For running tests we need pytest and sklearn as we test against sklearns implementation of coordinate descent. We should make this explicit in the pyproject.toml.
sklearn has quite a few dependencies and we currently only have it as "soft" dependency for testing (not explicity noted in the pyproject.toml). Allowing for pipelines would be cool though.
For running tests we need
pytest
andsklearn
as we test againstsklearn
s implementation of coordinate descent. We should make this explicit in thepyproject.toml
._Originally posted by @simon-hirsch in https://github.com/simon-hirsch/rolch/pull/25#discussion_r1768039197_