twitter-statuses-subject-recognizer

twitter-statuses-subject-recognizer lets you define subjects to track using Twitter Streaming APIs and recognize incoming statuses' subjects.

What problem does this module solve?

If you decide working with Twitter Streaming APIs and want to track statuses that relate to numerous topics (subjects) you will face the problem that Twitter pushes all statuses in one stream so that your application needs a way to identify the subject of each status. twitter-statuses-subject-recognizer solves that problem. Of course it may be used not only with Streaming APIs (e.g., assume you have large amount of tweets stored in database, and you want to filter those tweets that relate to one or various particular topics).

Install

$ npm install twitter-statuses-subject-recognizer

How to use

Lets say you want to track Twitter statuses related to GitHub, pastries, and Vladimir Putin. First you need to define your subjects and instantiate parser object:

var TSSR = require('twitter-statuses-subject-recognizer').TwitterStatusesSubjectRecognizer;

// Define subjects:
var subjects = [
  {
    name: "GitHub",
    keywords: ["github"]
  },
  {	
	name: "Pastries",
	keywords: ["banitsa", "eclair", "blachindla", "chouquette"]
  },
  {
	name: "Putin", 
	keywords: ["путин", "putin", "普京"]
  }
];

// Instantiate parser object:
var tssr = new TSSR(subjects);

In that case your request to Twitters API may look like:

https://stream.twitter.com/1.1/statuses/filter.json?track=github,banitsa,eclair,blachindla,chouquette,путин,putin,普京

where track parameter holds all keywords concatinated with commas. You don't have to hardcode track parameter, once parser object were instantiated it has trackString property which you can use as track parameter.

console.log(tssr.trackString);
-> 'github,banitsa,eclair,blachindla,chouquette,путин,putin,普京'

Once you have connected to streaming endpoint you can use parse method to recognize subjects of incoming tweets:

var augmentedTweet = tssr.parse(tweet);

That's it! tssr.parse method creates new object that has all properties of the original tweet and subjectsCollection property that holds subjects it's related to. To get names of related subjects use: augmentedTweet.subjectsCollection.names.

Example

This example utilizes twit – Twitter API Client for node.

var Twit = require('twit')
  , TSSR = require('twitter-statuses-subject-recognizer').TwitterStatusesSubjectRecognizer

// Instatiate Twitter client using your application credentials:
var T = new Twit({
  'consumer_key': "your consumer key",
  'consumer_secret': "your consumer secret",
  'access_token': "your access_token",
  'access_token_secret': "your access token secret"
});

// Define subjects to track
var subjects = [
  {
    name: "Putin", 
    keywords: ["путин", "putin", "普京"]
  }
];

// Instantiate parser object:
var tssr = new TSSR(subjects);

// Connect to 'statuses/filter' endpoint using twit module.
// Use tssr.trackString property as track parameter:
var stream = T.stream('statuses/filter', {track: tssr.trackString});

stream.on('tweet', function (tweet) {
  // Recognize subjects:
  augmentedTweet = tssr.parse(tweet);
  // Log subject names:
  console.log(augmentedTweet.subjectsCollection.names);
});