Compare similarity between 2 sentences using cosine similarity

compare_sentences_cosine.py

from gensim.parsing.preprocessing import STOPWORDS
import re
import numpy as np
from numpy.linalg import norm



def find_similarity(x,y):
#print(STOPWORDS)
    token_rules = r'\s+|[,\.]'

    x_token = re.split(token_rules,x)
    y_token = re.split(token_rules,y)

    print(x_token)
    print(y_token)

    x_set = {word for word in x_token if not word in STOPWORDS}
    y_set = {word for word in y_token if not word in STOPWORDS}

    print(x_set)
    print(y_set)

    rvector = x_set.union(y_set)
    print(rvector)

    l1 = []
    l2 = []

    for word in rvector:
        if word in x_set: l1.append(1)
        else: l1.append(0)
        if word in y_set: l2.append(1)
        else: l2.append(0)

    vec1 = np.array(l1)
    vec2 = np.array(l2)

    cosine = np.dot(vec1,vec2)/(norm(vec1)*norm(vec2))

    return (cosine)

print(find_similarity("I am a boy","I am a girl"))