Decorators can be thought of as functions which modify the functionality of another function. They help to make your code shorter and mote "Pythonic".
To properly explain decorators we will slowly build up from functions. Make sure to restart the Python and the Notebooks for this lecture to look the same on your own computer. So lets break down the steps:
In [1]:
def func():
return 1
In [2]:
func()
Out[2]:
In [6]:
s = 'Global Variable'
def func():
print locals()
Remember that Python functions create a new scope, meaning the function has its own namespace to find variable names when they are mentioned within the function. We can check for local variables and global variables with the local() and globals() functions. For example:
In [7]:
print globals()
Here we get back a dictionary of all the global variables, many of them are predefined in Python. So let's go ahead and look at the keys:
In [8]:
print globals().keys()
Note how s is there, the Global Variable we defined as a string:
In [10]:
globals()['s']
Out[10]:
Now lets run our function to check for any local variables in the func() (there shouldn't be any)
In [11]:
func()
Great! Now lets continue with building out the logic of what a decorator is. Remember that in Python everything is an object. That means functions are objects which can be assigned labels and passed into other functions. Lets start with some simple examples:
In [12]:
def hello(name='Jose'):
return 'Hello '+name
In [13]:
hello()
Out[13]:
Assign a label to the function. Note that e are not using parentheses here because we are not calling the function hello, instead we are just putting it into the greet variable.
In [14]:
greet = hello
In [15]:
greet
Out[15]:
In [16]:
greet()
Out[16]:
This assignment is not attached to the original function:
In [17]:
del hello
In [18]:
hello()
In [19]:
greet()
Out[19]:
In [26]:
def hello(name='Jose'):
print 'The hello() function has been executed'
def greet():
return '\t This is inside the greet() function'
def welcome():
return "\t This is inside the welcome() function"
print greet()
print welcome()
print "Now we are back inside the hello() function"
In [27]:
hello()
In [29]:
welcome()
In [34]:
def hello(name='Jose'):
def greet():
return '\t This is inside the greet() function'
def welcome():
return "\t This is inside the welcome() function"
if name == 'Jose':
return greet
else:
return welcome
In [39]:
x = hello()
Now lets see what function is returned if we set x = hello(), note how the closed parenthesis means that name ahs been defined as Jose.
In [40]:
x
Out[40]:
Great! Now we can see how x is pointing to the greet function inside of the hello function.
In [42]:
print x()
Lets take a quick look at the code again.
In the if/else clause we are returning greet and welcome, not greet() and welcome().
This is because when you put a pair of parentheses after it, the function gets executed; whereas if you don’t put parenthesis after it, then it can be passed around and can be assigned to other variables without executing it.
When we write x = hello(), hello() gets executed and because the name is Jose by default, the function greet is returned. If we change the statement to x = hello(name = "Sam") then the welcome function will be returned. We can also do print hello()() which outputs now you are in the greet() function.
In [43]:
def hello():
return 'Hi Jose!'
def other(func):
print 'Other code would go here'
print func()
In [45]:
other(hello)
Great! Note how we can pass the functions as objects and then use them within other functions. Now we can get started with writing our first decorator:
In [46]:
def new_decorator(func):
def wrap_func():
print "Code would be here, before executing the func"
func()
print "Code here will execute after the func()"
return wrap_func
def func_needs_decorator():
print "This function is in need of a Decorator"
In [47]:
func_needs_decorator()
In [50]:
# Reassign func_needs_decorator
func_needs_decorator = new_decorator(func_needs_decorator)
In [51]:
func_needs_decorator()
So what just happened here? A decorator simple wrapped the function and modified its behaviour. Now lets understand how we can rewrite this code using the @ symbol, which is what Python uses for Decorators:
In [52]:
@new_decorator
def func_needs_decorator():
print "This function is in need of a Decorator"
In [53]:
func_needs_decorator()
Great! You've now built a Decorator manually and then saw how we can use the @ symbol in Python to automate this and clean our code. You'll run into Decorators a lot if you begin using Python for Web Development, such as Flask or Django!