[Biopython-dev] biopython-dev BOSC BoF redux

[Jeffrey Chang]

Hello everybody,

Here is a summary of the Developer's BoF at BOSC.  It'll provide a
good roadmap of things that need to be done in the next few months.

Jeff



Andrew's work at EBI

- Andrew met bioperl and biojava developers
- Andrew wrote code to connect biopython objects to Ewan's bioperl-db


Documentation

- Brad needs some feedback, what's important to work on?
- we need documentation for Martel
- manual and tutorial should be separate
- A cookbook with mini blocks of working code is a good idea.
   Brad will investigate
- general dissatisfaction with Wiki.  Organization problems?


Reorganizing modules

Currently, modules are organized:
   Bio/
     [databases]
     Data/
     Tools/
       Classification/
       Clustering/
       Transcribe.py
       Translate.py
       listfns.py
       mathfns.py
       stringfns.py
       MultiProc/
     Seq.py
     SeqFeature.py
     utils.py
   Martel/

Discussion points:
- want a broad tree rather than deep tree -- stuff easier to find
- packages for end users should be on top, rather than buried deep inside
- should try to keep stuff in Bio namespace to avoid clashes with
other packages (Martel is an exception)
- directories within Tools/ should be moved up one level
- should add sequtils.py to include general sequence calculations


Johann's stuff
- has a DAS server, almost done
- has code to work with controlled vocabularies on expression data
- needs to get approval to release to biopython


Integrating current parsing API with Martel

We considered two alternatives to building the parsing API.  Both are
compatible with Martel:
Current API:
   parser = XXXParser()
   obj = parser.parse(handle)
Alternative API:
   p = format.make_parser()
   p.setContentHandler(obj)
   p.parseFile(handle)
   return obj.built

We agreed to map Martel into the current API, because it is simpler
and less tied to the Martel/SAX way of doing things.



Biopython.com

Technically, we talked about this in the general BoF, but I'm
including it here because of its importance.  Andrew had some
proposals for what to do with the biopython.com domain name.  We could
1) allow him to use it for his consulting company, or we can 2)
reserve it for companies that provide support and/or services for
biopython.  We agreed that the second would be the better choice, and
trust him to maintain the domain for those purposes.