@fusionary/keyword-extractor

Keyword Extractor

A simple NPM package for extracting keywords from a string (or array of strings) by removing stopwords.

Note: this is a fork of the Michael De Lorenzo's Keyword Extractor to include Chinese and allow for an array of strings as the first argument.

Installation

npm install keyword-extractor

Running the demo

This repo's demo html file is at demo/index.html. If you want to modify JavaScript for the demo (at demo/sample.js), make sure you've installed the development dependencies:

npm install

Then run browserify:

cd keyword-extractor/demo
browserify sample.js -o bundle.js

see browserify documentation for detailed usage

Running tests

To run the test suite, first install the development dependencies by running the following command within the package's directory.

npm install

To execute the package's tests, run:

make test

Usage of the Module

//  include the Keyword Extractor
var keyword_extractor = require("keyword-extractor");

//  Opening sentence to NY Times Article at
//  http://www.nytimes.com/2013/09/10/world/middleeast/surprise-russian-proposal-catches-obama-between-putin-and-house-republicans.html
var sentence = "President Obama woke up Monday facing a Congressional defeat that many in both parties believed could hobble his presidency."

//  Extract the keywords
var extraction_result = keyword_extractor.extract(sentence, {
    language:"english",
    remove_digits: true,
    return_changed_case:true,
    remove_duplicates: false

});

/*
  extraction result is:

  [
        "president",
        "obama",
        "woke",
        "monday",
        "facing",
        "congressional",
        "defeat",
        "parties",
        "believed",
        "hobble",
        "presidency"
    ]
*/

Options Parameters

The second argument of the extract method is an Object of configuration/processing settings for the extraction.

Parameter Name	Description	Permitted Values
language	The stopwords list to use.	english, chinese, danish, spanish, galician, german, french, italian, dutch, romanian, russian, polish, portuguese, swedish,
remove_digits	Removes all digits from the results if set to true	true or false
return_changed_case	The case of the extracted keywords. Setting the value to true will return the results all lower-cased, if false the results will be in the original case.	true or false
return_chained_words	Instead of returning each word separately, join the words that were originally together. Setting the value to true will join the words, if false the results will be splitted on each array element.	true or false
remove_duplicates	Removes the duplicate keywords	true , false (defaults to false )
return_max_ngrams	Returns keywords that are ngrams with size 0-integer	integer , false (defaults to false )

Credits

The initial stopwords lists are taken from the following sources:

• English [http://jmlr.org/papers/volume5/lewis04a/a11-smart-stop-list/english.stop]
• Spanish [https://stop-words.googlecode.com/svn/trunk/stop-words/stop-words/stop-words-spanish.txt]

Credits for other lists can be found in their respective files within the lib/stopwords/ directory