On 30 September 2010 03:45, Gary Martin <garycmar...@googlemail.com> wrote:
> However to be honest, I think the extra dependancies are the real issue > just at the moment (I haven't seen it working yet). I guess specific distros > could each individually package up and install the needed dependancies > (though that is not very child friendly, at least a deployment could choose > to include it). Unfortunately as it is, almost none of the existing Sugar > users will be able to use this activity. I try to keep a fairly clean Sugar > environment here for testing, and it fails right away with "ImportError: No > module named nltk.corpus" > > I made it as far as yum installing 15Mb for python-nltk (which pulled in > about 11 external dependancies), at which point the activity would run and > show its UI, but was still non-functional. I looked through the error logs > (attached below if you are curious) but didn't have time to try and go > through the dependencies trying to find what was still missing. > NLTK depends on much more than the software. It also needs the corpus data to run effectively[1][,2]. From memory, that's generally several hundred megs. However wordGroupz may not require the whole stack. NLTK is a highly modular application/framework and could be spliced and packaged by the wordGroupz maintainers. I've done a fair bit of work with NLTK this year and am fairly active in that community, however I don't have time to look into the complexities of this right now. Tim [1] http://www.nltk.org/data [2] http://nltk.googlecode.com/svn/trunk/nltk_data/index.xml
_______________________________________________ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel