Merge pull request 'Undefined variables and column aggregations' (#43) from fix/undefined-variable into main

Reviewed-on: #43

midas/checker/frames/column_groupby_methods.py

@@ -13,7 +13,6 @@ from midas.checker.types import (
     Function,
     ParamSpec,
     Type,
-    UnknownType,
 )
 
 if TYPE_CHECKING:
@@ -86,6 +85,9 @@ class ColumnGroupByMethodRegistry(MethodRegistry[Call]):
 
         # TODO: maybe better to filter arguments and pass some, in case the
         # return type depends on them
+
+        # Don't catch UndefinedMethodException because all aggregation
+        # methods should be defined on columns too
         returns: Type = self.typer.call_method(
             location=call.location,
             call_expr=call.call_expr,
@@ -94,11 +96,9 @@ class ColumnGroupByMethodRegistry(MethodRegistry[Call]):
             positional=[],
             keywords={},
         )
-        if not isinstance(returns, ColumnType):
-            returns = ColumnType(type=UnknownType())
         signature = Function(
             params=ParamSpec(mixed=real_params),
-            returns=returns,
+            returns=ColumnType(type=returns),
         )
 
         result: CallResult = self.dispatcher.get_result(

midas/checker/frames/column_methods.py

@@ -364,13 +364,11 @@ class ColumnMethodRegistry(MethodRegistry[Call]):
             Type: the result type
         """
 
-        returns: Type = ColumnType(type=TopType())
+        returns: Type = TopType()
         if formula:
-            returns = ColumnType(
-                type=self._resolve_formula_type(
-                    call,
-                    formula(call.column.type),
-                )
+            returns = self._resolve_formula_type(
+                call,
+                formula(call.column.type),
             )
         signature = Function(
             params=ParamSpec(

midas/checker/frames/frame_groupby_methods.py

@@ -53,6 +53,8 @@ class FrameGroupByMethodRegistry(MethodRegistry[Call]):
         for column in call.groupby.frame.columns:
             with self.reporter.with_context(f"in column '{column.name}'"):
                 column_groupby: ColumnGroupBy = ColumnGroupBy(column=column.type)
+                # Don't catch UndefinedMethodException because all aggregation
+                # methods should be defined on columns too
                 result_type: Type = self.typer.call_method(
                     location=call.location,
                     call_expr=call.call_expr,

midas/checker/midas.py

@@ -170,7 +170,10 @@ class MidasTyper(m.Stmt.Visitor[None], m.Expr.Visitor[Type], m.Type.Visitor[Type
             type = GenericType(name=name, params=params, body=type)
         else:
             type = DerivedType(name=name, type=type)
-        self.types.define_type(name, type)
+        try:
+            self.types.define_type(name, type)
+        except ValueError:
+            self.reporter.error(stmt.location, f"Type {name} already defined")
         self._local_variables.clear()
         self._current_name = None
 
@@ -178,7 +181,10 @@ class MidasTyper(m.Stmt.Visitor[None], m.Expr.Visitor[Type], m.Type.Visitor[Type
         name: str = stmt.name.lexeme
         self._current_name = name
         type: Type = stmt.type.accept(self)
-        self.types.define_type(name, type)
+        try:
+            self.types.define_type(name, type)
+        except ValueError:
+            self.reporter.error(stmt.location, f"Type {name} already defined")
         self._current_name = None
 
     def visit_member_stmt(self, stmt: m.MemberStmt) -> None: ...
@@ -201,6 +207,7 @@ class MidasTyper(m.Stmt.Visitor[None], m.Expr.Visitor[Type], m.Type.Visitor[Type
             )
 
     def visit_predicate_stmt(self, stmt: m.PredicateStmt) -> None:
+        name: str = stmt.name.lexeme
         for spec in stmt.params:
             for param in spec.mixed:
                 assert param.name is not None
@@ -222,14 +229,17 @@ class MidasTyper(m.Stmt.Visitor[None], m.Expr.Visitor[Type], m.Type.Visitor[Type
                     returns=type,
                 )
         self._predicate_params = {}
-        self.types.define_predicate(
-            stmt.name.lexeme,
-            Predicate(
-                type=type,
-                body=stmt.body,
-                alias=len(params) == 0,
-            ),
-        )
+        try:
+            self.types.define_predicate(
+                name,
+                Predicate(
+                    type=type,
+                    body=stmt.body,
+                    alias=len(params) == 0,
+                ),
+            )
+        except ValueError:
+            self.reporter.error(stmt.location, f"Predicate {name} already defined")
 
     def _is_valid_predicate(self, body: Type) -> bool:
         """Check whether the given type is valid as a predicate's body

midas/checker/python.py

@@ -48,7 +48,9 @@ TypedExpr = tuple[p.Expr, Type]
 
 
 class ReturnException(Exception):
-    pass
+    def __init__(self, stmt: p.Stmt):
+        super().__init__()
+        self.stmt: p.Stmt = stmt
 
 
 class UndefinedMethodException(Exception):
@@ -116,7 +118,10 @@ class PythonTyper(
         self.judgements = []
         self.evaluated_casts = []
 
-        self.check(stmts)
+        try:
+            self.check(stmts)
+        except ReturnException as e:
+            self.reporter.error(e.stmt.location, "Return statement outside of function")
 
         return TypedAST(
             stmts=stmts,
@@ -576,7 +581,7 @@ class PythonTyper(
     def visit_return_stmt(self, stmt: p.ReturnStmt) -> None:
         type: Type = self.type_of(stmt.value) if stmt.value is not None else UnitType()
         self.env.return_types.append(type)
-        raise ReturnException()
+        raise ReturnException(stmt)
 
     def visit_if_stmt(self, stmt: p.IfStmt) -> None:
         # Not evaluated in sub-environment because assignments in the test leak out of the if
@@ -599,7 +604,7 @@ class PythonTyper(
         else_returned: bool = self.process_block(stmt.orelse, env)
         self.env.return_types.extend(env.return_types)
         if body_returned and else_returned:
-            raise ReturnException()
+            raise ReturnException(stmt)
 
     def visit_pass(self, stmt: p.Pass) -> None:
         pass
@@ -627,7 +632,7 @@ class PythonTyper(
         self.env = outer_env
 
         if body_returned:
-            raise ReturnException()
+            raise ReturnException(stmt)
 
     def visit_import_stmt(self, stmt: p.ImportStmt) -> None:
         self._visit_imports(stmt.location, stmt.imports)
@@ -772,14 +777,21 @@ class PythonTyper(
         match expr.callee:
             case p.GetExpr(object=obj, name=method):
                 obj_type: Type = self.type_of(obj)
-                return self.call_method(
-                    location=expr.location,
-                    call_expr=expr,
-                    obj=(obj, obj_type),
-                    method_name=method,
-                    positional=positional,
-                    keywords=keywords,
-                )
+                try:
+                    return self.call_method(
+                        location=expr.location,
+                        call_expr=expr,
+                        obj=(obj, obj_type),
+                        method_name=method,
+                        positional=positional,
+                        keywords=keywords,
+                    )
+                except UndefinedMethodException:
+                    self.reporter.error(
+                        expr.location,
+                        f"Unknown method {method} on type {obj_type}",
+                    )
+                    return UnknownType()
 
         callee: Type = self.type_of(expr.callee)
         result: CallResult = self.dispatcher.get_result(
@@ -1001,7 +1013,11 @@ class PythonTyper(
 
         if len(node.args) != 0:
             args: list[Type] = [self.resolve_type_expr(arg) for arg in node.args]
-            return self.types.apply_generic(base, args)
+            try:
+                return self.types.apply_generic(base, args)
+            except Exception as e:
+                self.reporter.error(node.location, f"Cannot apply generic type: {e}")
+                return UnknownType()
         return base
 
     def visit_frame_column(self, node: p.FrameColumn) -> ColumnType:
@@ -1078,10 +1094,13 @@ class PythonTyper(
                 bound: Optional[Type] = None
                 variance: Variance = Variance.INVARIANT
                 if "bound" in call.keywords:
-                    bound_type: p.MidasType = self._parse_type_from_expr(
-                        call.keywords["bound"]
-                    )
-                    bound = self.resolve_type_expr(bound_type)
+                    try:
+                        bound_type: p.MidasType = self._parse_type_from_expr(
+                            call.keywords["bound"]
+                        )
+                        bound = self.resolve_type_expr(bound_type)
+                    except NotImplementedError:
+                        bound = UnknownType()
 
                 if is_kw_true("covariant"):
                     variance = Variance.COVARIANT
@@ -1096,7 +1115,13 @@ class PythonTyper(
                     else:
                         variance = Variance.CONTRAVARIANT
                 var: TypeVar = TypeVar(name=name, bound=bound, variance=variance)
-                self.types.define_type(name, var)
+                try:
+                    self.types.define_type(name, var)
+                except ValueError:
+                    self.reporter.error(
+                        call.location,
+                        f"A type or type variable with the name {name} is already defined",
+                    )
                 return var
 
             case _:
@@ -1127,6 +1152,14 @@ class PythonTyper(
                 return parser._parse_type(node.body)
             case p.VariableExpr(name=name):
                 return p.BaseType(location=location, base=name, args=())
+            case p.SubscriptExpr(object=p.VariableExpr(name=name), index=arg):
+                args: tuple[p.MidasType, ...] = (
+                    tuple(self._parse_type_from_expr(a) for a in arg.items)
+                    if isinstance(arg, p.TupleExpr)
+                    else (self._parse_type_from_expr(arg),)
+                )
+                return p.BaseType(location=location, base=name, args=args)
+
             case _:
                 raise NotImplementedError
 

midas/checker/resolver.py

@@ -97,6 +97,21 @@ class Resolver(p.Stmt.Visitor[None], p.Expr.Visitor[None]):
                 return True
         return False
 
+    def is_defined(self, name: str) -> bool:
+        """Check whether the given variable is defined
+
+        Args:
+            name (str): the name of the variable
+
+        Returns:
+            bool: `True` if the variable is defined in the closest scope where it
+                is declared, `False` otherwise
+        """
+        for scope in reversed(self.scopes):
+            if name in scope:
+                return scope[name]
+        return False
+
     def resolve_function(self, function: p.Function) -> None:
         """Resolve a function definition
 
@@ -230,7 +245,7 @@ class Resolver(p.Stmt.Visitor[None], p.Expr.Visitor[None]):
         pass
 
     def visit_variable_expr(self, expr: p.VariableExpr) -> None:
-        if len(self.scopes) != 0 and self.scopes[-1].get(expr.name) is False:
+        if self.is_declared(expr.name) and not self.is_defined(expr.name):
             self.reporter.error(
                 expr.location,
                 f"Variable '{expr.name}' is declared but may not be defined",

midas/checker/types.py

@@ -362,9 +362,6 @@ def substitute_typevars(type: Type, substitutions: dict[str, Type]) -> Type:
         case UnknownType() | UnitType():
             return type
 
-        case TopType() | GenericType():
-            raise NotImplementedError(f"Unsupported type {type}")
-
         # Ensure exhaustiveness
         case _:
             assert_never(type)

tests/cases/checker/09_frame_ops.py.ref.json

@@ -4305,9 +4305,7 @@
         "arguments": [],
         "keywords": {}
       },
-      "type": {
-        "type": {}
-      }
+      "type": {}
     },
     {
       "location": {
@@ -4342,9 +4340,7 @@
         "arguments": [],
         "keywords": {}
       },
-      "type": {
-        "type": {}
-      }
+      "type": {}
     },
     {
       "location": {
@@ -4380,9 +4376,7 @@
         "keywords": {}
       },
       "type": {
-        "type": {
-          "name": "int"
-        }
+        "name": "int"
       }
     },
     {
@@ -4419,9 +4413,7 @@
         "keywords": {}
       },
       "type": {
-        "type": {
-          "name": "float"
-        }
+        "name": "float"
       }
     },
     {
@@ -4458,9 +4450,7 @@
         "keywords": {}
       },
       "type": {
-        "type": {
-          "name": "int"
-        }
+        "name": "int"
       }
     },
     {
@@ -4497,9 +4487,7 @@
         "keywords": {}
       },
       "type": {
-        "type": {
-          "name": "int"
-        }
+        "name": "int"
       }
     },
     {
@@ -4536,9 +4524,7 @@
         "keywords": {}
       },
       "type": {
-        "type": {
-          "name": "int"
-        }
+        "name": "int"
       }
     },
     {
@@ -4575,9 +4561,7 @@
         "keywords": {}
       },
       "type": {
-        "type": {
-          "name": "int"
-        }
+        "name": "int"
       }
     },
     {
@@ -4614,9 +4598,7 @@
         "keywords": {}
       },
       "type": {
-        "type": {
-          "name": "int"
-        }
+        "name": "int"
       }
     },
     {
@@ -4652,9 +4634,7 @@
         "arguments": [],
         "keywords": {}
       },
-      "type": {
-        "type": {}
-      }
+      "type": {}
     },
     {
       "location": {
@@ -4690,9 +4670,7 @@
         "keywords": {}
       },
       "type": {
-        "type": {
-          "name": "int"
-        }
+        "name": "int"
       }
     },
     {
@@ -4728,9 +4706,7 @@
         "arguments": [],
         "keywords": {}
       },
-      "type": {
-        "type": {}
-      }
+      "type": {}
     },
     {
       "location": {