fix(checker): allow inferred return to by subtype of hint

gen/python.py

@@ -92,6 +92,10 @@ class ForStmt:
     body: list[Stmt]
 
 
+class RawStmt:
+    stmt: ast.stmt
+
+
 ###<
 
 
@@ -164,4 +168,8 @@ class SliceExpr:
     step: Optional[Expr]
 
 
+class RawExpr:
+    expr: ast.expr
+
+
 ###<

midas/ast/printer.py

@@ -613,6 +613,11 @@ class PythonAstPrinter(
                         self._mark_last()
                     body_stmt.accept(self)
 
+    def visit_raw_stmt(self, stmt: p.RawStmt) -> None:
+        self._write_line("RawStmt")
+        with self._child_level(single=True):
+            self._write_line(f"stmt: {ast.unparse(stmt.stmt)}")
+
     def visit_binary_expr(self, expr: p.BinaryExpr) -> None:
         self._write_line("BinaryExpr")
         with self._child_level():
@@ -756,3 +761,8 @@ class PythonAstPrinter(
             self._write_optional_child("lower", expr.lower)
             self._write_optional_child("upper", expr.upper)
             self._write_optional_child("step", expr.step, last=True)
+
+    def visit_raw_expr(self, expr: p.RawExpr) -> None:
+        self._write_line("RawExpr")
+        with self._child_level(single=True):
+            self._write_line(f"expr: {ast.unparse(expr.expr)}")

midas/ast/python.py

@@ -113,6 +113,9 @@ class Stmt(ABC):
         @abstractmethod
         def visit_for_stmt(self, stmt: ForStmt) -> T: ...
 
+        @abstractmethod
+        def visit_raw_stmt(self, stmt: RawStmt) -> T: ...
+
 
 @dataclass(frozen=True)
 class ExpressionStmt(Stmt):
@@ -202,6 +205,14 @@ class ForStmt(Stmt):
         return visitor.visit_for_stmt(self)
 
 
+@dataclass(frozen=True)
+class RawStmt(Stmt):
+    stmt: ast.stmt
+
+    def accept(self, visitor: Stmt.Visitor[T]) -> T:
+        return visitor.visit_raw_stmt(self)
+
+
 ###############
 # Expressions #
 ###############
@@ -254,6 +265,9 @@ class Expr(ABC):
         @abstractmethod
         def visit_slice_expr(self, expr: SliceExpr) -> T: ...
 
+        @abstractmethod
+        def visit_raw_expr(self, expr: RawExpr) -> T: ...
+
 
 @dataclass(frozen=True)
 class BinaryExpr(Expr):
@@ -373,3 +387,11 @@ class SliceExpr(Expr):
 
     def accept(self, visitor: Expr.Visitor[T]) -> T:
         return visitor.visit_slice_expr(self)
+
+
+@dataclass(frozen=True)
+class RawExpr(Expr):
+    expr: ast.expr
+
+    def accept(self, visitor: Expr.Visitor[T]) -> T:
+        return visitor.visit_raw_expr(self)

midas/checker/python.py

@@ -252,7 +252,7 @@ class PythonTyper(
         if returns_hint is not None:
             assert stmt.returns is not None
             returns = returns_hint
-            if returns != inferred_return:
+            if not self.is_subtype(inferred_return, returns):
                 self.reporter.error(
                     stmt.returns.location,
                     f"Return type mismatch, annotated {returns} but returns {inferred_return}",
@@ -370,6 +370,9 @@ class PythonTyper(
         if body_returned:
             raise ReturnException()
 
+    def visit_raw_stmt(self, stmt: p.RawStmt) -> None:
+        pass
+
     def visit_binary_expr(self, expr: p.BinaryExpr) -> Type:
         method: Optional[str] = OPERATOR_METHODS.get(expr.operator.__class__)
         if method is None:
@@ -566,6 +569,9 @@ class PythonTyper(
     def visit_slice_expr(self, expr: p.SliceExpr) -> Type:
         return self.types.get_type("slice")
 
+    def visit_raw_expr(self, expr: p.RawExpr) -> Type:
+        return UnknownType()
+
     def visit_base_type(self, node: p.BaseType) -> Type:
         base: Type
         try:

midas/checker/resolver.py

@@ -163,6 +163,9 @@ class Resolver(p.Stmt.Visitor[None], p.Expr.Visitor[None]):
         self.resolve(*stmt.body)
         self.end_scope()
 
+    def visit_raw_stmt(self, stmt: p.RawStmt) -> None:
+        pass
+
     def visit_binary_expr(self, expr: p.BinaryExpr) -> None:
         self.resolve(expr.left)
         self.resolve(expr.right)
@@ -221,3 +224,6 @@ class Resolver(p.Stmt.Visitor[None], p.Expr.Visitor[None]):
             self.resolve(expr.upper)
         if expr.step is not None:
             self.resolve(expr.step)
+
+    def visit_raw_expr(self, expr: p.RawExpr) -> None:
+        pass

midas/generator/generator.py

@@ -1,11 +1,33 @@
 import ast
 import shutil
+from dataclasses import dataclass, field
 from pathlib import Path
 
+from midas.ast.location import Location
 import midas.ast.python as p
+from midas.checker.types import (
+    AliasType,
+    AppliedType,
+    BaseType,
+    ComplexType,
+    ExtensionType,
+    Function,
+    GenericType,
+    OverloadedFunction,
+    TopType,
+    Type,
+    TypeVar,
+    UnitType,
+)
 from midas.utils import TypedAST
 
 
+@dataclass
+class Scope:
+    pre_assertions: list[ast.stmt] = field(default_factory=list)
+    aliases: list[str] = field(default_factory=list)
+
+
 class Generator(p.Stmt.Visitor[ast.stmt], p.Expr.Visitor[ast.expr]):
     def __init__(self, workdir: Path) -> None:
         self.workdir: Path = workdir.resolve()
@@ -13,19 +35,28 @@ class Generator(p.Stmt.Visitor[ast.stmt], p.Expr.Visitor[ast.expr]):
         if self.build_dir.exists():
             shutil.rmtree(self.build_dir)
         self.build_dir.mkdir(parents=True, exist_ok=True)
+        self.rel_src_path: Path = Path()
+
+        self._typed_ast: TypedAST = TypedAST(
+            stmts=[],
+            judgements=[],
+        )
+        self._alias_count: int = 0
+        self._scopes: list[Scope] = []
 
     def generate(self, typed_ast: TypedAST, src_path: Path) -> Path:
+        self.rel_src_path = src_path.relative_to(self.workdir)
+        self._typed_ast = typed_ast
         body: list[ast.stmt] = self._visit_body(typed_ast.stmts)
         module = ast.Module(body=body, type_ignores=[])
         module = ast.fix_missing_locations(module)
         compiled: str = ast.unparse(module)
-        rel_src_path: Path = src_path.relative_to(self.workdir)
+        out_path: Path = (self.build_dir / self.rel_src_path).resolve()
-        out_path: Path = (self.build_dir / rel_src_path).resolve()
         try:
             _ = out_path.relative_to(self.build_dir)
         except ValueError:
             raise ValueError(
-                f"Directory traversal, {rel_src_path} points outside of parent directory"
+                f"Directory traversal, {self.rel_src_path} points outside of parent directory"
             )
         out_path.parent.mkdir(parents=True, exist_ok=True)
         out_path.write_text(compiled)
@@ -80,8 +111,13 @@ class Generator(p.Stmt.Visitor[ast.stmt], p.Expr.Visitor[ast.expr]):
         )
 
     def visit_cast_expr(self, expr: p.CastExpr) -> ast.expr:
-        # TODO: insert assertion
+        expr2: ast.expr = expr.expr.accept(self)
-        return expr.expr.accept(self)
+        alias: ast.expr = self._make_alias(expr2)
+
+        type: Type = self._get_expr_type(expr)
+        self._make_cast_asserts(expr.location, alias, type)
+
+        return alias
 
     def visit_ternary_expr(self, expr: p.TernaryExpr) -> ast.expr:
         return ast.IfExp(
@@ -108,6 +144,9 @@ class Generator(p.Stmt.Visitor[ast.stmt], p.Expr.Visitor[ast.expr]):
             step=expr.step.accept(self) if expr.step is not None else None,
         )
 
+    def visit_raw_expr(self, expr: p.RawExpr) -> ast.expr:
+        return expr.expr
+
     def visit_expression_stmt(self, stmt: p.ExpressionStmt) -> ast.stmt:
         return ast.Expr(
             value=stmt.expr.accept(self),
@@ -169,5 +208,120 @@ class Generator(p.Stmt.Visitor[ast.stmt], p.Expr.Visitor[ast.expr]):
             orelse=[],
         )
 
+    def visit_raw_stmt(self, stmt: p.RawStmt) -> ast.stmt:
+        return stmt.stmt
+
     def _visit_body(self, stmts: list[p.Stmt]) -> list[ast.stmt]:
-        return [stmt.accept(self) for stmt in stmts]
+        generated: list[ast.stmt] = []
+        for stmt in stmts:
+            scope = Scope()
+            self._scopes.append(scope)
+
+            stmt2 = stmt.accept(self)
+            generated.extend(scope.pre_assertions)
+            generated.append(stmt2)
+            if len(scope.aliases) != 0:
+                generated.append(
+                    ast.Delete(targets=[ast.Name(id=alias) for alias in scope.aliases])
+                )
+            self._scopes.pop()
+
+        # Remove redundant pass statements
+        if len(generated) > 1:
+            generated = [stmt for stmt in generated if not isinstance(stmt, ast.Pass)]
+        return generated
+
+    def _make_alias(self, expr: ast.expr) -> ast.expr:
+        name: str = f"__midas_alias_{self._alias_count}__"
+        alias = ast.Name(id=name)
+        self._alias_count += 1
+        self._scopes[-1].aliases.append(name)
+        self._scopes[-1].pre_assertions.append(
+            ast.Assign(
+                targets=[alias],
+                value=expr,
+            )
+        )
+        return alias
+
+    def _add_assert(self, expr: ast.expr, message: str | ast.expr):
+        if isinstance(message, str):
+            message = ast.Constant(value=message)
+        self._scopes[-1].pre_assertions.append(
+            ast.Assert(
+                test=expr,
+                msg=message,
+            )
+        )
+
+    def _get_expr_type(self, query: p.Expr) -> Type:
+        for expr, type in self._typed_ast.judgements:
+            if expr == query:
+                return type
+        raise RuntimeError(f"Cannot get type judgement for {query}")
+
+    def _make_cast_asserts(self, src_location: Location, expr: ast.expr, type: Type):
+        match type:
+            case BaseType(name=name):
+                self._add_assert(
+                    ast.Call(
+                        func=ast.Name(id="isinstance"),
+                        args=[expr, ast.Name(id=name)],
+                        keywords=[],
+                    ),
+                    self._make_cast_assert_message(src_location, expr, type),
+                )
+
+            case AliasType(type=base):
+                self._make_cast_asserts(src_location, expr, base)
+
+            case UnitType():
+                self._add_assert(
+                    ast.Compare(
+                        left=expr,
+                        ops=[ast.Is()],
+                        comparators=[
+                            ast.Constant(value=None),
+                        ],
+                    ),
+                    self._make_cast_assert_message(src_location, expr, type),
+                )
+
+            case AppliedType():
+                self._make_cast_asserts(src_location, expr, type.body)
+
+            case (
+                TopType()
+                | Function()
+                | OverloadedFunction()
+                | ComplexType()
+                | ExtensionType()
+                | GenericType()
+            ):
+                raise NotImplementedError(f"Can't make assertion for type {type}")
+
+            case TypeVar():
+                raise RuntimeError("Unexpected TypeVar")
+
+    def _make_cast_assert_message(
+        self, location: Location, expr: ast.expr, type: Type
+    ) -> ast.expr:
+        loc_str: str = f"{self.rel_src_path}:L{location.lineno}:{location.col_offset+1}"
+        # f"file.py:L1:1: CastError: Cannot cast {type(expr).__name__} to Type"
+        return ast.JoinedStr(
+            values=[
+                ast.Constant(f"{loc_str}: CastError: Cannot cast "),
+                ast.FormattedValue(
+                    value=ast.Attribute(
+                        value=ast.Call(
+                            func=ast.Name(id="type"),
+                            args=[expr],
+                            keywords=[],
+                        ),
+                        attr="__name__",
+                    ),
+                    conversion=-1,
+                ),
+                ast.Constant(f" to {type}"),
+            ]
+        )

midas/parser/python.py

@@ -22,6 +22,8 @@ from midas.ast.python import (
     LiteralExpr,
     LogicalExpr,
     MidasType,
+    RawExpr,
+    RawStmt,
     ReturnStmt,
     SliceExpr,
     Stmt,
@@ -99,7 +101,7 @@ class PythonParser:
 
             case _:
                 print(f"Unsupported statement: {ast.unparse(node)}")
-                return None
+                return RawStmt(location=location, stmt=node)
 
     def parse_annotation_assign(self, node: ast.AnnAssign) -> list[Stmt]:
         statements: list[Stmt] = []
@@ -461,7 +463,8 @@ class PythonParser:
                 )
 
             case _:
-                raise UnsupportedSyntaxError(node)
+                print(f"Unsupported expression: {ast.unparse(node)}")
+                return RawExpr(location=location, expr=node)
 
     def parse_bool_op(self, node: ast.BoolOp) -> LogicalExpr:
         op: ast.boolop = node.op

tests/cases/python-parser/01_simple_types.py.ref.json

@@ -1,5 +1,9 @@
 {
   "stmts": [
+    {
+      "_type": "RawStmt",
+      "stmt": "from __future__ import annotations"
+    },
     {
       "_type": "TypeAssign",
       "name": "df",

tests/cases/python-parser/02_custom_types.py.ref.json

@@ -1,5 +1,9 @@
 {
   "stmts": [
+    {
+      "_type": "RawStmt",
+      "stmt": "from __future__ import annotations"
+    },
     {
       "_type": "TypeAssign",
       "name": "df",

tests/cases/python-parser/03_functions.py.ref.json

@@ -1,5 +1,9 @@
 {
   "stmts": [
+    {
+      "_type": "RawStmt",
+      "stmt": "from __future__ import annotations"
+    },
     {
       "_type": "Function",
       "name": "func",

tests/serializer/python.py

@@ -22,6 +22,8 @@ from midas.ast.python import (
     LogicalExpr,
     MidasType,
     Pass,
+    RawExpr,
+    RawStmt,
     ReturnStmt,
     SliceExpr,
     Stmt,
@@ -191,6 +193,12 @@ class PythonAstJsonSerializer(
             "body": self._serialize_list(stmt.body),
         }
 
+    def visit_raw_stmt(self, stmt: RawStmt) -> dict:
+        return {
+            "_type": "RawStmt",
+            "stmt": ast.unparse(stmt.stmt),
+        }
+
     def visit_binary_expr(self, expr: BinaryExpr) -> dict:
         return {
             "_type": "BinaryExpr",
@@ -284,3 +292,9 @@ class PythonAstJsonSerializer(
             "upper": self._serialize_optional(expr.upper),
             "step": self._serialize_optional(expr.step),
         }
+
+    def visit_raw_expr(self, expr: RawExpr) -> dict:
+        return {
+            "_type": "RawExpr",
+            "expr": ast.unparse(expr.expr),
+        }