[Biopython-dev] Active projects list? (BOSC Biopython Project Update)

[Peter]

On Sat, Jul 3, 2010 at 7:52 AM, Andrea Pierleoni
<andrea at biocomp.unibo.it> wrote:
> Dear Peter,
> I'm actually working on two more projects than the XML parsing, that could
> be useful in biopython.
>
> 1) together with Mauro Amico, we hare developing a graphical library very
> similar to the Bio::Graphics module pf BioPerl. The project is at good
> point, and will come with documentation and tutorial as a standalone
> package we call BioGraPy. I know that in biopython one can already use
> GenomeDiagram to draw, for example, seqrecord features, but this could
> extend biopython plotting capability significantly. You can use BioGraPy
> to plot a blast output (with its HTML map), to plot hydrophobicity plot
> along the sequence (read as per letter annotations), mRNA and CDS with
> their splicing sites, and so on... BioGrapy relies on matplotlib, so this
> will be an additional external dependence, but worthwhile in my opinion.

That does sound interesting. I'm not saying it couldn't be rolled into
Biopython, but perhaps shipping it a separate package building on
Biopython and matplotlib is a good plan. There are advantages either
way.

> 2) Since I'm working with the web2py web framework, and I work with biosql
> databases, I spent some time adapting the current BioSQL code to be used
> with the web2py DAL (Database Abstraction Layer). DAL is much more simpler
> (and sometimes faster) than SQLAlchemy, and its syntax and use are very
> similar to SQL queries, so it was very easy to adapt the current code to
> use the DAL. Main advantages of using the web2py DAL are that it can be
> used on almost any DB engine. listing from the web2py site: SQLite,
> PostgreSQL, MySQL, MSSQL, FireBird, Oracle, IBM DB2, Informix, Ingres, and
> Google App Engine. I've succesfully tested with both Postgres and SQLite,
> but should be tested for the other. Since the Web2py code is GPL2, I can
> incorporate the modules needed for DAL directly into Biopython, so there
> will be no external dependences. I know that Brad Chapman and some others
> were working on implementing BioSQL with SQLAlchemy, so let me know if
> this could be an addition to Biopython.

I'm not sure we can easily include GPL code in Biopython... it would
complicate things. Kyle has also been working on using the JVM DB
API for BioSQL under Jython - I'd rather we ended up with a runtime
choice of drivers (database specific like mysqldb, and others like the
abstractions SQLAlchemy or the web2py DAL) which would all be
external to Biopython.

Peter