[Biopython] Google Summer of Code

[Peter Cock]

On Thu, Mar 1, 2012 at 8:14 AM, Mic <mictadlo at gmail.com> wrote:
> Hello,
> Is it possible to use PyPy with:
> * BioPython
> * Pysam
> * Matplotlib
> * etc
>
> If not than it might be good idea to get a support for it with help of
> Google Summer of Code, because PyPy getting faster and faster.

Most of Biopython is working under PyPy (ignoring the C extensions,
much like our situation under Jython). This was mentioned in the
release notice for Bioython 1.59 - early adopters may be able to
find other problems that we're not aware of from the unit tests:
http://news.open-bio.org/news/2012/02/biopython-1-59-released/
I doubt there is enough work here alone to make a GSoC project.

I'm not sure about pysam under PyPy - but I would be interested
to know, because here interfacing with the samtools C code is the
essence of pysam. My impression from the PyPy mailing lists
calling external C libraries from PyPy is that this is another area
of active work.

For matplotlib, you would need NumPy under PyPy. That is an
area of active work for the PyPy team who are currently trying
to re-implement a pure-python version of NumPy which they are
calling NumPyPy (originally it was called micronumpy) sufficient
for other libraries using just the Python numpy API to run. A
problem with this is many Python libraries also use the NumPy
C API (e.g. bits of Biopython). See for example:
http://morepypy.blogspot.com/2012/01/numpypy-status-update.html
http://technicaldiscovery.blogspot.com/2011/10/thoughts-on-porting-numpy-to-pypy.html
I suggest reading the PyPy and NumPy mailing list archives for
more about this.

Peter