Decorator factories and stacking decorators

Sometimes the decorator itself needs an argument, like @repeat(3). The previous decorators just took a function. To accept an argument, you need one more layer.

A decorator factory is a function that returns a decorator. Three nested defs:

def repeat(times):
	def decorator(func):
		def wrapper(*a, **kw):
			return [func(*a, **kw) for _ in range(times)]
		return wrapper
	return decorator

Python

Now @repeat(3) works because repeat(3) returns a decorator:

def repeat(times):
	def decorator(func):
		def wrapper(*a, **kw):
			out = []
			for _ in range(times):
				out.append(func(*a, **kw))
			return out
		return wrapper
	return decorator
 
@repeat(3)
def hi():
	return "hi"
 
print(hi())

Python

Output

What will be the output?

def repeat(n):
	def deco(func):
		def wrapper(*a, **kw):
			return [func(*a, **kw) for _ in range(n)]
		return wrapper
	return deco
 
@repeat(2)
def go():
	return 5
 
print(go())

Python

You can stack multiple decorators by listing them above the function. They apply bottom up: the one nearest the def runs first.

Add one and double, applied in different orders:

def add_one(func):
	def w(*a, **kw):
		return func(*a, **kw) + 1
	return w
 
def double(func):
	def w(*a, **kw):
		return func(*a, **kw) * 2
	return w
 
@double
@add_one
def value():
	return 5
 
print(value())

Python

Output

Reading bottom up: value returns 5, add_one makes it 6, then double makes it 12. Reverse the order and you get (5 * 2) + 1 = 11.

What will be the output?

def add_one(func):
	def w(*a, **kw):
		return func(*a, **kw) + 1
	return w
 
def double(func):
	def w(*a, **kw):
		return func(*a, **kw) * 2
	return w
 
@add_one
@double
def value():
	return 5
 
print(value())

Python

What will be the output?

def repeat(n):
	def deco(func):
		def w(*a, **kw):
			return [func(*a, **kw) for _ in range(n)]
		return w
	return deco
 
@repeat(4)
def x():
	return 0
 
print(len(x()))

Python

What will be the output?

def add(func):
	def w(*a, **kw):
		return func(*a, **kw) + 10
	return w
 
@add
@add
def n():
	return 1
 
print(n())

Python

Decorator factories are useful when behaviour depends on a value (count, role name, retry limit). Stacking lets you compose small decorators into larger pipelines.

What will be the output?

def repeat(n):
	def deco(func):
		def w(*a, **kw):
			return [func(*a, **kw) for _ in range(n)]
		return w
	return deco
 
@repeat(3)
def f():
	return "x"
 
print(f())

Python