I have a dictionary that I use as my “config”. I don’t change values, so it could be immutable. I use its values as arguments for functions via key indexing, something like this:

from typing import Union

my_config: dict[str, Union[int, float, str]] = {
    "timeout": 60,
    "precision": 3.14,
    "greeting": "Hello",
}

def my_function(arg: int):
    print(f"The value is: {arg}")

my_function(arg=my_config["timeout"])

Since mypy is a static type checker, it cannot infer the types inside a dictionary. So it will complain:

Argument “arg” to “my_function” has incompatible type “Union[int, float, str]”; expected “int”

What is the proper way to deal with this situation? Some ideas:

from typing import cast
# type cast before call
my_function(arg=int(my_dict["key1"]))
my_function(arg=cast(int, my_dict["key1"]))

Is there more? What is the preferred way to deal with this? Do I have to go back a step and change my approach of using a dict for my_config? Maybe use a truly immutable data type in the first place?