PEP-6731 introduces the Self type and it’s coming to Python 3.11. However, you can already use that now via the typing_extenstions2 module.

The Self type makes annotating methods that return the instances of the corresponding classes trivial. Before this, you’d have to do some mental gymnastics to statically type situations as follows:

# src.py
from __future__ import annotations

from typing import Any


class Animal:
    def __init__(self, name: str, says: str) -> None:
        self.name = name
        self.says = says

    @classmethod
    def from_description(cls, description: str = "|") -> Animal:
        descr = description.split("|")
        return cls(descr[0], descr[1])


class Dog(Animal):
    def __init__(self, *args: Any, **kwargs: Any) -> None:
        super().__init__(*args, **kwargs)

    @property
    def legs(self) -> int:
        return 4


if __name__ == "__main__":
    dog = Dog.from_description("Matt | woof")
    print(dog.legs)  # Mypy complains here!

The class Animal has a from_description class method that acts as an additional constructor. It takes a description string, and then builds and returns an instance of the same class. The return type of the method is annotated as Animal here. However, doing this makes the child class Dog conflate its identity with the Animal class. If you execute the snippet, it won’t raise any runtime error. Also, Mypy will complain about the type:

src.py:27: error: "Animal" has no attribute "legs"
        print(dog.legs)  # Mypy complains here!
              ^
Found 1 error in 1 file (checked 1 source file)

To fix this, we’ll have to make sure that the return type of the from_description class method doesn’t confuse the type checker. This is one way to do this:

from __future__ import annotations

from typing import TypeVar

T = TypeVar("T", bound="Animal")


class Animal:
    def __init__(self, name: str, says: str) -> None:
        self.name = name
        self.says = says

    @classmethod  # In <Python3.9, Use typing.Type[T].
    def from_description(cls: type[T], description: str = "|") -> T:
        descr = description.split("|")
        return cls(descr[0], descr[1])


...

In the above snippet, first I had to declare a TypeVar and bind that to the Animal class. Then I had to explicitly type the cls variable in the from_description method. This time, the type checker will be happy. While this isn’t a lot of work, it surely goes against the community convention. Usually, we don’t explicitly type the self, cls variables and instead, let the type checker figure out their types. Also, subjectively, this sticks out like a sore thumb.

PEP-673 allows us to solve the issue elegantly:

# src.py
from __future__ import annotations

import sys

if sys.version_info >= (3, 11):
    from typing import Self
else:
    from typing_extensions import Self


class Animal:
    def __init__(self, name: str, says: str) -> None:
        self.name = name
        self.says = says

    @classmethod
    def from_description(cls, description: str = "|") -> Self:
        descr = description.split("|")
        return cls(descr[0], descr[1])


...

If you run Mypy against the second snippet, it won’t complain.

Typing instance methods that return self

Take a look at this:

# src.py
from __future__ import annotations

import sys

if sys.version_info >= (3, 11):
    from typing import Self
else:
    from typing_extensions import Self


class Counter:
    def __init__(self, start: int = 1) -> None:
        self.val = start

    def increment(self) -> Self:
        self.val += 1
        return self

    def decrement(self) -> Self:
        self.val -= 1
        return self

The increment and decrement method of the Counter class return the instance of the same class after performing the operations on the start value. This is a perfect case where the Self type can be useful.

Typing __new__ methods

You can also type the __new__ method easily:

from __future__ import annotations

import sys

if sys.version_info >= (3, 11):
    from typing import Self
else:
    from typing_extensions import Self

from typing import Any


class Config:
    def __new__(cls, var: int, *args: Any, **kwargs: Any) -> Self:
        """Validate the value before constructing the class."""

        if not 0 <= var < 10:
            raise TypeError(
                "'var' must be a positive integer between 0 and 9",
            )
        return super().__new__(cls)

    def __init__(self, var: int) -> None:
        self.var = var

The __new__ method in the Config class validates the var before constructing an instance of the class. The Self type makes it easy to annotate the method.

Recent posts

  • TypeIs does what I thought TypeGuard would do in Python
  • ETag and HTTP caching
  • Crossing the CORS crossroad
  • Dysfunctional options pattern in Go
  • Einstellung effect
  • Strategy pattern in Go
  • Anemic stack traces in Go
  • Retry function in Go
  • Type assertion vs type switches in Go
  • Patching pydantic settings in pytest