Description
As mentioned in #9958 (comment), people are more likely to bump into CONTRIBUTING.md than the contributing doc on scikit-learn.org before opening an issue or PR.
CONTRIBUTING.md should have very little content and points to the relevant portion of the scikit-lern.org dev (because stable can lag behind quite a bit) doc. Here are a few points that are worth mentioning:
Filing issues
Use the template provided: What you did, what you expect, what you got, scikit-learn and dependencies version. Full traceback if you get an error. A stand-alone snippet makes it a lot more likely to get good feedback.
Link to dev doc for all the details.
Pull requests
For contributing Pull Requests:
- if you are not familiar with Pull Request, relevant link to https://opensource.guide/ or relevant help.github.com link
- install the scikit-learn dev version, link to dev doc
- make sure you can run the tests localy, link to the dev doc
For more details, link to the dev doc
Other things:
- Other ways to contribute, reviewing pull request, improve the doc, answer questions on github and/or StackOverflow and/or the mailing list.
- Link to first good issue, easy issues, help needed issues, maybe bug issues
- other things I may have missed
I may have missed things, obviously comments more than welcome.
Related things were discussed in #9958 and probably others.
cc @mohamed-ali who said he may be interested to work on this.
A few contributing guide I could find:
numpy
scipy
pandas
dask
requests
django