openkinome / kinoml

Structure-informed machine learning for kinase modeling

https://openkinome.org/kinoml/

MIT License

51 stars 21 forks source link

Add notebook tests to codecov #117

Closed schallerdavid closed 2 years ago

schallerdavid commented 2 years ago

Description

This PR aims to add the notebook tests to the codecov report.

Todos

Notable points that this PR has either accomplished or will accomplish. ~~- [ ] codecov report includes notebook tests~~

[x] add tests for complex featurizers
[ ] tests pass

Other

[x] update notebooks

Status

[ ] Ready to go

codecov-commenter commented 2 years ago

Codecov Report

Merging #117 (c0776a8) into master (10ab0ef) will increase coverage by 15.13%. The diff coverage is n/a.

schallerdavid commented 2 years ago

Somehow it is really hard to get the results from notebook test via nbval and standard tests into one correct coverage report for codecove. Hence, I copied all tests from the notebooks to the standard tests to directory to have them also refected in the coverage report. Since this was mainly about testing of structural featurizers I do not expect a big time increase, since the tests are the same and the intermediate results are cached.

schallerdavid commented 2 years ago

Another take-home message was that testing is more stable when removing multiprocessing from featurization.

schallerdavid commented 2 years ago

72

review-notebook-app[bot] commented 2 years ago

Check out this pull request on

See visual diffs & provide feedback on Jupyter Notebooks.