[ty] Add subdiagnostic suggestion to unresolved-reference diagnostic when variable exists on self

[lipefree]

Summary

Implements astral-sh/ty#502.

In the following example:

class Foo:
    x: int

    def method(self):
        y = x

The user may intended to use y = self.x in method.

This is now added as a subdiagnostic in the following form :

info: An attribute with the same name as 'x' is defined, consider using 'self.x'

Currently the following will not raise the subdiagnostic which is what I am working on to close this PR:

class Foo:
    x: int = 0

    def method(self):
        y = x

Test Plan

I will add mdtests with snapshots to capture the subdiagnostic (I just have to find the right file).

[github-actions]

mypy_primer results

No ecosystem changes detected ✅

[lipefree]

Sorry I did not see the panic found by the fuzzing, I will work on it

[AlexWaygood]

if you make this change to the TypeInferenceBuilder::class_context_of_current_method() method on your PR branch, it will fix the fuzzing failure:

diff --git a/crates/ty_python_semantic/src/types/infer.rs b/crates/ty_python_semantic/src/types/infer.rs
index abb6f91206..8b627ac33c 100644
--- a/crates/ty_python_semantic/src/types/infer.rs
+++ b/crates/ty_python_semantic/src/types/infer.rs
@@ -1714,6 +1714,10 @@ impl<'db> TypeInferenceBuilder<'db> {
     fn class_context_of_current_method(&self) -> Option<ClassLiteral<'db>> {
         let current_scope_id = self.scope().file_scope_id(self.db());
         let current_scope = self.index.scope(current_scope_id);
+        if current_scope.kind() != ScopeKind::Function {
+            return None;
+        }
+
         let parent_scope_id = current_scope.parent()?;
         let parent_scope = self.index.scope(parent_scope_id);

It's a pre-existing bug that wasn't caused by your PR -- it's just been exposed by your PR!

[lipefree]

if you make this change to your PR branch, it will fix the fuzzing failure:

diff --git a/crates/ty_python_semantic/src/types/infer.rs b/crates/ty_python_semantic/src/types/infer.rs
index abb6f91206..8b627ac33c 100644
--- a/crates/ty_python_semantic/src/types/infer.rs
+++ b/crates/ty_python_semantic/src/types/infer.rs
@@ -1714,6 +1714,10 @@ impl<'db> TypeInferenceBuilder<'db> {
     fn class_context_of_current_method(&self) -> Option<ClassLiteral<'db>> {
         let current_scope_id = self.scope().file_scope_id(self.db());
         let current_scope = self.index.scope(current_scope_id);
+        if current_scope.kind() != ScopeKind::Function {
+            return None;
+        }
+
         let parent_scope_id = current_scope.parent()?;
         let parent_scope = self.index.scope(parent_scope_id);

It's a pre-existing bug that wasn't caused by your PR -- it's just been exposed by your PR!

Thank you so much @AlexWaygood, I was going for a full investigation but on my code only ! You are saving me so much time here ! I will then add it to my PR!

[carljm]

Looks great, thank you! Couple of minor nits.