functools¶

functools is a module that provides tools for working with functions and other callable objects.

partial¶

One of its useful is partial which allows write and maintain the existing functions to achieve new purpose. The following function shows how to rewrite the t-test in a way just calculate the p-value..

from functools import partial
import numpy as np
import scipy.stats
x, y = np.random.normal(0, 1, (100,2)).T
ttest_0=partial(scipy.stats.ttest_ind, equal_var=False)

def ttest_NE(x,y):
   return print("p-value=",ttest_0(x,y)[1])

ttest_NE(x,y)

wraps¶

wraps can be used to write a decorators in Python, by using @wraps, you can preserves that metadata by copying it from the original function to the wrapper function.

from functools import wraps

def do_nothing_decorator(f):
   @wraps(f)
   def wrapper(*args, **kwds):
          print('INSIDE DECORATOR: This is called before function')
          return f(*args, **kwds)
   return wrapper

@do_nothing_decorator
def hello_world():
  """This is a hello world function"""
  print("AAA")

Let run this function

>>> hello_world()
INSIDE DECORATOR: This is called before function
AAA
>>> print(f"Function Name: {hello_world.__name__}")
Function Name: hello_world
>>> print(hello_world.__doc__)
This is a hello world function

Now let run without the wrape

def do_nothing_decorator(f):
   def wrapper(*args, **kwds):
          print('INSIDE DECORATOR: This is called before function')
          return f(*args, **kwds)
   return wrapper

@do_nothing_decorator
def hello_world():
  """This is a hello world function"""
  print("AAA")

>>> hello_world()
INSIDE DECORATOR: This is called before function
AAA
>>> print(f"Function Name: {hello_world.__name__}")
Function Name: wrapper
>>> print(hello_world.__doc__)
None

So, wraps ensures the decorated function retains its original identity.