OSU Open Source Lab

OSL GSOC 2015 - Protein Geometry Database

What is the Protein Geometry Database?

The Protein Geometry Database project (PGD) is many things to many people.

The synopses on code.osuosl.org says:

"Protein Geometry Database is a specialized search engine for protein geometry. It allows you to explore either protein conformation or protein covalent geometry or the correlations between protein conformation and bond angles and lengths."

There's a lot of science in that paragraph; I speak code much better than I speak science, so let's look at the GitHub Repository. That page says things like...

• PGD has 6 open Pull Requests,
• 1 unresolved Issue,
• 1156 commits in the master branch,
• 5 branches, and
• 11 contributors.

It also describes the code as being:

• 59.2% Python,
• 27.2% HTML,
• 12.4% JavaScript, and
• 1.2% Other

Depending on what you use PGD for (if you use it at all) you have a different relationship with the project. What matters here is that PGD is a project that the OSL develops and maintains. This year a lot of great work was done on it for the 2015 Google Summer of Code.

What PGD Accomplished During GSOC 2015

This year's PGD GSOC project had five core goals, all of which got accomplished.

1. Revamping the current account system.
2. Building occupancy awareness into PGD.
3. Testing the current development branch of PGD.
4. Implementing a search by deposition date filter.
5. Upgrading PGD to Django 1.8 (from Django 1.6!)

The student for this project was S. Ramana Subramanyam. He is in his second year at the Birla Institute of Technology and Science in Goa, India, and was wonderful to work with. Despite a 12 hour time difference he was able to be productive the majority of the time.

Although none of the code developed for this year's GSOC has been merged into PGD, it has all been reviewed and will be merged over the next few months as the project lead (Jack Twilley) and I are able to work together on migrating the changes.

Overcoming Challenges

The largest challenges we were faced with in this project were scheduling.

The PGD Project Lead (Jack) got an amazing internship for his Food Science degree in California at a vineyard; as a result he was unable to work on PGD and his GSOC mentorship as much as was initially expected. While I was able to answer (or at least help with) many of the questions S. Ramana had, sometimes we were forced to throw up our hands, send an email to Jack, and wait.

This didn't stop S. Ramana from completing all of his goals for the GSOC project; there was always plenty to do so he could put one thing on the back-burner and focus on a new task. At the most it was a mild inconvenience but didn't get in the way too often.

Where PGD Stands

Once the code is merged and the inevitable version control conflicts are resolved, PGD will have some pretty neat new features:

1. Search results can be saved.
2. Search results can be saved as a PNG image.
3. Occupancy Awareness.
4. Deposition Date is now a search Filter.
5. PGD is pgraded to Django 1.8.

It took a lot of energy not to add ! to the end of each of those items.

Despite scheduling conflicts and the usual technical snafus that come with major engineering changes, I would say that this GSOC was a success for PGD and the OSL.

Personal Takeaways

This was my first time mentoring a student for GSOC and although I have had limited experience mentoring students with Devops Bootcamp, mentoring a student remotely with a 12 hour time difference is an entirely different can of worms.

My mentorship abilities were challenged but I learned a lot of great skills and added many tools to my belt when it comes to dealing with problems and knowing when/who to ask for help. If I am given the opportunity to be a GSOC mentor next year I will definitely jump on the opportunity to do so.