Functions defined with `@functions_framework.http` or `@functions_framework.typed` do not support type hints

[DevJake]

Simply put, defining a function with the @functions_framework.http decorator does not support type hints. With some digging, I stumbled upon the possibility of using @functions_framework.typed, but even this does not work, nor is it documented anywhere. Please see:

1. feat: Support strongly typed functions signature #208
2. Document functions_framework.typed #290

This is taken from my attempts to create and deploy a function running python version 3.12.x. There are no error messages indicating that the cause is the type hints, only that the function failed to be ready and accepting of traffic.

I would suggest the following:

1. Type hints be properly supported for Python functions running any compatible version, given they are now standard practice for developers
2. Versions that use type hinting when they shouldn't -- or otherwise cannot be built -- show an appropriate and relevant error message

As for an example, even the most basic functions such as the following will break:

from flask import Request

@functions_framework.http
def main(request: Request):
	pass

from flask import Request

@functions_framework.typed
def main(request: Request):
	pass

[AITechXcel]

Google has some of the most talented developers and extremely deep pockets and yet somehow this was overlooked. It is not a good look.

[frou]

Forgive me if I'm misunderstanding, but is this all about the general Python gotcha where a decorator doesn't propagate the static type signature of a function it wraps, unless in the definition of the decorator the author does a little dance involving typing.ParamSpec and typing.TypeVar?

[red8888]

The decorator raises linting errors like this.

Forgive me if I'm misunderstanding, but is this all about the general Python gotcha where a decorator doesn't propagate the static type signature of a function it wraps, unless in the definition of the decorator the author does a little dance involving typing.ParamSpec and typing.TypeVar?

Haven't tested this, but suggestion from gemini.

[red8888]

I am doing this, but not sure if it will cause issues? Is there any reason why I shouldn't do this?

from collections.abc import Callable
from typing import Any, TypeVar
from functions_framework import (
    http as functions_framework_http,  # pyright: ignore[reportUnknownVariableType]
)

F = TypeVar("F", bound=Callable[..., Any])

def functions_framework_typed(func: F) -> F:
    """Wrap functions_framework.http to support type hints.

    This decorator applies the correct typing to satisfy linters like Pylance,
    preventing "partially unknown type" errors, without modifying the underlying
    behavior of the original decorator.

    The original decorator just registers the function and returns it,
    so we can call it and then return the original function, which
    allows the type checker to preserve the original function's signature.
    """
    functions_framework_http(func)
    return func

Seems to work as a drop-in replacement

@functions_framework_typed
def main(request: Request) -> Response:
    #....