保留signatu的Python decorator（闭包版本）

Question

我的decorator是“闭包式”的；它在返回被修饰的函数之前做一些工作。你知道吗

借用这个著名的问题：Preserving signatures of decorated functions

def args_as_ints(f):

    time.sleep(1) # hard at work

    def g(*args, **kwargs):
        args = [int(x) for x in args]
        kwargs = dict((k, int(v)) for k, v in kwargs.items())
        return f(*args, **kwargs)
    return g

functools.wraps在python2中不保留签名。你知道吗

from functools import wraps

def args_as_ints(f):

    time.sleep(1) # hard at work

    @wraps(f) 
    def g(*args, **kwargs):
        args = [int(x) for x in args]
        kwargs = dict((k, int(v)) for k, v in kwargs.items())
        return f(*args, **kwargs)
    return g


@args_as_ints
def funny_function(x, y, z=3):
    """Computes x*y + 2*z"""
    return x*y + 2*z

help(funny_function)

演出

Help on function funny_function in module __main__:

funny_function(*args, **kwargs)
    Computes x*y + 2*z

decorator模块似乎不支持这种装饰器样式。你知道吗

同样相关：Preserve Signature in Decorator python 2

Answer 1

您可以使用^{}模块。你知道吗

请记住，wrapt修饰符与标准python修饰符具有不同的接口。我强烈建议您阅读^{} documentation。总之，这里是用wrapt.decorator实现装饰器的基本方法：

import wrapt

def args_as_ints(f):
    time.sleep(1) # hard at work

    @wrapt.decorator
    def g(f, instance, *args, **kwargs):
        args = [int(x) for x in args]
        kwargs = dict((k, int(v)) for k, v in kwargs.items())
        return f(*args, **kwargs)

    return g(f)  # apply the decorator to the function

请注意，我的实现完全忽略了instance参数，因此对于实例方法它无法正常工作。你知道吗

但是，它确实保留了修饰函数的签名：

Help on function funny_function in module __main__:

funny_function(x, y, z=3)
    Computes x*y + 2*z

Answer 2

设法在这里找到答案：https://decorator.readthedocs.io/en/latest/tests.documentation.html#dealing-with-third-party-decorators

def args_as_ints(f):

    time.sleep(1)

    def g(*args, **kwargs):
        args = [int(x) for x in args]
        kwargs = dict((k, int(v)) for k, v in kwargs.items())
        return f(*args, **kwargs)

    return decorator.FunctionMaker.create(
        f, 'return decfunc(%(signature)s)',
        dict(decfunc=g, __wrapped__=f))