Cosine Similarity with numpy

cosine_similarity.py

import numpy as np
import math

def get_cosine_sim(a, b):
    a = np.array(a)
    b = np.array(b)
    
    aq = np.square(np.abs(a))
    bq =np.square(np.abs(b))

    asum = np.sum(aq)
    bsum = np.sum(bq)

    asq = math.sqrt(np.sum(aq))
    bsq = math.sqrt(np.sum(bq))
    dot = a.dot(b)
    
    similarity = (dot / (asq * bsq))
    
    return similarity

cosine_similarity_no_numpy.py

import math

def get_cosine_sim(a, b):
  assert len(a) == len(b)
	
	dot = sum([v * b[i] for i,v in enumerate(a)])
	asq = math.sqrt(sum([v**2.0 for v in a]))
	bsq = math.sqrt(sum([v**2.0 for v in b]))
	
	similarity = (dot / (asq * bsq))
  
  return similarity

euclidean_distance.py

from typing import Sequence
import math

def euclidean_distance(vector1: Sequence[float], vector2: Sequence[float]):
    assert len(vector1) == len(vector2)
    squares = []
    
    for index, el in enumerate(vector1):
        subtraction = el - vector2[index]
        square = subtraction ** 2
        squares.append(square)
  
    square_sum =sum(squares)
    
    return math.sqrt(square_sum)

l2_norm.py

from typing import Sequence
import math

def l2_norm(vector: Sequence[float]) -> float:
    squares = [el * el for el in vector]
    square_sum = sum(squares)
    square_root = math.sqrt(square_sum) if square_sum > 0.0 else 1.0
    return square_root