refactor: Make use of is_typing for Literal/Annotated detection.

Makes soft use concept of `mapped_names` more generic and reuse it for
walking `Annotated` value expressions.

flake8_type_checking/checker.py

@@ -66,12 +66,17 @@ def ast_unparse(node: ast.AST) -> str:
         Import,
         ImportTypeValue,
         Name,
+        SupportsIsTyping,
     )
 
 
 class AnnotationVisitor(ABC):
     """Simplified node visitor for traversing annotations."""
 
+    @abstractmethod
+    def is_typing(self, node: ast.AST, symbol: str) -> bool:
+        """Check if the given node matches the given typing symbol."""
+
     @abstractmethod
     def visit_annotation_name(self, node: ast.Name) -> None:
         """Visit a name inside an annotation."""
@@ -80,6 +85,10 @@ def visit_annotation_name(self, node: ast.Name) -> None:
     def visit_annotation_string(self, node: ast.Constant) -> None:
         """Visit a string constant inside an annotation."""
 
+    def visit_annotated_value(self, node: ast.expr) -> None:
+        """Visit a value expression of `typing.Annotated[type, value]`."""
+        return
+
     def visit(self, node: ast.AST) -> None:
         """Visit relevant child nodes on an annotation."""
         if node is None:
@@ -95,15 +104,19 @@ def visit(self, node: ast.AST) -> None:
             self.visit(node.value)
         elif isinstance(node, ast.Subscript):
             self.visit(node.value)
-            if getattr(node.value, 'id', '') == 'Annotated' and isinstance(
+            if self.is_typing(node.value, 'Literal'):
+                return
+            elif self.is_typing(node.value, 'Annotated') and isinstance(
                 (elts_node := node.slice.value if py38 and isinstance(node.slice, Index) else node.slice),
                 (ast.Tuple, ast.List),
             ):
                 if elts_node.elts:
-                    # only visit the first element
-                    self.visit(elts_node.elts[0])
-                    # TODO: We may want to visit the rest as a soft-runtime use
-            elif getattr(node.value, 'id', '') != 'Literal':
+                    elts_iter = iter(elts_node.elts)
+                    # only visit the first element like a type expression
+                    self.visit(next(elts_iter))
+                    for value_node in elts_iter:
+                        self.visit_annotated_value(value_node)
+            else:
                 self.visit(node.slice)
         elif isinstance(node, (ast.Tuple, ast.List)):
             for n in node.elts:
@@ -300,18 +313,22 @@ def visit_AsyncFunctionDef(self, node: AsyncFunctionDef) -> None:
 class SQLAlchemyAnnotationVisitor(AnnotationVisitor):
     """Adds any names in the annotation to mapped names."""
 
-    def __init__(self, mapped_names: set[str]) -> None:
-        self.mapped_names = mapped_names
+    def __init__(self, sqlalchemy_plugin: SQLAlchemyMixin) -> None:
+        self.plugin = sqlalchemy_plugin
+
+    def is_typing(self, node: ast.AST, symbol: str) -> bool:
+        """Check if the given node matches the given typing symbol."""
+        return self.plugin.is_typing(node, symbol)
 
     def visit_annotation_name(self, node: ast.Name) -> None:
         """Add name to mapped names."""
-        self.mapped_names.add(node.id)
+        self.plugin.soft_uses.add(node.id)
 
     def visit_annotation_string(self, node: ast.Constant) -> None:
         """Add all the names in the string to mapped names."""
-        visitor = StringAnnotationVisitor()
+        visitor = StringAnnotationVisitor(self.plugin)
         visitor.parse_and_visit_string_annotation(node.value)
-        self.mapped_names.update(visitor.names)
+        self.plugin.soft_uses.update(visitor.names)
 
 
 class SQLAlchemyMixin:
@@ -333,23 +350,22 @@ class SQLAlchemyMixin:
         sqlalchemy_mapped_dotted_names: set[str]
         current_scope: Scope
         uses: dict[str, list[tuple[ast.AST, Scope]]]
+        soft_uses: set[str]
 
         def in_type_checking_block(self, lineno: int, col_offset: int) -> bool:  # noqa: D102
             ...
 
         def lookup_full_name(self, node: ast.AST) -> str | None:  # noqa: D102
             ...
 
+        def is_typing(self, node: ast.AST, symbol: str) -> bool:  # noqa: D102
+            ...
+
     def __init__(self, *args: Any, **kwargs: Any) -> None:
         super().__init__(*args, **kwargs)
 
-        #: Contains a set of all names used inside `Mapped[...]` annotations
-        # These are treated like soft-uses, i.e. we don't know if it will be
-        # used at runtime or not
-        self.mapped_names: set[str] = set()
-
         #: Used for visiting annotations
-        self.sqlalchemy_annotation_visitor = SQLAlchemyAnnotationVisitor(self.mapped_names)
+        self.sqlalchemy_annotation_visitor = SQLAlchemyAnnotationVisitor(self)
 
     def visit_AnnAssign(self, node: ast.AnnAssign) -> None:
         """Remove all annotations assigments."""
@@ -403,9 +419,9 @@ def handle_sqlalchemy_annotation(self, node: ast.AST) -> None:
             # add all names contained in the inner part of the annotation
             # since this is not as strict as an actual runtime use, we don't
             # care if we record too much here
-            visitor = StringAnnotationVisitor()
+            visitor = StringAnnotationVisitor(self)
             visitor.parse_and_visit_string_annotation(inner)
-            self.mapped_names.update(visitor.names)
+            self.soft_uses.update(visitor.names)
             return
 
         # we only need to handle annotations like `Mapped[...]`
@@ -780,9 +796,14 @@ def lookup(self, symbol_name: str, use: HasPosition | None = None, runtime_only:
 class StringAnnotationVisitor(AnnotationVisitor):
     """Visit a parsed string annotation and collect all the names."""
 
-    def __init__(self) -> None:
+    def __init__(self, typing_lookup: SupportsIsTyping) -> None:
         #: All the names referenced inside the annotation
         self.names: set[str] = set()
+        self._typing_lookup = typing_lookup
+
+    def is_typing(self, node: ast.AST, symbol: str) -> bool:
+        """Check if the given node matches the given typing symbol."""
+        return self._typing_lookup.is_typing(node, symbol)
 
     def parse_and_visit_string_annotation(self, annotation: str) -> None:
         """Parse and visit the given string as an annotation expression."""
@@ -814,7 +835,7 @@ def visit_annotation_string(self, node: ast.Constant) -> None:
 class ImportAnnotationVisitor(AnnotationVisitor):
     """Map all annotations on an AST node."""
 
-    def __init__(self) -> None:
+    def __init__(self, import_visitor: ImportVisitor) -> None:
         #: All type annotations in the file, without quotes around them
         self.unwrapped_annotations: list[UnwrappedAnnotation] = []
 
@@ -831,6 +852,12 @@ def __init__(self) -> None:
         # e.g. AnnAssign.annotation within a function body will never evaluate
         self.never_evaluates = False
 
+        self.import_visitor = import_visitor
+
+    def is_typing(self, node: ast.AST, symbol: str) -> bool:
+        """Check if the given node matches the given typing symbol."""
+        return self.import_visitor.is_typing(node, symbol)
+
     def visit(
         self,
         node: ast.AST,
@@ -864,12 +891,19 @@ def visit_annotation_string(self, node: ast.Constant) -> None:
         if getattr(node, BINOP_OPERAND_PROPERTY, False):
             self.invalid_binop_literals.append(node)
         else:
-            visitor = StringAnnotationVisitor()
+            visitor = StringAnnotationVisitor(self.import_visitor)
             visitor.parse_and_visit_string_annotation(node.value)
             (self.excess_wrapped_annotations if self.never_evaluates else self.wrapped_annotations).append(
                 WrappedAnnotation(node.lineno, node.col_offset, node.value, visitor.names, self.scope, self.type)
             )
 
+    def visit_annotated_value(self, node: ast.expr) -> None:
+        """Visit a value expression of `typing.Annotated[type, value]`."""
+        previous_context = self.import_visitor.in_soft_use_context
+        self.import_visitor.in_soft_use_context = True
+        self.import_visitor.visit(node)
+        self.import_visitor.in_soft_use_context = previous_context
+
 
 class ImportVisitor(
     DunderAllMixin, AttrsMixin, InjectorMixin, FastAPIMixin, PydanticMixin, SQLAlchemyMixin, ast.NodeVisitor
@@ -932,8 +966,12 @@ def __init__(
         #: List of all names and ids, except type declarations
         self.uses: dict[str, list[tuple[ast.AST, Scope]]] = defaultdict(list)
 
+        #: Contains a set of all names to be treated like soft-uses.
+        # i.e. we don't know if it will be used at runtime or not
+        self.soft_uses: set[str] = set()
+
         #: Handles logic of visiting annotation nodes
-        self.annotation_visitor = ImportAnnotationVisitor()
+        self.annotation_visitor = ImportAnnotationVisitor(self)
 
         #: Whether there is a `from __futures__ import annotations` is present in the file
         self.futures_annotation: Optional[bool] = None
@@ -951,6 +989,10 @@ def __init__(
         #: For tracking which comprehension/IfExp we're currently inside of
         self.active_context: Optional[Comprehension | ast.IfExp] = None
 
+        #: Whether or not we're in a context where uses count as soft-uses.
+        # E.g. the value expression of `typing.Annotated[type, value]`
+        self.in_soft_use_context: bool = False
+
     @contextmanager
     def create_scope(self, node: ast.ClassDef | Function, is_head: bool = True) -> Iterator[Scope]:
         """Create a new scope."""
@@ -1853,7 +1895,7 @@ def unused_imports(self) -> Flake8Generator:
         }
 
         all_imports = {name for name, imp in self.visitor.imports.items() if not imp.exempt}
-        unused_imports = all_imports - self.visitor.names - self.visitor.mapped_names
+        unused_imports = all_imports - self.visitor.names - self.visitor.soft_uses
         used_imports = all_imports - unused_imports
         already_imported_modules = [self.visitor.imports[name].module for name in used_imports]
         annotation_names = (

flake8_type_checking/types.py

@@ -24,5 +24,9 @@ def lineno(self) -> int:
         def col_offset(self) -> int:
             pass
 
+    class SupportsIsTyping(Protocol):
+        def is_typing(self, node: ast.AST, symbol: str) -> bool:
+            pass
+
 
 ImportTypeValue = Literal['APPLICATION', 'THIRD_PARTY', 'BUILTIN', 'FUTURE']

tests/test_name_extraction.py

@@ -1,28 +1,30 @@
+import ast
 import sys
 
 import pytest
 
-from flake8_type_checking.checker import StringAnnotationVisitor
+from flake8_type_checking.checker import ImportVisitor, StringAnnotationVisitor
 
 examples = [
-    ('', set()),
-    ('invalid_syntax]', set()),
-    ('int', {'int'}),
-    ('dict[str, int]', {'dict', 'str', 'int'}),
+    ('', set(), set()),
+    ('invalid_syntax]', set(), set()),
+    ('int', {'int'}, set()),
+    ('dict[str, int]', {'dict', 'str', 'int'}, set()),
     # make sure literals don't add names for their contents
-    ('Literal["a"]', {'Literal'}),
-    ("Literal['a']", {'Literal'}),
-    ('Literal[0]', {'Literal'}),
-    ('Literal[1.0]', {'Literal'}),
-    ('Literal[True]', {'Literal'}),
-    ('T | S', {'T', 'S'}),
-    ('Union[Dict[str, Any], Literal["Foo", "Bar"], _T]', {'Union', 'Dict', 'str', 'Any', 'Literal', '_T'}),
+    ('Literal["a"]', {'Literal'}, set()),
+    ("Literal['a']", {'Literal'}, set()),
+    ('Literal[0]', {'Literal'}, set()),
+    ('Literal[1.0]', {'Literal'}, set()),
+    ('Literal[True]', {'Literal'}, set()),
+    ('L[a]', {'L'}, set()),
+    ('T | S', {'T', 'S'}, set()),
+    ('Union[Dict[str, Any], Literal["Foo", "Bar"], _T]', {'Union', 'Dict', 'str', 'Any', 'Literal', '_T'}, set()),
     # for attribute access only everything up to the first dot should count
     # this matches the behavior of add_annotation
-    ('datetime.date | os.path.sep', {'datetime', 'os'}),
-    ('Nested["str"]', {'Nested', 'str'}),
-    ('Annotated[str, validator]', {'Annotated', 'str'}),
-    ('Annotated[str, "bool"]', {'Annotated', 'str'}),
+    ('datetime.date | os.path.sep', {'datetime', 'os'}, set()),
+    ('Nested["str"]', {'Nested', 'str'}, set()),
+    ('Annotated[str, validator(int, 5)]', {'Annotated', 'str'}, {'validator', 'int'}),
+    ('Annotated[str, "bool"]', {'Annotated', 'str'}, set()),
 ]
 
 if sys.version_info >= (3, 11):
@@ -31,8 +33,20 @@
     ])
 
 
-@pytest.mark.parametrize(('example', 'expected'), examples)
-def test_name_extraction(example, expected):
-    visitor = StringAnnotationVisitor()
+@pytest.mark.parametrize(('example', 'expected', 'soft_uses'), examples)
+def test_name_extraction(example, expected, soft_uses):
+    import_visitor = ImportVisitor(
+        cwd='fake cwd',  # type: ignore[arg-type]
+        pydantic_enabled=False,
+        fastapi_enabled=False,
+        fastapi_dependency_support_enabled=False,
+        cattrs_enabled=False,
+        sqlalchemy_enabled=False,
+        sqlalchemy_mapped_dotted_names=[],
+        injector_enabled=False,
+        pydantic_enabled_baseclass_passlist=[],
+    )
+    import_visitor.visit(ast.parse('from typing import Annotated, Literal, Literal as L'))
+    visitor = StringAnnotationVisitor(import_visitor)
     visitor.parse_and_visit_string_annotation(example)
     assert visitor.names == expected

tests/test_tc001_to_tc003.py

@@ -203,25 +203,29 @@ class Migration:
                 '''),
             set(),
         ),
+        # Annotated soft use
         (
             textwrap.dedent(f'''
                 from typing import Annotated
 
                 from {import_} import Depends
 
                 x: Annotated[str, Depends]
+                y: Depends
             '''),
             set(),
         ),
+        # This is not a soft-use, it's just a plain string
         (
             textwrap.dedent(f'''
                 from typing import Annotated
 
                 from {import_} import Depends
 
                 x: Annotated[str, "Depends"]
+                y: Depends
             '''),
-            set(),
+            {'4:0 ' + ERROR.format(module=f'{import_}.Depends')},
         ),
     ]