引文与描述：
Adding arbitrary metadata annotations to Python functions and variables
说说我的体会：
类似编译的作用，能够帮助你尽早地避免错误
1. 不支持 Python2+
>>> def test_annotation_py2(a_str: str):
File "<stdin>", line 1
def test_annotation_py2(a_str: str):
^
SyntaxError: invalid syntax
2. 代码检查，而且写的时候很容易，并且可以被 IDE 如 Pycharm 支持
3. 基本用法
>>> # all is python built-in type (single)
... def search_for(neddle: str, haystack: str) -> int:
... offset = haystack.find(needle)
... return offset
...
>>> # More complicated types
...
>>> # Python3.5 added the `typing` module, which both gives us a bunch of new names
... # for types, and tools to build our own types
...
>>> from typing import List
>>> def multisearch(needle: str, haystack: str) -> List[int]:
... offset = haystack.find(needle)
... if offset == -1:
... return []
... else:
... return [offset] + multisearch(needle, haystack[offset+1:])
...
>>> # In func multisearch, we define a new type List[int], `List` is from `typeing`, `int` is python built-in type.
# There are many of these -e.g. Dict[keytype, valuetype], if you need more, you can view `typing` documentation
...
>>> # A func reteurn different type, use `Union`
...
>>> from typing import Union
>>> def search_for(needle: str, haystack: str) -> Union[int, None]:
... offset = haystack.find(needle)
... if offset == -1:
... return None
... else:
... return offset
3. 注意事项
# 使用的是 [] 而不是 (): typing.List[] 而不是 typing.List()
# 类型混合，比如返回的是 (int, None) 或者是 (int, str)，那么可以写为
# typing.Tuple[int, typing.Union[str, None]] 或者
# typing.Union[typing.Tuple[int, str], typing.Tuple[int, None]]
有一个疑问，这样写与静态语言有什么区别？都是在运行前检查。
参考：