Python Decorators

The decorator pattern in Python is a means to add functionality to a function without actually changing the code of the decorated function.

The pattern

The decorator pattern consists of

  • Defining a decorator function
  • calling the decorator with the decorated function as parameter
  • and assigning the return value of this call to the decorated function

Decorator function

In its basic form, a decorator is a function that receives a function (decorated function) as parameter, defines an internal function which wraps this decorated function and then returns the wrapper function

Here is a basic example of the pattern :

from functools import wraps

#this is the function to be decorated
def decorated(a,b):
  print ("I am a decorated function")
  print ("There is nothing special about me")
  print (f"my arguments are a= {a} and b={b}")

#START Decorator Pattern

#1- define a decorator function

def my_decorator(funct):
  @wraps(funct)   # see below for more on @wraps
  def wrapper(*args,**kwargs):
    # code to be executed before executing funct
    print(f"before {funct.__name__}") 
    funct(args,kwargs)
    # code to be executed after executing funct
    print (f"after {funct.__name__}")
  return wrapper

#2- call the decorator with the decorated function as parameter
#3- and assign the return of the call to the decorated function

decorated = my_decorator(decorated)

#END Decorator Pattern

#calling the decorated function
decorated ("one", b = "two")

This produces:

before decorated 
I am a decorated function 
There is nothing special about me 
my arguments are a= one and b=two 
after decorated  

@decorator syntax

Steps 2 and 3 of the pattern can be shorthanded by using the @decorator syntax. Right before the definition of the function to be decorated, insert a line containing @ followed by the name of the decorator function.

#this line decorates decorated_1 with my_decorator
@my_decorator
def decorated_2(a,b):
    print ("I am also a decorated function")
    print ("but I use the @decorator syntax")
    print (f"arguments are {a} and {b}")

decorated_2("hello","world")

This will produce:

before decorated_2 
I am also a decorated function 
but I use the @decorator syntax 
arguments are hello and world                                   
after decorated_2  

Good uses for decorators are:

  • Logging
  • Timing a function
  • IO
  • Network communication

Using @wraps

The wrapper function of the decorator is usually itself decorated with the @wraps(funct) decorator. This prevents the decorated function from having its attribute __name__ changed to the name of the wrapper function