Standard library — math, random, os, sys, datetime, time, collections, itertools, functools

1.3.2 · Coding › Python Intermediate

Drawers ka map (woh 20% jo 80% deta hai)

Module	Yeh KISLIYE hai (ek word)	Killer tools
`math`	Pure-number math	`sqrt, floor, ceil, gcd, factorial, pi, comb, log`
`random`	Pseudo-randomness	`random, randint, choice, shuffle, sample, seed`
`os`	OS / files se baat karna	`getcwd, listdir, path.join, environ, makedirs`
`sys`	Interpreter se baat karna	`argv, exit, path, maxsize`
`datetime`	Calendar dates & times	`datetime.now, timedelta, strftime, strptime`
`time`	Clock / sleep / timing	`time(), sleep(), perf_counter()`
`collections`	Better containers	`Counter, defaultdict, deque, namedtuple`
`itertools`	Lazy looping algebra	`count, cycle, product, permutations, combinations, chain, groupby`
`functools`	Functions par kaam karne wale tools	`reduce, lru_cache, partial, wraps`

math — exact numeric helpers

Worked example Do points ke beech distance

import math
d = math.hypot(3, 4)      # 5.0

hypot kyun? Yeh $3^{2} + 4^{2}$ compute karta hai lekin internally scaling karke overflow avoid karta hai — math.sqrt(3**2+4**2) yahan kaam karta hai lekin bade numbers ke liye break ho jaata hai.

random — controllable chance

Worked example

import random
random.seed(42)
random.randint(1, 6)       # dice roll, dono ends inclusive
random.choice(['a','b','c'])
random.sample(range(10), 3)  # 3 unique items (no repeats)
random.shuffle(my_list)      # in-place

randint(1,6) mein 6 kyun include hai? Kyunki yeh human "1 se 6" sochne ke liye design kiya gaya hai — range ke unlike, upper end inclusive hai.

os & sys — environment vs interpreter

Worked example Safe path banao

import os
os.path.join('data', 'raw', 'file.csv')   # 'data/raw/file.csv' on Linux, '\\' on Windows

Sirf 'data/' + 'file.csv' kyun nahi? Kyunki separator OS ke hisaab se alag hota hai. os.path.join portable hai — yahi toh poora point hai.

Worked example Command-line arguments

import sys
# run: python script.py hello 7
sys.argv          # ['script.py', 'hello', '7']
sys.exit(0)       # 0 = success, nonzero = error

argv[0] script ka naam kyun hai? C se inherited convention hai — program "apna naam jaanta hai."

datetime & time — do alag clocks

Worked example Date arithmetic

from datetime import datetime, timedelta
now = datetime.now()
tomorrow = now + timedelta(days=1)
now.strftime('%Y-%m-%d')               # format → string
datetime.strptime('2024-01-31', '%Y-%m-%d')  # parse string → object

timedelta kyun? Date mein 1 add karna ambiguous hai (1 second? day?). timedelta unit ko explicit banata hai. Memory trick: format = string mein, parse = string se (strf vs strp).

Worked example Code ko time karo

import time
t0 = time.perf_counter()
work()
print(time.perf_counter() - t0, "seconds")

perf_counter kyun, time.time() nahi? perf_counter highest-resolution monotonic clock hai — agar system clock adjust ho jaaye toh yeh kabhi backward nahi jaata.

collections — superpowers wale containers

Worked example Words count karna

from collections import Counter, defaultdict, deque
Counter("banana")              # Counter({'a':3,'n':2,'b':1})
Counter("banana").most_common(2)  # [('a',3),('n',2)]
 
d = defaultdict(list)
d['fruits'].append('apple')    # key auto-created as []
 
q = deque([1,2,3])
q.appendleft(0); q.pop()       # dono sides par fast

Queue ke liye list ki jagah deque kyun? List ke front se remove karna $O (n)$ hai (sab kuch shift karna padta hai). deque.popleft() $O (1)$ hai.

itertools — lazy looping algebra

Worked example

from itertools import product, permutations, combinations, chain, groupby, count
list(product([0,1], repeat=2))      # [(0,0),(0,1),(1,0),(1,1)]
list(permutations('ABC', 2))        # 6 ordered pairs
list(combinations('ABC', 2))        # 3 unordered: AB AC BC
list(chain([1,2],[3,4]))            # [1,2,3,4]  (flatten)

functools — functions PAR operate karne wale tools

Worked example

lru_cache se Memoization

from functools import lru_cache, reduce
 
@lru_cache(maxsize=None)
def fib(n):
    return n if n < 2 else fib(n-1) + fib(n-2)
 
reduce(lambda a,b: a*b, [1,2,3,4])   # 24  (factorial-style product)

lru_cache kyun? Naive fib same subproblems exponentially recompute karta hai ( $O (2^{n})$ ). Cache har result ek baar store karta hai → $O (n)$ . LRU = least-recently-used items cache full hone par evict ho jaate hain.

Figure — Standard library — math, random, os, sys, datetime, time, collections, itertools, functools

Common mistakes (Steel-man + fix)

Active-recall flashcards

Counter, defaultdict, deque ke liye kaun sa module?

collections

math.comb(n,k) kya return karta hai, aur kis type mein?

(k n) = \frac{n !}{k ! ( n - k )!}

, ek exact integer ke roop mein

Kya random.randint(1,6) mein 6 inclusive hai?

Haan — dono endpoints inclusive hain

strftime aur strptime mein kya fark hai?

strf = format object→string; strp = parse string→object

String + ki jagah os.path.join kyun use karo?

Yeh correct OS separator use karta hai → portable code

Timing ke liye time.time() ki jagah time.perf_counter() kyun?

Highest-resolution monotonic clock; kabhi backward nahi jaata

permutations(n, r) ka size?

n! / (n - r)!

combinations(n, r) ka size?

(r n) = n! / (r! (n - r)!)

functools.reduce(f,[a,b,c]) kya compute karta hai?

f (f (a, b), c)

— left-fold accumulation

Recursive fib ke liye lru_cache kya karta hai?

Results memoize karta hai →

O (2^{n})

O (n)

mein badal deta hai

defaultdict ko list pass karo, list() kyun nahi?

defaultdict khud factory call karta hai jab koi key missing hoti hai

sys.argv[0] kya hai?

Script ka apna naam

deque.popleft() list.pop(0) se better kyun hai?

O (1)

O (n)

(koi shifting nahi)

Recall Feynman: ek 12-saal ke bacche ko explain karo

Socho ek bada toolbox hai jo tumhari cycle ke saath free mein aata hai. Ek drawer (math) mein number gadgets hain, ek (random) ek dice cup hai, ek (os) tumhare computer se files ke baare mein baat karta hai, ek (collections) mein magic boxes hain jo khud cheezein count karte hain, aur ek (itertools) ek machine hai jo tumhare stickers ke har possible combination banaati hai bina tumhare haath lagaye. Tumhe yeh tools banane nahi hain — woh box mein pehle se hain. Bas tumhe yeh jaanna hai ki kaun sa drawer kholna hai.

Connections

Python Iterators and Generators — itertools lazy kyun hai
Big-O Notation — deque/lru_cache complexity kyun change karte hain
Recursion and Memoization — lru_cache deep dive
Combinatorics — math.comb, permutations, combinations formulas
File Handling in Python — os aur os.path ke saath pair hota hai
Decorators — functools.wraps, lru_cache, partial