[Openchemistry-developers] GSoc'17 Project: RDKit-MMTF Integration

[Greg Landrum]

[replying from my gmail account, where I do most of the RDKit work]

Dear Ritesh,

Thanks for expressing an interest! I've CC'ed the openchemistry-developers
list (http://public.kitware.com/mailman/listinfo/openchemistry-developers)
on this message because as a first-time GSoC mentor I'm hoping to get a bit
of mentoring myself from the community. It's also a good place for these
discussions. You will definitely want to subscribe to that mailing list if
you haven't already.

A good way to get started is to gain a bit of familiarity with the RDKit
itself and to start by fixing a bug or implementing a small feature and
submitting a PR with your change. That let's you start to learn the code
base and us (as well as the folks who evaluate the actual applications) get
to know you and your coding style a bit. You will want to take a look at
the RDKit source and try to maintain a consistent style in terms of the fix
and how it is tested. I'm happy to help with that (probably best done using
the rdkit-devel mailing list:
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss)

The open RDKit issues are all in the github tracker where you can take a
look, but I'd suggest that this one:
https://github.com/rdkit/rdkit/issues/888 is of reasonable scope and
shouldn't be overly difficult.

With respect to the MMTF project itself: We are looking for an reader for
MMTF -> RDKit (i.e. the input is MMTF data and the output is an RDKit
molecule) and a writer that goes RDKit -> MMTF. The implementation should
be in C++ and be inspired by the existing PDB reader/writer. Wrappers to
access the functionality from Python and Java/C# will also be needed, again
very analogous to the existing PDB reader and writer. As well as the base
implementation, we'd expect a solid set of unit tests for the code.

I hope this helps,
-greg




On Sat, Mar 4, 2017 at 11:56 AM, Greg Landrum <
greg.landrum at t5informatics.com> wrote:

>
>
> ---------- Forwarded message ----------
> From: RITESH PRASAD <ritesh.prasad289 at gmail.com>
> Date: Sat, Mar 4, 2017 at 10:44 AM
> Subject: GSoc'17 Project: RDKit-MMTF Integration
> To: greg.landrum at t5informatics.com
>
>
> Sir,
>
> *Introduction:*
> Myself *Ritesh Prasad*. I'm a *second year student(Computer Science and
> Engineering) at National Institute of Technology, Jamshedpur, India*.
>
> *Programming background:*
> I prefer programming in *C++* and *python*. I'm also quite comfortable
> with *Javascript*.
> *Projects:*
>
>    - I'm currently working on a *project to identify gestures from images
>    using deep learning, specifically convolutional neural networks*.
>    - I've been a part of a open source project to *create a chrome
>    extension to enable cut,copy and paste* on the websites where it is
>    disabled.(https://github.com/ritesh-nitjsr/CopyPaste)
>    - I've developed *an online judge for competitive programming using
>    NodeJS, ExpressJS, MongoDB, HTML and CSS* (https://github.com/ritesh
>    -nitjsr/codeknights).
>
>
> I would like to work on the project proposed by you *to implement MMTF
> reader and writer for RDKit*. I did a bit of research on MMTF file
> format. However I could not get much insights about the project. Please
> provide some appropriate insights and suggestions so that I can get myself
> working on it.
>
> Anticipating a quick reply.
>
> Regards,
> Ritesh Prasad
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/openchemistry-developers/attachments/20170304/24082af4/attachment.html>