Improve typing of inference functions#2166
Conversation
jacobtylerwalls
added
the
Maintenance
Co-authored-by: Daniël van Noord <13665637+DanielNoord@users.noreply.github.com>
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## main #2166 +/- ##
=======================================
Coverage 92.53% 92.54%
=======================================
Files 94 94
Lines 10804 10807 +3
=======================================
+ Hits 9998 10001 +3
Misses 806 806
Flags with carried forward coverage won't be shown. Click here to find out more.
|
for more information, see https://pre-commit.ci
|
@jacobtylerwalls think this is correct. Let me know what you think! |
jacobtylerwalls
left a comment
jacobtylerwalls
left a comment
There was a problem hiding this comment.
New mypy messages as of b3b6c53:
astroid/nodes/node_ng.py:147: error: Argument 1 to "__call__" of "InferFn" has incompatible type "NodeNG"; expected "Self" [arg-type]
astroid/nodes/node_ng.py:149: error: Argument 1 to "__call__" of "InferFn" has incompatible type "NodeNG"; expected "Self" [arg-type]
I'll change it to be Node again. |
This seems to be a |
|
Works for me. The blame will point back to this discussion in case mypy improves before we turn it on. |
Implemented! |
jacobtylerwalls
left a comment
jacobtylerwalls
left a comment
There was a problem hiding this comment.
LGTM, I can't click Approve as the OP. Happy to squash merge once you ✅
| **kwargs: Any, | ||
| ) -> Generator[InferenceResult, None, None]: | ||
| partial_cache_key = (func, node) | ||
| if partial_cache_key in _CURRENTLY_INFERRING: |
There was a problem hiding this comment.
Unrelated, but isn't this what the path wrapper does? Can we use that here?
There was a problem hiding this comment.
Perhaps, but at glance, it looks like that one is sensitive to specific InferenceContexts. Here, we're not ready to unleash recursive inference with every slightly different context.
|
Thanks for the review and for taking over! |
3 participants
Type of Changes
Description
Follow-up to #2158.
Before, this was quite vague: