[ty] Simple subtyping support for bidirectional inference #21747
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ibraheemdev
requested review from
AlexWaygood,
carljm,
dcreager and
sharkdp
as code owners
ibraheemdev
force-pushed
the
ibraheem/bidi-subtyping
branch
from
51d8ef8 to
0e577a3
Compare
Diagnostic diff on typing conformance testsChanges were detected when running ty on typing conformance tests--- old-output.txt 2025-12-02 22:03:19.784386714 +0000
+++ new-output.txt 2025-12-02 22:03:23.396430298 +0000
@@ -815,7 +815,7 @@
overloads_evaluation.py:291:33: error[invalid-return-type] Function always implicitly returns `None`, which is not assignable to return type `T@example6`
protocols_class_objects.py:58:16: error[invalid-assignment] Object of type `<class 'ConcreteA'>` is not assignable to `ProtoA1`
protocols_class_objects.py:59:16: error[invalid-assignment] Object of type `<class 'ConcreteA'>` is not assignable to `ProtoA2`
-protocols_definition.py:30:11: error[invalid-argument-type] Argument to function `close_all` is incorrect: Expected `Iterable[SupportsClose]`, found `list[Unknown | int]`
+protocols_definition.py:30:11: error[invalid-argument-type] Argument to function `close_all` is incorrect: Expected `Iterable[SupportsClose]`, found `list[SupportsClose | int]`
protocols_definition.py:114:22: error[invalid-assignment] Object of type `Concrete2_Bad1` is not assignable to `Template2`
protocols_definition.py:115:22: error[invalid-assignment] Object of type `Concrete2_Bad2` is not assignable to `Template2`
protocols_definition.py:116:22: error[invalid-assignment] Object of type `Concrete2_Bad3` is not assignable to `Template2`
|
|
|
| Lint rule | Added | Removed | Changed |
|---|---|---|---|
invalid-argument-type |
26 | 80 | 168 |
invalid-assignment |
5 | 25 | 66 |
unsupported-operator |
0 | 1 | 39 |
type-assertion-failure |
0 | 13 | 21 |
invalid-return-type |
1 | 4 | 14 |
possibly-missing-attribute |
2 | 11 | 2 |
unresolved-attribute |
14 | 0 | 0 |
no-matching-overload |
7 | 5 | 0 |
unused-ignore-comment |
5 | 6 | 0 |
non-subscriptable |
1 | 4 | 0 |
not-iterable |
0 | 2 | 0 |
unsupported-base |
0 | 1 | 0 |
| Total | 61 | 152 | 310 |
ibraheemdev
commented
Dec 2, 2025
crates/ty_python_semantic/resources/mdtest/type_compendium/tuple.md
Outdated
Show resolved
Hide resolved
CodSpeed Performance ReportMerging #21747 will degrade performances by 7.79%Comparing Summary
Benchmarks breakdown
Footnotes
|
ibraheemdev
commented
Dec 2, 2025
ibraheemdev
force-pushed
the
ibraheem/bidi-subtyping
branch
2 times, most recently
from
ceff10e to
83ba34d
Compare
ibraheemdev
force-pushed
the
ibraheem/bidi-subtyping
branch
from
83ba34d to
19c5cd8
Compare
AlexWaygood
approved these changes
Dec 2, 2025
Comment on lines
+1055
to
+1069
| // TODO: We should use the constraint solver here to determine the type mappings for more | ||
| // complex subtyping relationships, e.g., `type[C[T]]` to `Callable[..., T]`, or unions | ||
| // containing multiple generic elements. |
Member
There was a problem hiding this comment.
you could possibly also mention protocols here, since although this PR tackles simple cases where a nominal type actually subclasses a protocol, there are also lots of cases where a nominal type is a subtype of a protocol without actually subclassing that protocol
| timestamp: datetime = field(default_factory=datetime.now, init=False) | ||
|
|
||
| # revealed: (self: Data, content: list[int] = list[int]) -> None | ||
| # revealed: (self: Data, content: list[int] = Unknown) -> None |
Member
There was a problem hiding this comment.
any idea what's going on here?
Comment on lines
+1077
to
+1093
| for (base_typevar, base_ty) in specialization | ||
| .generic_context(db) | ||
| .variables(db) | ||
| .zip(specialization.types(db)) | ||
| { | ||
| if *base_ty == Type::TypeVar(bound_typevar) { | ||
| if !visited.insert((origin, base_typevar.identity(db))) { | ||
| return None; | ||
| } | ||
|
|
||
| if let Some(ty) = | ||
| self.find_type_var_from_impl(db, base_typevar, origin, visited) | ||
| { | ||
| return Some(ty); | ||
| } | ||
| } | ||
| } |
Member
There was a problem hiding this comment.
we do a similar loop through the MRO in the (old) constraint solver which doesn't seem to need to do the FxHashSet memoization thing -- could you also take a similar approach here? Might that be more performant?
ruff/crates/ty_python_semantic/src/types/generics.rs
Lines 1598 to 1641 in 2250fa6
Comment on lines
6128
to
6134
| // ```py | ||
| // class tuple: | ||
| // class tuple(Sequence[_T_co]): | ||
| // @overload | ||
| // def __new__(cls) -> tuple[()]: ... | ||
| // @overload | ||
| // def __new__(cls, iterable: Iterable[object]) -> tuple[object, ...]: ... | ||
| // def __new__(cls, iterable: Iterable[_T_co]) -> tuple[_T_co, ...]: ... | ||
| // ``` |
Member
There was a problem hiding this comment.
oops, sorry... I guess I baked this bug in a long time ago for you to find now :-)
Member
|
(this will also help with #21499, where I'm currently trying (not very successfully) to do some awkward workarounds due to this missing feature 😆) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Summary
This solves most of astral-sh/ty#1576. More advanced cases likely require the new constraint solver.