refactor: introduce parametrised nplike types #2795
Merged
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The existing type hints for
nplike
were watered down to support a singleArrayLike
protocol. This prevents us from imposing the constraint that a CuPy array can only be handled by a CuPy nplike.This PR restores the original parametrised-nplike protocol, so that it's possible to declare an nplike as operating upon fixed array types. Unfortunately, mypy doesn't support higher-kinded-types, which are required to compose generic types. This means that we can't e.g. have
and so on. There are higher-kinded-type works related to mypy, e.g. https://sobolevn.me/2020/10/higher-kinded-types-in-python, but they don't look like a good fit for our needs; they have some drawbacks and impose constraints on the runtime logic.
So, this PR won't have huge sweeping impact, but it will allow us to explicitly type an nplike as a typetracer nplike, e.g.
such that we can use fewer assertions.