Keyword_Extract.py

# -*- coding: utf-8 -*-
"""
Created on Mon Mar  4 13:25:43 2019

@author: Shatadru Majumdar
"""

import numpy as np
from sklearn.feature_extraction.text import TfidfVectorizer


def keyword_extraction(tfidf_vector,document_list):
    keywords=[]
    for i in range(0,len(document_list)):
        response2=tfidf_vector.transform([document_list[i]])
        feature_array = np.array(tfidf_vector.get_feature_names())
        tfidf_sorting = np.argsort(response2.toarray()).flatten()[::-1]
        n=12
        top_n = feature_array[tfidf_sorting][:n]
        keywords.append(top_n)
    return keywords

fileName = "report4.txt"
d=open(fileName,"r",encoding='utf8')
document=d.read()
document_list=document.split("\n\n")
tfidf_vector=TfidfVectorizer(stop_words='english',ngram_range=(1,3))
response=tfidf_vector.fit_transform(document_list)
keywords=keyword_extraction(tfidf_vector,document_list)
keywords=np.asarray(keywords)
print(keywords)