Adding DrugBank DDI and Two Sides.

[benedekrozemberczki]

Summary

• Adding DrugBank DDI and TwoSides.

• Code passes all tests

• Unit tests provided for these changes

• Documentation and docstrings added for these changes

Changes

• Adds the DrugBank DDI and TwoSides Datasets.
• Adds the loaders.
• Adds tests.
• Adds documentation of data cleaning.
• Adds the data cleaning scripts.

[cthoyt]

I’d highly request provenance information on how these new datasets were constructed. Were they automatically downloaded from Some external repo? Was processing done to them?

[benedekrozemberczki]

Yes, @cthoyt I will do that in a moment.

[benedekrozemberczki]

There will be a whole Appendix section about this in the paper.

[cthoyt]

While you’re thinking about it maybe also consider doing the same for the previous two datasets as well :)

[benedekrozemberczki]

@cthoyt How about a dataset preprocessing section in the documentation?

[cthoyt]

Tbh the only important documentation of data preprocessing to me is code that can exactly reproduce it. Let’s start there and backfill prose-based documentation if there are any places where it can’t be better documented in code itself

[benedekrozemberczki]

Added the cleaning scripts.

[cthoyt]

It appears you merged the branch with failing tests. This shouldn’t be allowed/possible - the solution is to add some branch protection rules in the settings for the repository

[benedekrozemberczki]

Does it require changes to Github actions?

[cthoyt]

Nope that looks right!