topliceanu/sum

    _____                       _
   / ____|                     (_)
  | (___   _   _  _ __ ___      _  ___
   \___ \ | | | || '_ ` _ \    | |/ __|
   ____) || |_| || | | | | | _ | |\__ \
  |_____/  \__,_||_| |_| |_|(_)| ||___/
                              _/ |
                             |__/

Sum.js

Indicator
continuous integration
dependency management
change log	CHANGELOG Releases

A simple function for summarizing text e.g. for automatically determining the sentences that are most relevant to the context of the corpus. This library depends on the underscore, underscore.string and porter-stemmer.

Install in node.js

sudo npm install -g sum

Install in browser

<script src="/lib/underscore.js"></script>
<script src="/lib/underscore.string.js"></script>
<script src="/lib/porter-stemmer.js"></script>
<script src="/sum.js"></script>

Quick Start

var sum = require( 'sum' );
var bigString = "....";
var abstract = sum({ 'corpus': bigString });
// `abstract` is an object w/ format `{"summary":String, "sentences":Array<String>}`
// where summary is the concatenation of the array of sentences.

Further Options

var sum = require( 'sum' );
var anotherBigString = "...";
var abstract = sum({
    /**
     * `corpus`: String - is the string you want to summarize
     */
    'corpus': anotherBigString,

    /**
     * `nSentences`: Number - controls the number of sentences from the original text included in the abstact
     */
    'nSentences': 3,

    /**
     * `nWords`: Number - controls the length in words of the nGram output. Output might be larger as some words are ignored in the algorithm but present in the abstract, for ex. prepositions. When `nWords` is set, `nSentences` is ignored
     */
    'nWords': 5,

    /**
     * `exclude`: Array[String] - sum.js allows you to exclude from the final abstract, sentences or nGrams that contain any of the words in the `exclude` param
     */
    'exclude': ['polar', 'bear'],

    /**
     * `emphasise`: Array[String] - forces sum.js to include in the summary the sentences or nGrams that contain any the words specified by `emphasise` param.
     */
    'emphasise': ['magic']
});

//`abstract` is an object with format {'sentences':Array<String>, 'summary':String} where summary is just the concatenation of the sentences, for convenience.
console.log("The short version of corpus is ", abstract.summary);

Running tests

Run /tests/browser/specrunner.html in your favourite browser.

To run node tests, run npm run test.

Goals

This library is intended to be fully embeddable. It's purpose is to be used primarly on the client-side. It should be self-contained so no API calls to external services. It should be as light as possible, both in terms of code size and dependencies and above all it must be fast. Because of these constraints, the algorithm used is purely statistical, using TF IDF to calculate abstracts. Other methods of text summarization proposed by researchers in NLP and ML produce better results but are not (to my best of knowledge) practical in the browser context as many of them require intense computation to produce their output.

TODO

1. add tests to verify the correctness of the actual output
2. currenty the output does not preserve the ending chars of the original sentences
3. make the lib more plugable, e.g allow plugin of custom algorithms, string cleaning rutines, etc.
4. better control for the length of the summary, by words, by letters
5. add more algorithms to calculate abstracts
6. Make the library's inner utility functions available as some of them might be usefull

Licence

(The MIT License)

Copyright (c) Alex Topliceanu [email protected]

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.