Python dynamic decorators - why so many wraps?

Question

So I\'m still kind of new to Python decorators - I\'ve used them before, but I\'ve never made my own. I\'m reading this tutorial (that particular paragraph) and I don\'t seem to

Answer 1

A decorator modifies a function by adding a wrapper to it. At the time you decorate the function, it isn't being called yet, so you don't have any arguments (or keyword arguments) to look at. All you can do for now is create a new function that will handle those arguments when it finally gets them.

Answer 2

Well, what would happen if you called that decorator on a function?

@decorator
def foo(): pass

This code would immediately call foo, which we don't want. Decorators are called and their return value replaces the function. It's the same as saying

def foo(): pass
foo = decorator(foo)

So if we have a decorator that calls foo, we probably want to have a function that returns a function that calls foo -- that function that it returns will replace foo.

def decorator(f):
    def g(*args, **kwargs):
        return f(*args, **kwargs)
    return g

Now, if we want to pass options to the decorator, we can't exactly pass them ins ide-by-side with the function like in your example. There's no syntax for it. So we define a function that returns a parameterized decorator. The decorator it returns will be a closure.

def argument_decorator(will_I_call_f):
    def decorator(f):
        def g(*args, **kwargs):
            if will_I_call_f: return f(*args, **kwargs)
        return g
    return decorator

so we can do

decorator = argument_decorator(True)
@decorator
def foo(): pass

And Python offers the convenience syntax where you inline the function call:

@argument_decorator(True)
def foo(): pass

And all this is syntax sugar for the non-decorator syntax of

def foo(): pass
foo = argument_decorator(True)(foo)