Skip to content

rinzeb/calais-entity-extractor

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

calais-entity-extractor

An npm package that provides an easy way to extract entities from blocks of text using Open Calais. A valid Calais key is required. You can get a free one at the Open Calais site. This module was inspired by node-calais, but that project doesn't (as of 10/6/2015) support the Calais API changes.

We perform named entity recognition and output clean entity markup tags and socialTags in JSON.

Installation

npm install calais-entity-extractor

Usage

var Calais = require('calais-entity-extractor').Calais;
var calais = new Calais('ACCESS TOKEN');  //See valid options below

// You can set options after the constructor using .set(option, value). The example below sets
// the text that we want to analyze.
calais.set('content', 'The awesome text to analyze. News stories work great.');


calais.extractFromText(function(result, err) {     //perform the request
    if (err) {
        console.log('Uh oh, we got an error! : ' + err);
        return;
    }

    //Take a look at the results!
    var util = require('util');

    //The results have two fields: 'entities' and 'tags'

    //'entities' contains a list of the detected entities, and gives basic info & confidence
    console.log('Entities: ' + util.inspect(result.entities, false, null));

    //'tags' are a list of string tags (the "socialTags" from Calais).
    console.log('\nTags: ' + util.inspect(result.tags, false, null));
});

Example output of the above example on a news story:

Entities: [ { type: 'Company',
        name: 'Toyota',
        fullName: 'Toyota Motor Corp',
        confidence: '0.903' },
    { type: 'Person',
        name: 'Matthias Mueller',
        fullName: 'Matthias Mueller',
        confidence: '0.999' },
    { type: 'Company',
            name: 'Volkswagen',
        fullName: 'Volkswagen AG',
        confidence: '0.985' },
    { type: 'Person',
        name: 'Max Warburton',
        fullName: 'Max Warburton',
        confidence: '0.997' },
    { type: 'Person',
        name: 'Martin Winterkorn',
        fullName: 'Martin Winterkorn',
        confidence: '0.995' },
    { type: 'Company',
        name: 'Sanford C. Bernstein',
        fullName: 'Sanford C Bernstein Fund II Inc',
        confidence: '0.999' 
    } 
]

Tags: [ 'Volkswagen Group',
    'Volkswagen',
    'Martin Winterkorn',
    'Volkswagen emissions violations' 
]

Valid options and their default values are:

apiHost                 : 'api.thomsonreuters.com',
apiPath                 : '/permid/calais',
contentTy pe            : 'text/raw',          // [text/html, text/xml, text/raw, application/pdf]
language                : 'English'            // [English, Spanish, French],
minConfidence           : 0.75                 // Anything that has less than this confidence level is ignored

We also support analyzing text directly from webpages. Set up the calais objects just like in the previous example, and perform a query like this:

calais.extractFromUrl(url, function(result, err) {
    ...
 });

The results are returned in the same way as the extractFromText method.

For working examples, see example.js

Tests

expresso test/calais.test.js

About

Extract entities from text using Open Calais

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • JavaScript 100.0%