Should mypy be added to dev deps?

[korikuzma]

I think therapy normalizer is the only project that uses mypy. Should we add mypy to our other python projects for consistency?

[jsstevenson]

A few stray thoughts

• I think many of our type errors would elude type checks because we use so many collection types (eg dicts, triples) and the error might be, like, in the type of the second item in the tuple. I think there are strictness checks you can enforce so that you have to exhaustively define things but that might take some time to catch up to.
• some people prefer pyright. Some people just roll both since they aren't mutually exclusive and each has their own strengths.
• The way VRS is structured, especially VRS 2, is really annoying for type checking, because so many attributes can either be an ID/reference or an actual object, but at runtime we tend to assume that the full object is there. Pyright/mypy light up on projects like the mave mapping lib because of this. So for projects that deal with VRS or sequence type objects you're signing up for a lot of 'type: ignore's, which starts to beg the question of how helpful it is to run the checks in the first place.

Personally, I have pyright running in my editor (anyone with VScode probably has it or Pylance as well) and I use it as insights to flag possible errors, but I haven't been in any rush to push for expanding our mypy usage in CI beyond therapy (have thought about removing it to be consistent tbh)

[github-actions[bot]]

This issue is stale because it has been open 45 days with no activity. Please make a comment for triaging or closing the issue.

[korikuzma]

I think we can close this. We can re-open if we'd like to add it in the future.