Skip to content

🦆 Better duck-typing with mypy-compatible extensions to Protocol

License

Notifications You must be signed in to change notification settings

exoriente/quacks

 
 

Repository files navigation

🦆 Quacks

If it walks like a duck and it quacks like a duck, then it must be a duck

Thanks to PEP544, Python now has protocols: a way to define duck typing statically. This library gives you some niceties to make common idioms easier.

Installation

pip install quacks

⚠️ For type checking to work with mypy, you'll need to enable the plugin in your mypy config file:

[mypy]
plugins = quacks.mypy

Features

Easy read-only protocols

Defining read-only protocols is great for encouraging immutability and working with frozen dataclasses. Use the readonly decorator:

from quacks import readonly

@readonly
class User(Protocol):
    id: int
    name: str
    is_premium: bool

Without this decorator, we'd have to write quite a lot of cruft, reducing readability:

class User(Protocol):
    @property
    def id(self) -> int: ...
    @property
    def name(self) -> str: ...
    @property
    def is_premium(self) -> bool: ...

Partial protocols (work in progress)

What if you want to reuse parts of a protocol? Imagine we have several functions who use various properties of User. With partial protocols you can reuse attributes without having to define many overlapping protocols. Inspired by clojure spec.

(exact syntax TBD)

class User(Protocol):
    id: int
    name: str
    is_premium: bool
    address: Address

class Address(Protocol):
    street: str
    city: str
    country: str

from quacks import _

def determine_discount(u: User[_.id.is_premium]) -> int:
    ...  # access `id` and `is_premium` attributes

def greet(u: User[_.name.address[_.country]]) -> None:
    ...  # access `name` and `address.country` attributes

u: User = ...

determine_discount(u)
greet(u)

About

🦆 Better duck-typing with mypy-compatible extensions to Protocol

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 89.7%
  • Makefile 10.3%