illinois-twitter-client
A thin client for accessing Twitter's (filtered) firehose. Uses the Hosebird Client twitter library.
Requirements
To use the client, you need to get authentication tokens from Twitter, by registering your app here.
Once you have your tokens, store them in config/auth.properties using the following keys:
consumerKey = 1234key
consumerSecret = 1234secret
token = 1234token
secret = 1234secretRunning
A typical scenario involves instantiating the TwitterClient with some filters (locations, search terms, languages)
and creating a MessageHandler thread that can access the queue of messages coming from the stream.
public static void main(String[] args) {
// Set up location filters
List<Location> locations = Arrays.asList(Locations.URBANA_CHAMPAIGN, Locations.EDINBURGH);
// Set up search-term filters
//List<String> terms = Arrays.asList("machine learning", "natural language processing");
// Set up language filters
//List<String> languages = Arrays.asList("en", "es");
TwitterClient client = new TwitterClient(null, locations, null);
// A separate thread for handling the queue of tweets
MessageHandler simpleMessageHandler = new SimpleMessageHandler(client.getMsgQueue(), client.getClient());
Thread thread = new Thread(simpleMessageHandler);
thread.start();
}
public class SimpleMessageHandler extends MessageHandler {
public SimpleMessageHandler(BlockingQueue<String> msgQueue, BasicClient client) {
super(msgQueue, client);
}
public void run() {
while (!client.isDone()) {
try {
String msg = msgQueue.take();
Utils.printInfo(new JSONObject(msg));
} catch (InterruptedException e) {
e.printStackTrace();
}
}
client.stop();
}
}ClassifierMessageHandler is another example of MessageHandler where an LBJava Classifier is used called on
every message in the stream.
Note: you need to train the classifier by running mvn lbjava:clean lbjava:compile. The folder data contains sample
sentiment annotated data from here.
public void run() {
while (!client.isDone()) {
try {
String msg = msgQueue.take();
Utils.printInfo(new JSONObject(msg));
String text = Utils.getCleanText(new JSONObject(msg));
// Need to convert the text to a Tweet datastructure for the LBJava classifer to work
String decision = classifier.discreteValue(new Tweet(text));
System.out.println("\t***Sentiment classification: " + decision);
} catch (InterruptedException e) {
e.printStackTrace();
}
}
client.stop();
}