def find_cos_sim(url1,url2):
text1 = te.text_extract(url1)
text2 = te.text_extract(url2)
from nltk.corpus import stopwords
cachedStopWords = stopwords.words("english")
text1= ' '.join([word for word in text1.split() if word not in stopwords.words("english")])
text2= ' '.join([word for word in text2.split() if word not in stopwords.words("english")])
from nltk import PorterStemmer
text1 = PorterStemmer().stem(text1)
text2 = PorterStemmer().stem(text2)
text1_vec = text1.split(' ')
text2_vec = text2.split(' ')
for i in range(0,len(text1_vec)):
text1_vec[i] = PorterStemmer().stem(text1_vec[i])
for i in range(0,len(text2_vec)):
text2_vec[i] = PorterStemmer().stem(text2_vec[i])
text1_str = ""
text2_str = ""
for i in text1_vec:
text1_str = text1_str + " " + i
for i in text2_vec:
text2_str = text2_str + " " + i
vector1 = text_to_vector(text1_str)
vector2 = text_to_vector(text2_str)
cosine = get_cosine(vector1, vector2)
return cosine
