Python - advanced cont.

Higher Order Functions

• functions can:
  • accept functions as arguments
  • return functions as values

def is_even(x):
    return x % 2 == 0


def take_while(check,it): #accepts function check()
    '''yields items from it until check() returns False for some item'''
    for i in it:
        if check(i):
            yield i
        else:
            break


vals = [0, 2, 4, 6, 7, 8, 10]
list(take_while(is_even,vals)) #[0, 2, 4, 6]

def new_print(prefix):
    def inner(*args, **kwargs):
        '''
        assigns prefix passed to beginning of print,
        takes *args and **kwargs
        '''
        print(prefix,*args,**kwargs)
    return inner #returns function inner()


debug = new_print("debug: ")
#debug is new function that has "debug: " as prefix
warning = new_print("warning: ")

debug("reached here") #debug: reached here

warning("did the thing") #warning: did the thing

def f():
    i=0
    def g():
        print(i)
        return i
    return g


h=f()

h() #prints i, returns i


def f():
    i = 0
    def g():
        i += 1
        return i
    return g


h=f()
h() #Unbound local variable

Lambda functions

• Lamdas are anonymous functions

• Lambdas take zero of more arguments and return a values

• That’s it

• Syntax: lambda [arglist]: expression

  • lambda x: x + 1
  • lambda x,y: x + y

• Don’t abuse them, only really used when you need a one-time function

dogs = [Dog("frank",150), Dog("bob",110),..] #not gonna write a whole list
# assume no __lt__
dogs2 = sorted(dogs) #TypeError

dogs3 = sorted(dogs,key=lambda x: x.weight) #extremely common usage

map, filter, and reduce

• Holy trinity of math

• All are higher order functions
  • work a lot like generator expressions
• map(func, iterable)
  • returns a new iterable that yields func -> x for each x in iterable
  • map(str,lst) and (str(x) for x in lst) are about the same
• filter(func, iterable)
  • returns a new iterable that yields x for each x in iterable if func(x) is truthy
  • filter(is_even,lst) and (x for x in lst if is_even(x)) are similar
• reduce(func, iterable[,initial]) #import functools
  • applies a function of two arguments cumulatively to the items of iterable from left to right so as to reduce the iterable into a single value
    • vals = [0, 2, 4, 6] ```

    functools.reduce(lambda x,y: x + y, vals) #12 ```

Partial Function Application

Partially applies parameters to function, creates new function that is already partly applied, so takes fewer parameters.

import functools

def add(x,y):
    return x + y

add3 = functools.partial(add,3) #partially applies 3 to add

add3(7) #10

Function Decoration

• Wraps a function to modify its behavior

• Super common in Python Standard Library

• Don’t abuse these

import time

def time_this(func):
    def wrapper(*args,**kwargs):
        start = time.time()
        retval = func(*args,**kwargs)
        stop = time.time()
        print("{} took {:0.3f}sec".format(func.__name__,stop - start))
        return retval
    return wrapper

Extras

Ternery

• works just like in c++
• syntax: true if condition else false