This chapter is a work in progress; it's probably better if you don't begin making changes until I've finished the original version, which is being posted as a series on my weblog.
This amazing feature appeared in the language almost apologetically and with concern that it might not be that useful.
I predict that in time it will be seen as one of the more powerful features in the language. The problem is that all the introductions to decorators that I have seen have been rather confusing, so I will try to rectify that here.
Decorators vs. the Decorator Pattern
First, you need to understand that the word "decorator" was used with some trepidation in Python, because there was concern that it would be completely confused with the Decorator pattern from the Design Patterns book. At one point other terms were considered for the feature, but "decorator" seems to be the one that sticks.
Indeed, you can use Python decorators to implement the Decorator pattern, but that's an extremely limited use of it. Python decorators, I think, are best equated to macros.
History of Macros
The macro has a long history, but most people will probably have had experience with C preprocessor macros. The problems with C macros were (1) they were in a different language (not C) and (2) the behavior was sometimes bizarre, and often inconsistent with the behavior of the rest of C.
Both Java and C# have added annotations, which allow you to do some things to elements of the language. Both of these have the problems that (1) to do what you want, you sometimes have to jump through some enormous and untenable hoops, which follows from (2) these annotation features have their hands tied by the bondage-and-discipline (or as Martin Fowler gently puts it: "Directing") nature of those languages.
In a slightly different vein, many C++ programmers (myself included) have noted the generative abilities of C++ templates and have used that feature in a macro- like fashion.
Many other languages have incorporated macros, but without knowing much about it I will go out on a limb and say that Python decorators are similar to Lisp macros in power and possibility.
The Goal of Macros
I think it's safe to say that the goal of macros in a language is to provide a way to modify elements of the language. That's what decorators do in Python -- they modify functions, and in the case of class decorators, entire classes. This is why they usually provide a simpler alternative to metaclasses.
The major failings of most language's self-modification approaches are that they are too restrictive and that they require a different language (I'm going to say that Java annotations with all the hoops you must jump through to produce an interesting annotation comprises a "different language").
Python falls into Fowler's category of "enabling" languages, so if you want to do modifications, why create a different or restricted language? Why not just use Python itself? And that's what Python decorators do.
What Can You Do With Decorators?
Decorators allow you to inject or modify code in functions or classes. Sounds a bit like Aspect-Oriented Programming (AOP) in Java, doesn't it? Except that it's both much simpler and (as a result) much more powerful. For example, suppose you'd like to do something at the entry and exit points of a function (such as perform some kind of security, tracing, locking, etc. -- all the standard arguments for AOP). With decorators, it looks like this:
@entryExit def func1(): print("inside func1()") @entryExit def func2(): print("inside func2()")
The @ indicates the application of the decorator.
A function decorator is applied to a function definition by placing it on the line before that function definition begins. For example:
@myDecorator def aFunction(): print("inside aFunction")
When the compiler passes over this code, aFunction() is compiled and the resulting function object is passed to the myDecorator code, which does something to produce a function-like object that is then substituted for the original aFunction().
What does the myDecorator code look like? Well, most introductory examples show this as a function, but I've found that it's easier to start understanding decorators by using classes as decoration mechanisms instead of functions. In addition, it's more powerful.
The only constraint upon the object returned by the decorator is that it can be used as a function -- which basically means it must be callable. Thus, any classes we use as decorators must implement __call__.
What should the decorator do? Well, it can do anything but usually you expect the original function code to be used at some point. This is not required, however:
# PythonDecorators/my_decorator.py class my_decorator(object): def __init__(self, f): print("inside my_decorator.__init__()") f() # Prove that function definition has completed def __call__(self): print("inside my_decorator.__call__()") @my_decorator def aFunction(): print("inside aFunction()") print("Finished decorating aFunction()") aFunction()
When you run this code, you see:
inside my_decorator.__init__() inside aFunction() Finished decorating aFunction() inside my_decorator.__call__()
Notice that the constructor for my_decorator is executed at the point of decoration of the function. Since we can call f() inside __init__(), it shows that the creation of f() is complete before the decorator is called. Note also that the decorator constructor receives the function object being decorated. Typically, you'll capture the function object in the constructor and later use it in the __call__() method (the fact that decoration and calling are two clear phases when using classes is why I argue that it's easier and more powerful this way).
When aFunction() is called after it has been decorated, we get completely different behavior; the my_decorator.__call__() method is called instead of the original code. That's because the act of decoration replaces the original function object with the result of the decoration -- in our case, the my_decorator object replaces aFunction. Indeed, before decorators were added you had to do something much less elegant to achieve the same thing:
def foo(): pass foo = staticmethod(foo)
With the addition of the @ decoration operator, you now get the same result by saying:
@staticmethod def foo(): pass
This is the reason why people argued against decorators, because the @ is just a little syntax sugar meaning "pass a function object through another function and assign the result to the original function."
The reason I think decorators will have such a big impact is because this little bit of syntax sugar changes the way you think about programming. Indeed, it brings the idea of "applying code to other code" (i.e.: macros) into mainstream thinking by formalizing it as a language construct.
Slightly More Useful
Now let's go back and implement the first example. Here, we'll do the more typical thing and actually use the code in the decorated functions:
# PythonDecorators/entry_exit_class.py class entry_exit(object): def __init__(self, f): self.f = f def __call__(self): print("Entering", self.f.__name__) self.f() print("Exited", self.f.__name__) @entry_exit def func1(): print("inside func1()") @entry_exit def func2(): print("inside func2()") func1() func2()
The output is:
Entering func1 inside func1() Exited func1 Entering func2 inside func2() Exited func2
You can see that the decorated functions now have the "Entering" and "Exited" trace statements around the call.
The constructor stores the argument, which is the function object. In the call, we use the __name__ attribute of the function to display that function's name, then call the function itself.
Using Functions as Decorators
The only constraint on the result of a decorator is that it be callable, so it can properly replace the decorated function. In the above examples, I've replaced the original function with an object of a class that has a __call__() method. But a function object is also callable, so we can rewrite the previous example using a function instead of a class, like this:
# PythonDecorators/entry_exit_function.py def entry_exit(f): def new_f(): print("Entering", f.__name__) f() print("Exited", f.__name__) return new_f @entry_exit def func1(): print("inside func1()") @entry_exit def func2(): print("inside func2()") func1() func2() print(func1.__name__)
new_f() is defined within the body of entry_exit(), so it is created and returned when entry_exit() is called. Note that new_f() is a closure, because it captures the actual value of f.
Once new_f() has been defined, it is returned from entry_exit() so that the decorator mechanism can assign the result as the decorated function.
The output of the line print(func1.__name__) is new_f, because the new_f function has been substituted for the original function during decoration. If this is a problem you can change the name of the decorator function before you return it:
def entry_exit(f): def new_f(): print("Entering", f.__name__) f() print("Exited", f.__name__) new_f.__name__ = f.__name__ return new_f
The information you can dynamically get about functions, and the modifications you can make to those functions, are quite powerful in Python.
Review: Decorators without Arguments
If we create a decorator without arguments, the function to be decorated is passed to the constructor, and the __call__() method is called whenever the decorated function is invoked:
# PythonDecorators/decorator_without_arguments.py class decorator_without_arguments(object): def __init__(self, f): """ If there are no decorator arguments, the function to be decorated is passed to the constructor. """ print("Inside __init__()") self.f = f def __call__(self, *args): """ The __call__ method is not called until the decorated function is called. """ print("Inside __call__()") self.f(*args) print("After self.f(*args)") @decorator_without_arguments def sayHello(a1, a2, a3, a4): print('sayHello arguments:', a1, a2, a3, a4) print("After decoration") print("Preparing to call sayHello()") sayHello("say", "hello", "argument", "list") print("After first sayHello() call") sayHello("a", "different", "set of", "arguments") print("After second sayHello() call")
Any arguments for the decorated function are just passed to __call__(). The output is:
Inside __init__() After decoration Preparing to call sayHello() Inside __call__() sayHello arguments: say hello argument list After self.f(*args) After first sayHello() call Inside __call__() sayHello arguments: a different set of arguments After self.f(*args) After second sayHello() call
Notice that __init__() is the only method called to perform decoration, and __call__() is called every time you call the decorated sayHello().
Decorators with Arguments
The decorator mechanism behaves quite differently when you pass arguments to the decorator.
Let's modify the above example to see what happens when we add arguments to the decorator:
# PythonDecorators/decorator_with_arguments.py class decorator_with_arguments(object): def __init__(self, arg1, arg2, arg3): """ If there are decorator arguments, the function to be decorated is not passed to the constructor! """ print("Inside __init__()") self.arg1 = arg1 self.arg2 = arg2 self.arg3 = arg3 def __call__(self, f): """ If there are decorator arguments, __call__() is only called once, as part of the decoration process! You can only give it a single argument, which is the function object. """ print("Inside __call__()") def wrapped_f(*args): print("Inside wrapped_f()") print("Decorator arguments:", self.arg1, self.arg2, self.arg3) f(*args) print("After f(*args)") return wrapped_f @decorator_with_arguments("hello", "world", 42) def sayHello(a1, a2, a3, a4): print('sayHello arguments:', a1, a2, a3, a4) print("After decoration") print("Preparing to call sayHello()") sayHello("say", "hello", "argument", "list") print("after first sayHello() call") sayHello("a", "different", "set of", "arguments") print("after second sayHello() call")
From the output, we can see that the behavior changes quite significantly:
Inside __init__() Inside __call__() After decoration Preparing to call sayHello() Inside wrapped_f() Decorator arguments: hello world 42 sayHello arguments: say hello argument list After f(*args) after first sayHello() call Inside wrapped_f() Decorator arguments: hello world 42 sayHello arguments: a different set of arguments After f(*args) after second sayHello() call
Now the process of decoration calls the constructor and then immediately invokes __call__(), which can only take a single argument (the function object) and must return the decorated function object that replaces the original. Notice that __call__() is now only invoked once, during decoration, and after that the decorated function that you return from __call__() is used for the actual calls.
Although this behavior makes sense -- the constructor is now used to capture the decorator arguments, but the object __call__() can no longer be used as the decorated function call, so you must instead use __call__() to perform the decoration -- it is nonetheless surprising the first time you see it because it's acting so much differently than the no-argument case, and you must code the decorator very differently from the no-argument case.
Decorator Functions with Decorator Arguments
Finally, let's look at the more complex decorator function implementation, where you have to do everything all at once:
# PythonDecorators/decorator_function_with_arguments.py def decorator_function_with_arguments(arg1, arg2, arg3): def wrap(f): print("Inside wrap()") def wrapped_f(*args): print("Inside wrapped_f()") print("Decorator arguments:", arg1, arg2, arg3) f(*args) print("After f(*args)") return wrapped_f return wrap @decorator_function_with_arguments("hello", "world", 42) def sayHello(a1, a2, a3, a4): print('sayHello arguments:', a1, a2, a3, a4) print("After decoration") print("Preparing to call sayHello()") sayHello("say", "hello", "argument", "list") print("after first sayHello() call") sayHello("a", "different", "set of", "arguments") print("after second sayHello() call")
Here's the output:
Inside wrap() After decoration Preparing to call sayHello() Inside wrapped_f() Decorator arguments: hello world 42 sayHello arguments: say hello argument list After f(*args) after first sayHello() call Inside wrapped_f() Decorator arguments: hello world 42 sayHello arguments: a different set of arguments After f(*args) after second sayHello() call
The return value of the decorator function must be a function used to wrap the function to be decorated. That is, Python will take the returned function and call it at decoration time, passing the function to be decorated. That's why we have three levels of functions; the inner one is the actual replacement function.
Because of closures, wrapped_f() has access to the decorator arguments arg1, arg2 and arg3, without having to explicitly store them as in the class version. However, this is a case where I find "explicit is better than implicit," so even though the function version is more succinct I find the class version easier to understand and thus to modify and maintain.
- More examples of decorators. Note the number of these examples that use classes rather than functions as decorators.
- Matt Wilson's Decorators Are Fun.
- Another introduction to decorators.
- Philip Eby introduces decorators.
- Class Decorators.
- Michele Simionato's decorator module wraps functions for you. The page includes an introduction and some examples.
- Kevin Djangoor's replacement for make; heavy use of decorators.
- Doug Hellman describes the experience of converting from make to paver.
- Class decorators