在运行时将参数传递给装饰器

Question

如果我像这样创建一个python decorator函数

def retry_until_true(tries, delay=60):
    """
    Decorator to rety a function or method until it returns True.
    """
    def deco_retry(f):
        def f_retry(*args, **kwargs):
            mtries  = tries
            rv = f(*args, **kwargs)
            while mtries > 0:
                if rv is True:
                    return True
                mtries -= 1
                time.sleep(delay)
                rv = f(*args, **kwargs)
            return False
        return f_retry
    return deco_retry

我可以像这样使用它

    @retry_until_true(20, delay=30)
    def check_something_function(x, y):
        ...
        return True

但有没有办法在运行时将“尝试”和“延迟”的不同值传递给装饰器，以便20和30是变量？

Answer 1

您可以使用类作为装饰器，具有tries和delay实例变量：

class RetryUntilTrue(object):
    def __init__(self, f=None, tries=10, delay=30):
        self.f = f
        self.tries = tries
        self.delay = delay

    def __call__(self, *args, **kwargs):
        if self.f is None:
            self.f = args[0]
        else:
            tries = self.tries
            while tries:
                if self.f(*args, **kwargs):
                    return True

                tries -= 1
                time.sleep(self.delay)

用法：

@RetryUntilTrue
def foo(x):
    pass

foo.tries = 20
foo.delay = 1

@RetryUntilTrue(tries=100, delay=9999)
def bar(x):
    pass

bar.tries -= 1
bar.delay = foo.delay

Answer 2

当然可以，只需将函数定义嵌套在另一个函数中，例如：

def explicit_setup_func(tries, delay=60):
    @retry_until_true(tries, delay)
    def check_something_function(x, y):
        # Code

但是，类装饰器解决方案更实用。