We love open-source contributions and they are very welcome 🤗!
We have a queue of open things we are working on here and in the open issues. Feel free to open an issue or contact @csinva (cs1@berkeley.edu or https://www.linkedin.com/in/csinva/) if you want to contribute!
Before contributing, it would be good to read the sklearn estimator contributing guide and generally be familiar with sklearn.
- For examples, functions/classes that are not meant for external use should start with an underscore (e.g.
_Rule)
Docs are built using pdoc. Build them by changing to the docs directory and then running ./build_docs.sh.
Tests are run with pytest (e.g. run pytest in the repo directory) - make sure they pass before pushing code, and that new models pass a reasonable set of tests. Note that you might need to install some additional dependencies in order to get the tests to pass.
The model is on pypi. Packaged following this tutorial. Relevant commands:
pip install twine wheel
rm -rf build dist
python setup.py sdist bdist_wheel
twine check dist/*
twine upload dist/*Some models, e.g. FIGS and hierarchical shrinkage have their own dedicated doc pages.
To add a doc page like this, copy docs/figs.html into a new file and then add in the relevant content. You will also need to manually edit the TOC under "Our favorite models" of each of the html.mako file in this repo (and other existing tutorials).
You may also need to clean up a string in style_docs.py...