Skip to content

Yunrui-Liu/research_code

 
 

Repository files navigation

Writing good research code

This repo contains the slides and code for a presentation on writing research software I first gave in January 2021 to the PhD students in neuro at Harvard. It's a compendium of 5 lessons I learned the hard way about writing research code that won't bite back.

  • The slides are here
  • The rest of the repo contains supporting code in the format advocated in the first lesson.
  • You can see the full presentation recorded at NMA 2021 and a short version focused on testing recorded at Brainhack MTL 2021.

For the book version of these slides, see goodresearch.dev.

Organization

This repo follows the organization of shablona. All the code and tests are under research_code. research_code is itself a Python package.

For the package, we use the same setup as this tutorial on setuptools, and is compatible with it - this repo is publishable to PyPI directly!

To install the package locally in development mode

cd into this directory, then run:

pip install -e .

In Python:

import research_code

To test

cd into the research_code/tests directory, then run each file individually, or run nose2.

CI

While shablona recommended the use of Jenkins for continuous integration (CI), we showcase instead Github actions, which don't require additional accounts/software. The workflow, which runs tests, is located in .github/workflows/ci.yml.

About

Writing good research code

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 54.7%
  • Python 45.3%