I'm using Pydantic to create a number of custom types with validation and serialization logic. They look like this:

MyType = Annotated[BackedType, myvalidator(arg), myserializer(arg), myschema(arg)]

where myvalidator, myserializer, myschema return instances of BeforeValidator, PlainSerializer and GetPydanticSchema respectively. That seems to be the preferred way to do it, ref documentation.

Now a lot of this is repetitive, so I would like to do something like this:

MyType = Annotated[BackedType, *everything_in_one(arg)]

But that's illegal syntax in a type parameter context: can't splat there. So I thought, since Annotated[BackedType, ...] is equivalent to BackedType in a type checking context, can I not do something like this?

def make_annotated(arg) -> Type[BackedType]:
    return Annotated[BackedType, ...etc]

MyType = make_annotated(arg)  # mypy understands that this is BackedType.

But I can't get this to type check. Apparently mypy thinks the return value from the function is just object, and if I cast it, mypy claims that this is not valid as a type - nor does it work if i annotate it as a TypeAlias (the expression is an invalid type alias).

Is there a way to make this sort of thing work?