RFC: Add MypyTypeInferenceProvider

This change is RFC (please read whole change message).

Add `MypyTypeInferenceProvider` as an alternative for
`TypeInferenceProvider`. The provider infers types using mypy as
library. The only requirement for the usage is to have the latest mypy
installed. Types inferred are mypy types, since mypy type system is well
designed, to avoid the conversion, and also to keep it simple. For
compatibility and extensibility reasons, these types are stored in
separate field `MypyType.mypy_type`.

Let's assume we have the following code in the file `x.py` which we want
to inspect:
```python
x = [42]

s = set()

from enum import Enum

class E(Enum):
    f = "f"

e = E.f
```

Then to get play with mypy types one should use the code like:
```python
import libcst as cst

from libcst.metadata import MypyTypeInferenceProvider

filename = "x.py"
module = cst.parse_module(open(filename).read())
cache = MypyTypeInferenceProvider.gen_cache(".", [filename])[filename]
wrapper = cst.MetadataWrapper(
    module,
    cache={MypyTypeInferenceProvider: cache},
)

mypy_type = wrapper.resolve(MypyTypeInferenceProvider)
x_name_node = wrapper.module.body[0].body[0].targets[0].target
set_call_node = wrapper.module.body[1].body[0].value
e_name_node = wrapper.module.body[-1].body[0].targets[0].target

print(mypy_type[x_name_node])
 # prints: builtins.list[builtins.int]

print(mypy_type[x_name_node].fullname)
 # prints: builtins.list[builtins.int]

print(mypy_type[x_name_node].mypy_type.type.fullname)
 # prints: builtins.list

print(mypy_type[x_name_node].mypy_type.args)
 # prints: (builtins.int,)

print(mypy_type[x_name_node].mypy_type.type.bases[0].type.fullname)
 # prints: typing.MutableSequence

print(mypy_type[set_call_node])
 # prints: builtins.set

print("issuperset" in mypy_type[set_call_node].mypy_type.names)
 # prints: True

print(mypy_type[set_call_node.func])
 # prints: typing.Type[builtins.set]

print(mypy_type[e_name_node].mypy_type.type.is_enum)
 # prints: True
```

Why?

1. `TypeInferenceProvider` requires pyre (`pyre-check` on PyPI) to be
   installed. mypy is more popular than pyre. If the organization uses
   mypy already (which is almost always the case), it may be difficult
   to assure colleagues (including security team) that "we need yet
   another type checker". `MypyTypeInferenceProvider` requires the
   latest mypy only.
2. Even though it is possible to run pyre without watchman installation,
   this is not advertised. watchman installation is not always possible
   because of system requirements, or because of the security
   requirements like "we install only our favorite GNU/Linux
   distribution packages".
3. `TypeInferenceProvider` usage requires `pyre start` command to be run
   before the execution, and `pyre stop` - after the execution. This may
   be inconvenient, especially for the cases when pyre was not used
   before.
4. Types produced by pyre in `TypeInferenceProvider` are just strings.
   For example, it's not easily possible to infer that some variable is
   enum instance. `MypyTypeInferenceProvider` makes it easy, see the
   code above.

Drawback:

1. Speed. mypy is slower than pyre, so is `MypyTypeInferenceProvider`
   comparing to `TypeInferenceProvider`.
   How to partially solve this:
   1. Implement AST tree caching in mypy. It may be difficult, however
      this will lead to speed improvements for all the projects that use
      this functionality.
   2. Implement inferred types caching inside LibCST. As far as I know,
      no caching at all is implemented inside LibCST, which is the
      prerequisite for inferred types caching, so the task is big.
   3. Implement LibCST CST to mypy AST. I am not sure if this possible
      at all. Even if it is possible, the task is huge.
2. Two providers are doing similar things in LibCST will be present,
   this can potentially lead to the situation when there is a need
   install two typecheckers to get all codemods from the library
   running.
   Alternatives considered:
   1. Put `MypyTypeInferenceProvider` inside separate library (say,
       LibCST-mypy or `libcst-mypy` on PyPI). This will explicitly
       separate `MypyTypeInferenceProvider` from the rest of LibCST.
      Drawbacks:
      1. The need to maintain separate library.
      2. Limited fame (people need to know that the library exists).
      3. Since some codemods cannot be implemented easily without the
         library, for example, `if-elif-else` to `match` converter
	 (it needs powerful type inference), they are doomed to not be
	 shipped with LibCST, which makes the latter less attractive for
	 end users.
   2. Implement base class for inferred type, which inherits from `str`
      (to keep the compatibility with the existing codebase) and
      the mechanism for dynamically selecting `TypeInferenceProvider`
      typechecker (mypy or pyre; user can do this via enviromental
      variable). If the code inside LibCST requires just shallow type
      information (so, just `str` is enough), then the code can run with
      any typechecker. The remaining code (such as `if-elif-else` to
      `match` converter) will still require mypy.

Misc:

Code does not lint in my env, by some reason `pyre check` cannot find
`mypy` library.

Related to:

* Instagram#451
* pyastrx/pyastrx#40
* python/mypy#12513
* python/mypy#4868

Author: Roman Inflianskas

libcst/metadata/__init__.py

@@ -17,6 +17,7 @@
     ExpressionContextProvider,
 )
 from libcst.metadata.full_repo_manager import FullRepoManager
+from libcst.metadata.mypy_type_inference_provider import MypyTypeInferenceProvider
 from libcst.metadata.name_provider import (
     FullyQualifiedNameProvider,
     QualifiedNameProvider,
@@ -74,6 +75,7 @@
     "ClassScope",
     "ComprehensionScope",
     "ScopeProvider",
+    "MypyTypeInferenceProvider",
     "ParentNodeProvider",
     "QualifiedName",
     "QualifiedNameSource",

libcst/metadata/mypy_type_inference_provider.py

@@ -0,0 +1,96 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the MIT license found in the
+# LICENSE file in the root directory of this source tree.
+
+from pathlib import Path
+from typing import Dict, List, Mapping, Optional, TYPE_CHECKING
+
+import libcst as cst
+from libcst._position import CodeRange
+from libcst.helpers import calculate_module_and_package
+from libcst.metadata.base_provider import BatchableMetadataProvider
+from libcst.metadata.position_provider import PositionProvider
+
+try:
+    import mypy
+
+    MYPY_INSTALLED = True
+except ImportError:
+    MYPY_INSTALLED = False
+
+
+if TYPE_CHECKING:
+    import mypy.nodes
+
+    import libcst.metadata.mypy_utils
+
+
+def raise_on_mypy_non_installed() -> None:
+    if not MYPY_INSTALLED:
+        raise RuntimeError("mypy is not installed, please install it")
+
+
+class MypyTypeInferenceProvider(
+    BatchableMetadataProvider["libcst.metadata.mypy_utils.MypyType"]
+):
+    """
+    Access inferred type annotation through `mypy <http://mypy-lang.org/>`_.
+    """
+
+    METADATA_DEPENDENCIES = (PositionProvider,)
+
+    @classmethod
+    def gen_cache(
+        cls, root_path: Path, paths: List[str], timeout: Optional[int] = None
+    ) -> Mapping[
+        str, Optional["libcst.metadata.mypy_utils.MypyTypeInferenceProviderCache"]
+    ]:
+        raise_on_mypy_non_installed()
+
+        import mypy.build
+        import mypy.main
+
+        from libcst.metadata.mypy_utils import MypyTypeInferenceProviderCache
+
+        targets, options = mypy.main.process_options(paths)
+        options.preserve_asts = True
+        options.fine_grained_incremental = True
+        options.use_fine_grained_cache = True
+        mypy_result = mypy.build.build(targets, options=options)
+        cache = {}
+        for path in paths:
+            module = calculate_module_and_package(str(root_path), path).name
+            cache[path] = MypyTypeInferenceProviderCache(
+                module_name=module,
+                mypy_file=mypy_result.graph[module].tree,
+            )
+        return cache
+
+    def __init__(
+        self,
+        cache: Optional["libcst.metadata.mypy_utils.MypyTypeInferenceProviderCache"],
+    ) -> None:
+        from libcst.metadata.mypy_utils import CodeRangeToMypyNodesBinder
+
+        super().__init__(cache)
+        self._mypy_node_locations: Dict[CodeRange, "mypy.nodes.Node"] = {}
+        if cache is None:
+            return
+        code_range_to_mypy_nodes_binder = CodeRangeToMypyNodesBinder(cache.module_name)
+        code_range_to_mypy_nodes_binder.visit_mypy_file(cache.mypy_file)
+        self._mypy_node_locations = code_range_to_mypy_nodes_binder.locations
+
+    def _parse_metadata(self, node: cst.CSTNode) -> None:
+        range = self.get_metadata(PositionProvider, node)
+        if range in self._mypy_node_locations:
+            self.set_metadata(node, self._mypy_node_locations.get(range))
+
+    def visit_Name(self, node: cst.Name) -> Optional[bool]:
+        self._parse_metadata(node)
+
+    def visit_Attribute(self, node: cst.Attribute) -> Optional[bool]:
+        self._parse_metadata(node)
+
+    def visit_Call(self, node: cst.Call) -> Optional[bool]:
+        self._parse_metadata(node)

libcst/metadata/mypy_utils.py

@@ -0,0 +1,145 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the MIT license found in the
+# LICENSE file in the root directory of this source tree.
+from dataclasses import dataclass, field
+from typing import Dict, Optional, Union
+
+import mypy.build
+import mypy.main
+import mypy.modulefinder
+import mypy.nodes
+import mypy.options
+import mypy.patterns
+import mypy.traverser
+import mypy.types
+import mypy.typetraverser
+
+from libcst._add_slots import add_slots
+from libcst._position import CodePosition, CodeRange
+
+
+@add_slots
+@dataclass(frozen=True)
+class MypyTypeInferenceProviderCache:
+    module_name: str
+    mypy_file: mypy.nodes.MypyFile
+
+
+@add_slots
+@dataclass(frozen=True)
+class MypyType:
+    is_type_constructor: bool
+    mypy_type: Optional[Union[mypy.types.Type, mypy.nodes.TypeInfo]] = None
+    fullname: str = field(init=False)
+
+    def __post_init__(self) -> None:
+        if isinstance(self.mypy_type, mypy.types.Type):
+            fullname = str(self.mypy_type)
+        else:
+            fullname = self.mypy_type.fullname
+        if self.is_type_constructor:
+            fullname = f"typing.Type[{fullname}]"
+        object.__setattr__(self, "fullname", fullname)
+
+    def __str__(self) -> str:
+        return self.fullname
+
+
+class CodeRangeToMypyNodesBinder(
+    mypy.traverser.TraverserVisitor, mypy.typetraverser.TypeTraverserVisitor
+):
+    def __init__(self, module_name: str) -> None:
+        super().__init__()
+        self.locations: Dict[CodeRange, MypyType] = {}
+        self.in_type_alias_expr = False
+        self.module_name = module_name
+
+    # Helpers
+
+    @staticmethod
+    def get_code_range(o: mypy.nodes.Context) -> CodeRange:
+        return CodeRange(
+            start=CodePosition(o.line, o.column),
+            end=CodePosition(o.end_line, o.end_column),
+        )
+
+    @staticmethod
+    def check_bounds(o: mypy.nodes.Context) -> bool:
+        return (
+            (o.line is not None)
+            and (o.line >= 1)
+            and (o.column is not None)
+            and (o.column >= 0)
+            and (o.end_line is not None)
+            and (o.end_line >= 1)
+            and (o.end_column is not None)
+            and (o.end_column >= 0)
+        )
+
+    def record_type_location_using_code_range(
+        self,
+        code_range: CodeRange,
+        t: Optional[Union[mypy.types.Type, mypy.nodes.TypeInfo]],
+        is_type_constructor: bool,
+    ) -> None:
+        if t is not None:
+            self.locations[code_range] = MypyType(
+                is_type_constructor=is_type_constructor, mypy_type=t
+            )
+
+    def record_type_location(
+        self,
+        o: mypy.nodes.Context,
+        t: Optional[Union[mypy.types.Type, mypy.nodes.TypeInfo]],
+        is_type_constructor: bool,
+    ) -> None:
+        if self.check_bounds(o):
+            self.record_type_location_using_code_range(
+                code_range=self.get_code_range(o),
+                t=t,
+                is_type_constructor=is_type_constructor,
+            )
+
+    def record_location_by_name_expr(
+        self, code_range: CodeRange, o: mypy.nodes.NameExpr, is_type_constructor: bool
+    ) -> None:
+        if isinstance(o.node, mypy.nodes.Var):
+            self.record_type_location_using_code_range(
+                code_range=code_range, t=o.node.type, is_type_constructor=False
+            )
+        elif isinstance(o.node, mypy.nodes.TypeInfo):
+            self.record_type_location_using_code_range(
+                code_range=code_range, t=o.node, is_type_constructor=is_type_constructor
+            )
+
+    # Actual visitors
+
+    def visit_var(self, o: mypy.nodes.Var) -> None:
+        super().visit_var(o)
+        self.record_type_location(o=o, t=o.type, is_type_constructor=False)
+
+    def visit_name_expr(self, o: mypy.nodes.NameExpr) -> None:
+        super().visit_name_expr(o)
+        # Implementation in base classes is omitted, record it if it is variable or class
+        self.record_location_by_name_expr(
+            self.get_code_range(o), o, is_type_constructor=True
+        )
+
+    def visit_member_expr(self, o: mypy.nodes.MemberExpr) -> None:
+        super().visit_member_expr(o)
+        # Implementation in base classes is omitted, record it
+        # o.def_var should not be None after mypy run, checking here just to be sure
+        if o.def_var is not None:
+            self.record_type_location(o=o, t=o.def_var.type, is_type_constructor=False)
+
+    def visit_call_expr(self, o: mypy.nodes.CallExpr) -> None:
+        super().visit_call_expr(o)
+        if isinstance(o.callee, mypy.nodes.NameExpr):
+            self.record_location_by_name_expr(
+                code_range=self.get_code_range(o), o=o.callee, is_type_constructor=False
+            )
+
+    def visit_instance(self, o: mypy.types.Instance) -> None:
+        super().visit_instance(o)
+        self.record_type_location(o=o, t=o, is_type_constructor=False)

libcst/metadata/tests/test_mypy_type_inference_provider.py

@@ -0,0 +1,65 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the MIT license found in the
+# LICENSE file in the root directory of this source tree.
+
+import sys
+from pathlib import Path
+from unittest import skipIf
+
+import libcst as cst
+from libcst import MetadataWrapper
+from libcst.metadata.mypy_type_inference_provider import MypyTypeInferenceProvider
+from libcst.testing.utils import data_provider, UnitTest
+from libcst.tests.test_pyre_integration import TEST_SUITE_PATH
+
+
+def _test_simple_class_helper(test: UnitTest, wrapper: MetadataWrapper) -> None:
+    mypy_nodes = wrapper.resolve(MypyTypeInferenceProvider)
+    m = wrapper.module
+    assign = cst.ensure_type(
+        cst.ensure_type(
+            cst.ensure_type(
+                cst.ensure_type(m.body[1].body, cst.IndentedBlock).body[0],
+                cst.FunctionDef,
+            ).body.body[0],
+            cst.SimpleStatementLine,
+        ).body[0],
+        cst.AnnAssign,
+    )
+    self_number_attr = cst.ensure_type(assign.target, cst.Attribute)
+    test.assertEqual(str(mypy_nodes[self_number_attr]), "builtins.int")
+
+    # self
+    test.assertEqual(
+        str(mypy_nodes[self_number_attr.value]), "libcst.tests.pyre.simple_class.Item"
+    )
+    collector_assign = cst.ensure_type(
+        cst.ensure_type(m.body[3], cst.SimpleStatementLine).body[0], cst.Assign
+    )
+    collector = collector_assign.targets[0].target
+    test.assertEqual(
+        str(mypy_nodes[collector]), "libcst.tests.pyre.simple_class.ItemCollector"
+    )
+    items_assign = cst.ensure_type(
+        cst.ensure_type(m.body[4], cst.SimpleStatementLine).body[0], cst.AnnAssign
+    )
+    items = items_assign.target
+    test.assertEqual(
+        str(mypy_nodes[items]), "typing.Sequence[libcst.tests.pyre.simple_class.Item]"
+    )
+
+
+class MypyTypeInferenceProviderTest(UnitTest):
+    @data_provider(
+        ((TEST_SUITE_PATH / "simple_class.py", TEST_SUITE_PATH / "simple_class.json"),)
+    )
+    def test_simple_class_types(self, source_path: Path, data_path: Path) -> None:
+        file = str(source_path)
+        repo_root = Path(__file__).parents[3]
+        cache = MypyTypeInferenceProvider.gen_cache(repo_root, [file])
+        wrapper = MetadataWrapper(
+            cst.parse_module(source_path.read_text()),
+            cache={MypyTypeInferenceProvider: cache[file]},
+        )
+        _test_simple_class_helper(self, wrapper)

requirements-dev.txt

@@ -7,6 +7,7 @@ hypothesis>=4.36.0
 hypothesmith>=0.0.4
 jupyter>=1.0.0
 maturin>=0.8.3,<0.14
+mypy>=0.991
 nbsphinx>=0.4.2
 prompt-toolkit>=2.0.9
 pyre-check==0.9.9; platform_system != "Windows"