[BioPython] script for clustal alignment format

[Jeffrey Chang]

On Tue, 24 Oct 2000, Andrew Dalke wrote:

> Jeff:
> >Andrew, how do you feel about bundling mxTT with Martel/Biopython?
> 
> Not a problem.  The other option is to provide an alternate location for
> downloading mxTT.

Yes, definitely.  My preference, though, would be to go ahead and bundle
it (if possible etc etc), since it's small, and we could merge the
installations.  I suspect that the parsers are/will be Biopython's most
used code, so this would make things easier for a lot of users.


> >Also, when should we add Martel to the Biopython distribution?  I'm
> >planning on making one more release soon, and we could work on integration
> >after that.
> 
> I take it we've decided that Martel is usable and has the needed
> performance?
> I don't think there's been real solid testing and characterization by anyone
> other than me, although looking at the list traffic there have been a couple
> of people who've tried it out.

There's already code in Biopython that requires Martel (Brad's ClustalW
stuff), and I'm reasonably convinced that it can do the stuff already done
with the parsers, with at least comparable or better execution speed.

Although some of the existing stuff may never be moved to Martel (e.g. the
blast parser, which would be hard to duplicate/replace given the amount of
testing it's undergone), it'll probably lower the barrier to create new
parsers.  For example, having Martel around may finally encourage someone
to write a robust Genbank parser (hint hint :).


> I still have the nagging feeling that Martel doesn't support some class of
> file formats (tree-based phylogeny ones?).  But that hasn't yet happened, so
> I shouldn't be too concerned, and I've ideas for other ways to handle those.

Sure, but that's OK.  Even if it's not a be-all-end-all of parsers (which
probably doesn't even exist), it seems to do what it does extremely
well.  I'm not concerned either.



> Development-wise, the API still needs to be converted to SAX 2.0 and support
> added for namespaces.  There also needs to be a way to convert the callback
> form into an iterator one.  The code isn't too complicated, but as you
> all can tell I haven't been able to work much on the project :(  Maybe
> two more weeks?  Three?

Great!  Looking forward to it.

Jeff