hydra.resolution module

Type dereference, lookup, requirements, and instantiation.

hydra.resolution.dereference_type(cx: T0, graph: Graph, name: Name) → object

Dereference a type name to get the actual type (Either version).

hydra.resolution.f_type_is_polymorphic(typ: Type) → bool

Test whether a given System F type is polymorphic (i.e., a forall type).

hydra.resolution.field_map(fields: Sequence[Field]) → Mapping[Name, Term]

Build a map from field name to field term, given a list of fields.

hydra.resolution.field_type_map(fields: Sequence[FieldType]) → Mapping[Name, Type]

Build a map from field name to field type, given a list of field types.

hydra.resolution.field_types(cx: T0, graph: Graph, t: Type) → object

Get field types from a record or union type (Either version).

hydra.resolution.find_field_type(cx: T0, fname: Name, fields: Sequence[FieldType]) → object

Find a field type by name in a list of field types.

hydra.resolution.fully_strip_and_normalize_type(typ: Type) → Type

Fully strip a type of forall quantifiers, normalizing bound variable names for alpha-equivalence comparison.

hydra.resolution.fully_strip_type(typ: Type) → Type

Fully strip a type of forall quantifiers.

hydra.resolution.instantiate_type(cx: InferenceContext, typ: Type) → tuple[Type, InferenceContext]

Instantiate a type by replacing all forall-bound type variables with fresh variables, threading InferenceContext.

hydra.resolution.instantiate_type_scheme(cx: InferenceContext, scheme: TypeScheme) → tuple[TypeScheme, InferenceContext]

Instantiate a type scheme with fresh variables, threading InferenceContext.

hydra.resolution.map_key_resolves_to_string(types: Mapping[Name, Type], key_type: Type) → bool

Test whether a map’s key type resolves to string, following aliases and wrappers. Used to decide whether a map may use the compact JSON object encoding.

hydra.resolution.nominal_application(tname: Name, args: Sequence[Type]) → Type

Apply type arguments to a nominal type.

hydra.resolution.require_record_type(cx: T0, graph: Graph, name: Name) → object

Require a name to resolve to a record type.

hydra.resolution.require_row_type(cx: T0, label: str, getter: Callable[[Type], object], graph: Graph, name: Name) → object

Require a name to resolve to a row type.

hydra.resolution.require_schema_type(cx: InferenceContext, types: Mapping[Name, TypeScheme], tname: Name) → object

Look up a schema type and instantiate it, threading InferenceContext.

hydra.resolution.require_type(cx: T0, graph: Graph, name: Name) → object

Require a type by name.

hydra.resolution.require_union_field(cx: T0, graph: Graph, tname: Name, fname: Name) → object

Require a field type from a union type.

hydra.resolution.require_union_type(cx: T0, graph: Graph, name: Name) → object

Require a name to resolve to a union type.

hydra.resolution.resolve_base_type(types: Mapping[Name, Type], typ: Type) → Type

Resolve a type to its base, following annotations, type aliases (Type.variable dereferenced against the given type map), and Type.wrap bodies, transitively. An unresolvable variable name (not found in the map) is returned as-is, since the caller is in the best position to decide whether that is an error. As with hydra.sources.kernel.terms.lexical.dereferenceSchemaType, this assumes type aliases form an acyclic reference graph, as guaranteed by schema construction; a genuinely self-referential alias is not a supported input and would not terminate.

hydra.resolution.resolve_type(graph: Graph, typ: Type) → object

Resolve a type, dereferencing type variables.

hydra.resolution.type_to_type_scheme(t0: Type) → TypeScheme

Convert a (System F -style) type to a type scheme.