Last active
September 17, 2023 13:54
-
-
Save B-R-P/31619a0600c85db338c6abcabb8b19d4 to your computer and use it in GitHub Desktop.
Summarize text using nltk and numpy
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import nltk | |
import numpy as np | |
from nltk.stem import PorterStemmer | |
from nltk.text import TextCollection | |
from nltk.corpus import stopwords as sw | |
from nltk.tokenize import word_tokenize, sent_tokenize | |
stem = PorterStemmer().stem | |
nltk.download('stopwords') | |
stopWords = sw.words('english') | |
normalize = lambda array: np.divide(array, np.max(array)) | |
def RankText(text): | |
"Function to rank the sentences in the text based on their importance" | |
sentenceCollection = TextCollection(tuple(map(# Create a TextCollection object to work with sentences | |
lambda sentence: [ | |
stem(word) | |
for word in word_tokenize(sentence.lower()) | |
if (word.isalpha() and word not in stopWords) | |
], | |
sentences := sent_tokenize(text) | |
))) | |
RankedWords = tuple(zip(# Rank words based on their importance using TF-IDF | |
uniqueWords := set(sentenceCollection.tokens), | |
normalize(tuple(map(sentenceCollection.idf, uniqueWords))) | |
)) | |
return ( | |
tuple(zip(# Rank sentences based on the importance of words they contain | |
sentences, # Current sentence | |
normalize(np.sum([ | |
np.divide([ | |
rank | |
if word in sentence | |
else 0 | |
for word, rank in RankedWords | |
], | |
float(len(sentence)) # Avoid long sentence bias | |
) | |
if len(sentence) != 0 # Avoid division by zero | |
else np.zeros(len(RankedWords)) # Create a zero vector for empty sentences | |
for sentence in sentenceCollection._texts | |
], | |
axis=1 # Sum of each row | |
)) | |
)), | |
dict(RankedWords) | |
) | |
def getAboveAvg(Sranks): | |
"Function to get sentences above the average importance" | |
avg = sum(map(lambda i: i[1], Sranks)) / len(Sranks) | |
return tuple(filter(lambda i: i[1] > avg, Sranks)) | |
Sranks, Wranks = RankText(open('sample.txt').read()) | |
print("".join(map(lambda i:i[0],getAboveAvg(getAboveAvg(Sranks))))) | |
print(Wranks) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment