search

NCBI-Hackathons / Pharmacogenomics_Prediction_Pipeline_P3

Creative Commons Zero v1.0 Universal
28 stars 14 forks source link

P3 Post-hackathon Refactoring #9

Open khughitt opened 8 years ago

khughitt commented 8 years ago

This issue will help to organize some early refactoring efforts.

khughitt commented 8 years ago

While we are on the topic of refactoring, do we plan to continue used a mix of R and Python for the intermediate scripts? Although my initial thoughts are to keep things in a single language for consistency, maintainability, and portability, using both languages would provide some flexibility in choosing the best language for a given task and allow people to contribute in whichever language they feel comfortable with. As long as the list doesn't start to grow to include scripts in ten other languages, I'd be fine with either approach.

daler commented 8 years ago

A typical RNA-seq pipeline uses code written in various languages, but the thing that makes them usable together is that they all have a command line interface.

I think having all scripts conform to these ten recommendations should keep things usable while allowing everyone to work in their preferred language.

khughitt commented 8 years ago

Sounds good to me. Perhaps we can just create a dev guide somewhere where we describe things like coding standards for each language used.

khughitt commented 8 years ago

Regarding the docs, it looks like you have started some .rst files for various pipeline components.

What do you think about moving those files to doc/src and linking them together in an index.rst?

The architecture diagrams, etc. currently in doc/ could be moved to doc/extra.

daler commented 8 years ago

Sure. I meant to bring that up but forgot: should docs live on the github wiki or in text files in the repo?

If the latter, I'm a big fan of sphinx for docs and I can get that set up.

khughitt commented 8 years ago

My preference is for repo docs. I'm not a huge fan of Github wiki pages.. I usually don't ever notice them when they exist and have trouble navigating them even when I look for them.

I was also thinking Sphinx + readthedocs would be a good way to go. That way we can have the docs regenerate on each commit.