python 类型标注
1. 基本用法
from typing import List, Dict, Callable, Tuple
def list_word() -> List[str]:
return ["a", "b"]
def fix_max(cache: Dict[str, Tuple[float, float]], weight_func: Callable[[Tuple[float, float]], float]) -> str:
""" """
if not cache:
return
before_list = [(key, value) for key, value in cache.items()]
sorted_list = sorted(before_list, key=lambda x: weight_func(x[1]), reverse=True)
return sorted_list[0][0]
assert fix_max(cache={"A": (1, 2), "B": (1, 5), "C": (2, 2)}, weight_func=lambda x: 10 * x[0] + x[1]) == "C"
2. 别名
类型别名可用于简化复杂类型签名, 例如:
from typing import Dict, Tuple, Sequence
ConnectionOptions = Dict[str, str]
Address = Tuple[str, int]
Server = Tuple[Address, ConnectionOptions]
def broadcast_message(message: str, servers: Sequence[Server]) -> None:
...
# The static type checker will treat the previous type signature as
# being exactly equivalent to this one.
def broadcast_message(
message: str,
servers: Sequence[Tuple[Tuple[str, int], Dict[str, str]]]) -> None:
...
3. NewType
使用 NewType() 辅助函数创建不同的类型:
from typing import NewType
UserId = NewType('UserId', int)
some_id = UserId(524313)
4. Callable
期望特定签名的回调函数的框架可以将类型标注为 Callable[[Arg1Type, Arg2Type], ReturnType]。
def feeder(get_next_item: Callable[[], str]) -> None:
# Body
def async_query(on_success: Callable[[int], None],
on_error: Callable[[int, Exception], None]) -> None:
# Body
通过用文字省略号替换类型提示中的参数列表： Callable[…，ReturnType]，可以声明可调用的返回类型，而无需指定调用签名。
5. 标注类和实例
from typing import Type
class A(object):
pass
class B(A):
pass
class C(A):
pass
def get_a(a_cls: Type[A]) -> A:
return a_cls()