Decorators are a structural design pattern which wrap the original object, to add extra functionality dynamically, without modifying the original object. Decorators are usually used to run code before and/or after a function. Thus, they decorate the function with additional functionality, without changing the original function. They can also be used to alter the functionality of methods and classes.

Creating Decorators

Functions in Python are first class objects: they can be assigned to variables, passed into functions as arguments, returned from functions and declared inside functions.

With that in mind, a decorator can be created using functions or classes.

Decorator using Functions

The general syntax for creating a decorator using functions is as follows:

def decorator_name(original_function):
def wrapper():
# Code can be executed before the original function call
# Code can be executed after the original function call
return wrapper

The original_function(), in this case, is the function that is decorated with the decorator_name().

For example, a decorator that upper cases the returned values from any function can be created as follows:

# Define a function called upper_case_decorator() which will act as a decorator with parameter original_function
def upper_case_decorator(original_function):
# Creating the wrapper to wrap additional code around the original function
def wrapper():
# Code can be executed before the original function
print(f"'{original_function.__name__}' function execution will start now.")
# Original function call with the upper cased argument
result = original_function(arg1)
# Code can be executed after the original function
result = result.upper()
return result
# Return the wrapper function
return wrapper

Decorator using Classes

The general syntax for creating a decorator using classes is as follows:

class decorator:
def __init__(self, original_function):
self.original_function = original_function
def __call__(self):
# Code can be executed before the original function
# Original function call with the upper cased argument
# Code can be executed after the original function

For example, a decorator created using class based syntax that upper cases the return values from any function:

Applying Decorator to a Function

The decorators created above accept an argument called original_function which can be any function. Therefore, we create a greeting() function and decorate it with the @upper_case_decorator.

def greeting():
return "hello Dharma"

greeting() function can be decorated with the common python syntax by calling the decorator function and assigning the result back to greeting.

# Without using the @decorator syntax
greeting = upper_case_decorator(greeting)

However, Python provides a simpler way to achieve the same result using the @ symbol syntax. Add the line @upper_case_decorator at the start of function definition to decorate it.

# Using the @decorator syntax
def greeting():
return 'hello Dharma'

@upper_case_decorator or @upper_case_decorator_with_class can be used to decorate other functions as well.

def farewell():
return 'bye'
# Output: BYE

The full code with a decorator created using function will be:

Passing arguments to the Decorated functions

Arguments can be passed to the original function by creating parameters in the wrapper() function.

Eg: arg1 is the parameter that the original function greeting() requires. Thus, the wrapper() function receives it and passes it along to the original function call.

General purpose decorators can be also be created that can accept any number of arguments by packing positional arguments using *args and by packing keyword arguments using **kwargs.

def general_decorator(original_function):
# accepting variable number of arguments
def wrapper(*args, **kwargs):
func(args, kwargs)
return wrapper
def original_function(args, kwargs):
# code for the function

An example of variable arguments accepting decorator is as follows:

Passing Arguments to Decorators

Arguments can also be passed to the decorator itself. To pass arguments to the decorator, there should be a decorator_wrapper_to_accept_decorator_arguments() function wrapping the original decorator.

def decorator_wrapper_to_accept_decorator_arguments(decorator_arg1, decorator_arg2):
def decorator(original_function):
def wrapper(*args, **kwargs):
original_function(args, kwargs)
return wrapper
return decorator
@decorator_wrapper_to_accept_decorator_arguments('value 1', 'value 2')
def function_name(arg):
# statements

Eg: With the greeting function, the decorator could be setup to use a number_of_times_to_greet parameter to greet once or more than once.

def upper_case_decorator_wrapper(number_of_times_to_greet):
def upper_case_decorator(original_function):
def wrapper(*arg1, **kwargs):
result = ""
for i in range(number_of_times_to_greet):
result += original_function(*arg1, **kwargs)
result = result.upper()
return result
return wrapper
return upper_case_decorator
def greeting(arg1):
return arg1 + "!"
def greeting_twice(arg1):
return arg1 + "!"
# output:
# HI!
# output:
# HI!HI!

Even though, greeting() and greeting_twice() have the exact same function definition, their outputs are different because the decorator is executing these functions as per the number_of_times_to_greet argument passed to the decorator.

help() with Decorators

The original function name and docstring are lost when it’s wrapped by a decorator.

The name of the function is greeting() and not wrapper(). The decorator completely hides the function and it’s docstring. This also means that the help(greeting) call will be completely useless since it’ll have no information about the greeting() function.

To solve this problem, python provides a decorator. @functools.wraps() decorator can be used to decorate the wrapper() function.

Example Decorators

Decorators can be used to measure the performance of a function

Decorators are used to create specialized functions

Special functions like ‘once’ that would only execute one time even if they’re called more than once, can be created using decorators.

Interested in helping build Docs? Read the Contribution Guide or share your feedback form.

Learn Python on Codecademy