Couldn't seem to find a definitive answer. I want to do a type hint for a function and the type being some custom class that I have defined, called it CustomClass().
And then let's say in some function, call it FuncA(arg), I have one argument named arg. Would the correct way to type hint FuncA be:
def FuncA(arg: CustomClass):
Or would it be:
from typing import Type
def FuncA(Arg:Type[CustomClass]):
The former is correct, if arg accepts an instance of CustomClass:
def FuncA(arg: CustomClass):
# ^ instance of CustomClass
In case you want the class CustomClass itself (or a subtype), then you should write:
from typing import Type # you have to import Type
def FuncA(arg: Type[CustomClass]):
# ^ CustomClass (class object) itself
Like it is written in the documentation about Typing:
class typing.Type(Generic[CT_co])A variable annotated with
Cmay accept a value of typeC. In contrast, a variable annotated withType[C]may accept values that are classes themselves - specifically, it will accept the class object ofC.
The documentation includes an example with the int class:
a = 3 # Has type 'int' b = int # Has type 'Type[int]' c = type(a) # Also has type 'Type[int]'
Update 2024: Type is now deprecated in favour of type
def foo(bar: 'Qux') is equivalent to def foo(bar: Qux) except that it does not require loading the type immediately.
BaseEstimator and a ClassifierMixin? Can type take both so def foo(a: type[BaseEstimator, ClassifierMixin]): and if so, does the order matter?Willem Van Onsem's answer is of course correct, but I'd like to offer a small update. In PEP 585, type hinting generics were introduced in standard collections. For example, whereas we previously had to say e.g.
from typing import Dict
foo: Dict[str, str] = { "bar": "baz" }
we can now forgo the parallel type hierarchy in the typing module and simply say
foo: dict[str, str] = { "bar": "baz" }
This feature is available in python 3.9+, and also in 3.7+ if using from __future__ import annotations.
In terms of this specific question, it means that instead of from typing import Type, we can now simply annotate classes using the built-in type:
def FuncA(arg: type[CustomClass]):
typing.Type shown in Willem Van Onsem's answer is deprecated.