post.

Jarrod Millman

Since you liked zip, why not to like reduce (altho...

I would use itertools.chain: from itertools impor...

another variation of solution 7: reduce(lambda ou...

Or using a bit from the previous comment (and rela...

Here's another simple one (note that you are t...

Good thing there's only one way to do it with ...

A variation of solution 6: [item for items in (...

I also like solutions 5 as a general solution, tho...

As far as I understand it from the NumPy mailing l...

I believe the current development version of numpy...

Any word on the status of the 3.0 port?

Please take a look at the FOSSEE website. You can...

Starting at the end of May 2009, Prabhu very qu...

I can't find any mention of this on LinkedIn--...

Cool! I blogged about the event myself: http://ne...

s this greatest effort by him will be fruitful to ...

Hi, I'm used to high level JAVA programming bu...

As one of the thousands who admire SciPy from afar...

Thanks, I've joined the mailing list now and am ca...

Hello Barry,Thanks for the interest in wor...

I'm a student planning on applying for the Summer ...

NumPy is quite far from being a dinosaur. It is a...

I have the feeling that is something wrong with Nu...

hi,with pygame, we have made our 'pgreload...

As I understand it, Cython is already working to g...

Hello Barry,

Thanks for the interest in wor...

hi,

with pygame, we have made our 'pgreload...