summarize.py

Simple multi-language Python and NLTK-based implementation of text summarization.

Installation

$ pip install pysummarize

Setup

Before using, make sure you have stopwords and punkt NLTK packages downloaded:

import nltk
nltk.download(['stopwords', 'punkt'])

Quick start

from summarize import summarize
summarize("Alice and Bob are friends. Alice is fun and cuddly."
          " Bob is cute and quirky. Together they go on wonderful"
          " adventures in the land of tomorrow. Alice's cuddlines"
          " and Bob's cuteness allow them to reach their goals."
          " But before they get to them, they have to go past their"
          " mortal enemy - Mr. Boredom. He is ugly and mean. They"
          " will surely defeat him. He is no match for their abilities.")

Usage

summarize(text[, sentence_count=5, language='english'])

text

text to be summarized
sentence_count

number of sentences needed
language

language that the text is in (lowercase)

Supported languages

In theory, any language with full support in NLTK (stemming, sentence tokenization and stopwords) should work.

Working

English
Spanish

Supported by NLTK, but untested results

Danish
Dutch
Finnish
French
German
Italian
Norwegian
Portuguese
Swedish

Online demo

http://summarize.plansfortheday.org/ (source)

Credits

Original description of the algorithm: https://web.archive.org/web/20170825024342/http://engineering.flipboard.com/2014/10/summarization/