[Biopython-dev] Proposal for GSoC 2017

[Peter Cock]

On Thu, Mar 2, 2017 at 1:25 PM, Sourav Singh <ssouravsingh12 at gmail.com> wrote:
> Respected Sir,
>

You don't need to call me Sir ;)

> Thanks for the mail!
>
> Regarding the part about NumPy support for PyPy, I had heard that like two
> years ago. So I am not sure if things have changed in terms of NumPy's
> support for PyPy.

I don't think NumPy are making any special effort to be compatible with PyPy,
rather the PyPy team want to improve their compatibility with CPython in
general, and seem to be using NumPy as a good test case with lots of
potential users.

Quoting a recent PyPy project blog post in November 2012,

"We continue to make incremental improvements to our C-API
compatibility layer (cpyext). We pass all but 12 of the over-6000
tests in the upstream NumPy test suite, and have begun examining what
it would take to support Pandas and PyQt."

https://morepypy.blogspot.co.uk/2016/11/pypy27-v56-released-stdlib-2712-support.html

Perhaps the Biopython C code can also take advantage of cpyext
to work under PyPy (something I have not looked at)?

We have an open issue on this: https://github.com/biopython/biopython/issues/953

> As for specific modules, I am also thinking of Bio.Statistics.lowess,
> Bio.HMM, Bio.SVDSuperimposer among many others.
>
> My main criteria for choosing the module for Numba is-
>
> 1) The modules which could be necessary for LLVM speedups like Bio.HMM
>
> 2) The modules which can demonstrate a significant speedup using Numba.
>
> Regards,
>
> Sourav

Thanks for the clarification - hopefully this will catch the eye of any
potential mentors.

Peter