sandeepk

In this post, we will talk about the standard python library decorators and one or more things about the decorators . If you haven't read the previous blog post about decorator, go check out that here. I will be waiting...

We will talk about functools.lru_cache Decorator from Python standard Library, where lru means Least Recently Used.

lru_cache as the name suggested, it saves the previous result of the function expression based on argument and uses that result if the same argument passed. To save expensive calculations.

functools.lru_cache(maxsize=128, typed=False)

maxsize means that numbers of cache result which can be cached, once the cache is full the older result is discarded. One should use maxsize value as a power of 2 for optimal performance.

type true means argument will be treated differently as int and float values as 1 and 1.0 are treated the same, but if type value is set to true it will be treated differently.

>> 1 == 1.0
>> True

lru_cache use dict to the save the argument as position and keyword-based so all the argument passed to the decorator should be hash-able.

Some point as notes to remember about the decorators

• Decorators are executed when the module is loaded by Python and decorated function only executed if explicitly invoked.
• Decorators have the power to return the entirely a different function.
• We can also have a parameterized decorator as we have seen in the lru_cache decorator.
• Stocked Decorators means when more then one decorator is applied to a function, then the order of execution, is from the decorator nearest to the function definition to outside. Let seen an example

@d2
@d1
def func:
    print('f')

func = d2(d1(func))

so that wrap from my side on the topic Decorators.

#python

Decorators attach additional responsibility to the object dynamically. A decorator takes other function as an argument which it processes and returns that function or any other callable object.

So how the decorator looks like

@clean_strings
def get_full_name(first_name, middle_name, last_name):
    return first_name + middle_name + last_name

In above code snippet we have a decorator clean_strings, this can also be written in this way.

def get_full_name(first_name, middle_name, last_name):
    return first_name + middle_name + last_name

get_full_name = clean_strings(get_full_name)

There are one two things we will talk about decorator after understanding.

• Variable Scope
• Closure

Variable Scope

In every language, the variable has a scope where they are accessible and where not. So here we will talk about the local scope and global scope lets jump right into the code to see

def show(a):
    print(a)
    print(b)
>>> show(10) 
10
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<stdin>", line 3, in show
NameError: name 'b' is not defined 

We got the error as b is not defined in the scope

b = 101
def show(a):
    print(a)
    print(b)
>>> show(10) 
10
101

here its work fine as b is defined in the global scope, which can be accessed from within the function. let see another code snippet.

b = 101
def show(a):
    print(a)
    print(b)
    b =190
>>> show(10) 
10
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<stdin>", line 3, in show
UnboundLocalError: local variable 'b' referenced before assignment

here we face error as code interpret b as local variable of the function which being accessed before declaring as it is defined in the scope of the function. To treat b as global variable despite the assignment in function we can use global declaration.

b = 101
def show(a):
    global b
    print(a)
    print(b)
    global b =190
>>> show(10) 
10
101
>>> b
190

Closure

Closure are the function which have access to the non global variables referenced in the body of function.

Figure from Fluent Python Book, chapter 7

In Python3 nonlocal was introduced which allows assigning the variable inside the scope.

Consider an avg function to compute the mean of an ever-increasing series of values; for example, the average closing price of a commodity over its entire history. Every day a new price is added, and the average is computed taking into account all prices so far.

def avg_series():
    count = 0
    total = 0
    def averager():
        nonlocal count, total
        count += 1
        total += new_value
        return total / count

    return averager

The need of using nonlocal here is that if we don't, Python assumes that count and total are the local variable of averager method, which will break our logic.

Code example is taken from Fluent Python Book.

Now lets build a decorator that can logs the runtime for the function.

import time
def log_time(func):
    def clocked(*args):
        start_time = time.time()
        result = func(*args)
        elapsed_time = time.time() - start_time
        print("Elapsed time: {}".format(elapsed_time))
        return result
    return clocked    

There are also built-in decorators in Python Standard Library which we will discuss in the next blog post, so stay tuned till then cheers.

#python

Function are the code block which contain a logic to process on certain set of given input and return an output. Functions in Python are the First Class Object which basically means function as entity can be

• Create at Run-time.
• Passed as argument to function.
• Return as result from the function.
• Assigned to the variables.

Some functions are also called as Higher Order Function which means that a function which take other function as an argument or return a function as result. Example of higher order function in Python are Map, Filter, Sorted ...

Let see in Python function are classes or not and try to prove above all points to show in Python Function are First Class Object.

Creating function at run-time in console.

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

add(2, 4)
6
type(add)
<class 'function'>

Assigned to the variable

sum = add
sum
<function add at 0x7f2199555b70>
// notice above sum variable pointing to the add function.
sum(3, 4)
7

Passing function as argument.

list(map(add, range(5), range(5))) // here we pass *add* function as argument to the *map* function.
[0, 2, 4, 6, 8]

Returning function as result.

def factorial(x):
    if x < 1:
        return 1
    else:
        return x * factorial(x-1) // here we are returning a function

Above code snippets clearly show that function in Python are First Class Object.

#python