[red-knot] Check subtype relation between callable types #16804

dhruvmanila · 2025-03-17T14:22:21Z

Summary

This PR adds support for checking the subtype relationship between the two callable types.

The main source of reference used for implementation is https://typing.python.org/en/latest/spec/callables.html#assignability-rules-for-callables.

The implementation is split into two phases:

Check all the positional parameters which includes positional-only, standard (positional or keyword) and variadic kind
Collect all the keywords in a HashMap to do the keyword parameters check via name lookup

For (1), there's a helper struct which is similar to .zip_longest (from itertools) except that it allows control over one of the iterator as that's required when processing a variadic parameter. This is required because positional parameters needs to be checked as per their position between the two callable types. The struct also keeps track of the current iteration element because when the loop is exited (to move on to the phase 2) the current iteration element would be carried over to the phase 2 check.

This struct is internal to the is_subtype_of method as I don't think it makes sense to expose it outside. It also allows me to use "self" and "other" suffixed field names as that's only relevant in that context.

Test Plan

Add extensive tests in markdown.

Converted all of the code snippets from https://typing.python.org/en/latest/spec/callables.html#assignability-rules-for-callables to use knot_extensions.is_subtype_of and verified the result.

github-actions · 2025-03-17T14:31:32Z

`mypy_primer` results

No ecosystem changes detected ✅

carljm

This is looking really good! I don't have any significant advice based on what's here already; happy to consider any specific problems you're having with completing the PR.

I think that when we add assignability for callable types, it will make most sense to just expand this function to handle gradual types also and rename it is_assignable_to. Then the function can work both for is_subtype_of and is_assignable_to, in the former case it will just never receive a non-fully-static callable type.

crates/red_knot_python_semantic/resources/mdtest/type_properties/is_subtype_of.md

github-actions · 2025-03-18T10:42:30Z

`ruff-ecosystem` results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

Formatter (stable)

✅ ecosystem check detected no format changes.

Formatter (preview)

✅ ecosystem check detected no format changes.

dhruvmanila · 2025-03-18T12:23:47Z

I think that when we add assignability for callable types, it will make most sense to just expand this function to handle gradual types also and rename it is_assignable_to. Then the function can work both for is_subtype_of and is_assignable_to, in the former case it will just never receive a non-fully-static callable type.

Yeah, I think that makes sense. For now, I did make the assumption that we're only getting fully static types so I avoided the fallback to Unknown if there is no annotated type, mainly to avoid any surprise behavior but that shouldn't be too difficult to change to unwrap_or(Type::unknown()). And, if the entire parameter type is using a gradual form (...) then we can directly short-circuit the implementation.

crates/red_knot_python_semantic/src/types.rs

carljm

This looks really good!! Quite a complex feature, thanks for working through it so thoroughly.

crates/red_knot_python_semantic/resources/mdtest/type_properties/is_subtype_of.md

carljm · 2025-03-19T16:23:23Z

crates/red_knot_python_semantic/src/types.rs

+        let (self_parameters, other_parameters) = parameters.into_remaining();
+
+        // Collect all the keyword-only parameters and the unmatched standard parameters.
+        let mut self_keywords = HashMap::new();


It's possible that argument lists will tend to be short enough that it will be cheaper to use a Vec and just iterate through it than to pay the overhead of hashing. But I'm fine leaving this for a future performance investigation if we see this code as a hot spot. A mapping is semantically the appropriate data structure here.

Yeah, another option is to use smallmap crate.

crates/red_knot_python_semantic/src/types.rs

carljm

This is really excellent!!

crates/red_knot_python_semantic/src/types.rs

* main: (26 commits) Use the common `OperatorPrecedence` for the parser (#16747) [red-knot] Check subtype relation between callable types (#16804) [red-knot] Check whether two callable types are equivalent (#16698) [red-knot] Ban most `Type::Instance` types in type expressions (#16872) Special-case value-expression inference of special form subscriptions (#16877) [syntax-errors] Fix star annotation before Python 3.11 (#16878) Recognize `SyntaxError:` as an error code for ecosystem checks (#16879) [red-knot] add test cases result in false positive errors (#16856) Bump 0.11.1 (#16871) Allow discovery of venv in VIRTUAL_ENV env variable (#16853) Split git pathspecs in change determination onto separate lines (#16869) Use the correct base commit for change determination (#16857) Separate `BitXorOr` into `BitXor` and `BitOr` precedence (#16844) Server: Allow `FixAll` action in presence of version-specific syntax errors (#16848) [`refurb`] Fix starred expressions fix (`FURB161`) (#16550) [`flake8-executable`] Add pytest and uv run to help message for `shebang-missing-python` (`EXE003`) (#16855) Show more precise messages in invalid type expressions (#16850) [`flake8-executables`] Allow `uv run` in shebang line for `shebang-missing-python` (`EXE003`) (#16849) Add `--exit-non-zero-on-format` (#16009) [red-knot] Ban list literals in most contexts in type expressions (#16847) ...

## Summary Part of #15382 This PR adds support for checking the assignability of two general callable types. This is built on top of #16804 by including the gradual parameters check and accepting a function that performs the check between the two types. ## Test Plan Update `is_assignable_to.md` with callable types section.

sharkdp · 2025-03-24T14:10:09Z

Very nice work! One minor question: should Callables be a subtype of object? Should this pass?

from knot_extensions import static_assert, is_subtype_of
from typing import Callable

static_assert(is_subtype_of(Callable[[int], str], object))

https://playknot.ruff.rs/2fd03ad6-c4d9-45e5-bde5-8f2d1ec55abc

dhruvmanila · 2025-03-24T14:56:37Z

One minor question: should Callables be a subtype of object? Should this pass?

Hmm, good question. I think not because object is not callable.

AlexWaygood · 2025-03-24T14:59:22Z

Hmm, good question. I think not because object is not callable.

That's opposite to the way you should think about it -- all fully static types should be a subtype of object.

Consider the way that all instance types are a subtype of object. For example, the bool type extends the features and properties of object (it is an (indirect) subtype of object); thus bool is a subtype of object even though not all instances of object are instances of bool.

carljm · 2025-03-24T15:00:58Z

Callables (fully static ones) should be a subtype of object. The type object is inhabited by all Python objects, so all (fully static) types are subtypes of object.

It is true that all objects are not callable, but this means that object is not a subtype of Callable.

In my understanding this PR wasn't aiming to be comprehensive yet, it was just handling the core calculation of signature subtyping. We have other relations missing as well, eg a FunctionLiteral can be a subtype of a Callable type. Also an Instance type with a __call__ method. And a ClassLiteral type (based on its constructor signature.) And arguably SubclassOf as well, though this one is not sound since we won't enforce Liskov on constructor signatures.

dhruvmanila · 2025-03-24T15:15:34Z

Callables (fully static ones) should be a subtype of object. The type object is inhabited by all Python objects, so all (fully static) types are subtypes of object.

Yeah, I think I got it backwards, thanks.

In my understanding this PR wasn't aiming to be comprehensive yet, it was just handling the core calculation of signature subtyping. We have other relations missing as well, eg a FunctionLiteral can be a subtype of a Callable type. Also an Instance type with a __call__ method. And a ClassLiteral type (based on its constructor signature.) And arguably SubclassOf as well, though this one is not sound since we won't enforce Liskov on constructor signatures.

Yes, that is correct. I'll open an issue to keep track of the remaining relationship, it might be something that a contributor could pick up.

Edit: #16953

dhruvmanila added the ty Multi-file analysis & type inference label Mar 17, 2025

dhruvmanila changed the title ~~[WIP] [red-knot] Fix fully static check for callable type~~ [WIP] [red-knot] Check subtype relation between callable types Mar 17, 2025

dhruvmanila force-pushed the dhruv/callable-static-fix branch from e6e65d7 to 6192776 Compare March 17, 2025 14:24

dhruvmanila force-pushed the dhruv/callable-subtype branch from 8e60a84 to eb8f3ca Compare March 17, 2025 14:25

Base automatically changed from dhruv/callable-static-fix to main March 17, 2025 14:31

dhruvmanila force-pushed the dhruv/callable-subtype branch from eb8f3ca to f9ee155 Compare March 17, 2025 14:31

carljm reviewed Mar 18, 2025

View reviewed changes

dhruvmanila force-pushed the dhruv/callable-subtype branch from 12dc9eb to 98da221 Compare March 19, 2025 04:44

dhruvmanila changed the title ~~[WIP] [red-knot] Check subtype relation between callable types~~ [red-knot] Check subtype relation between callable types Mar 19, 2025

dhruvmanila force-pushed the dhruv/callable-subtype branch from 98da221 to 79a4afc Compare March 19, 2025 10:50

dhruvmanila changed the base branch from main to dhruv/callable-is-equivalent March 19, 2025 10:50

dhruvmanila mentioned this pull request Mar 19, 2025

[red-knot] Support callable types and typing.Callable #15382

Closed

19 tasks

dhruvmanila marked this pull request as ready for review March 19, 2025 10:54

dhruvmanila requested review from AlexWaygood, sharkdp and dcreager as code owners March 19, 2025 10:54

dhruvmanila commented Mar 19, 2025

View reviewed changes

crates/red_knot_python_semantic/src/types.rs Outdated Show resolved Hide resolved

carljm approved these changes Mar 19, 2025

View reviewed changes

dhruvmanila force-pushed the dhruv/callable-is-equivalent branch from dfe5fcc to 6a1f150 Compare March 20, 2025 12:10

dhruvmanila force-pushed the dhruv/callable-subtype branch from 12ff6b9 to a519c2e Compare March 20, 2025 12:10

carljm approved these changes Mar 20, 2025

View reviewed changes

carljm reviewed Mar 20, 2025

View reviewed changes

crates/red_knot_python_semantic/src/types.rs Outdated Show resolved Hide resolved

Base automatically changed from dhruv/callable-is-equivalent to main March 21, 2025 03:19

dhruvmanila added 4 commits March 21, 2025 08:51

[red-knot] WIP

0f1e60d

Handle keyword-only and keyword-variadic parameters

a8692d8

Add tests for keyword-only, keyword-variadic

c2ce05a

Update on latest main

3095412

dhruvmanila added 6 commits March 21, 2025 08:51

Cleanup

2067685

Fix unmatched variadic param, more tests

e9f660f

Avoid duplicate kind

bacde26

Use nested Option for keyword-variadic

ef665aa

Address review comments

b78fcc4

Expand callable subtype todo

3b1e030

dhruvmanila force-pushed the dhruv/callable-subtype branch from a519c2e to 3b1e030 Compare March 21, 2025 03:23

dhruvmanila enabled auto-merge (squash) March 21, 2025 03:23

dhruvmanila merged commit 04a8756 into main Mar 21, 2025
22 checks passed

dhruvmanila deleted the dhruv/callable-subtype branch March 21, 2025 03:27

dhruvmanila mentioned this pull request Mar 21, 2025

[red-knot] Check assignability for two callable types #16845

Merged

dhruvmanila mentioned this pull request Mar 24, 2025

[red-knot] Implement subtyping/assignability between different callable objects #16953

Closed

4 tasks

[red-knot] Check subtype relation between callable types #16804

[red-knot] Check subtype relation between callable types #16804

Uh oh!

Conversation

dhruvmanila commented Mar 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Test Plan

Uh oh!

github-actions bot commented Mar 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

mypy_primer results

Uh oh!

carljm left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions bot commented Mar 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

ruff-ecosystem results

Linter (stable)

Linter (preview)

Formatter (stable)

Formatter (preview)

Uh oh!

dhruvmanila commented Mar 18, 2025

Uh oh!

Uh oh!

carljm left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

carljm Mar 19, 2025

Choose a reason for hiding this comment

Uh oh!

dhruvmanila Mar 20, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

carljm left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

sharkdp commented Mar 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dhruvmanila commented Mar 24, 2025

Uh oh!

AlexWaygood commented Mar 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

carljm commented Mar 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dhruvmanila commented Mar 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

dhruvmanila commented Mar 17, 2025 •

edited

Loading

github-actions bot commented Mar 17, 2025 •

edited

Loading

`mypy_primer` results

github-actions bot commented Mar 18, 2025 •

edited

Loading

`ruff-ecosystem` results

sharkdp commented Mar 24, 2025 •

edited

Loading

AlexWaygood commented Mar 24, 2025 •

edited

Loading

carljm commented Mar 24, 2025 •

edited

Loading

dhruvmanila commented Mar 24, 2025 •

edited

Loading