Apache Storm on Twitter

In this chapter, we will discuss real-time applications of Apache Storm. We will see how Storm is used in Twitter.

Twitter

Twitter is an online social networking service that provides a platform for sending and receiving user tweets. Registered users can read and post tweets, but unregistered users can only read tweets. Hashtag is used to categorize tweets by keyword by appending # before relevant keywords. Now let’s look at a real-time scenario to find the most used hashtags for each topic.

Spout was created to receive tweets submitted by people as quickly as possible. Twitter provides the "Twitter Streaming API", a web service-based tool for retrieving tweets submitted by people in real time. The Twitter Streaming API can be accessed using any programming language.

twitter4j

is an open source and unofficial Java library that provides a Java-based module to easily access the Twitter Streaming API. twitter4j provides a listener-based framework to access tweets. To access the Twitter Streaming API, we need to log in to the Twitter developer account and obtain the following OAuth authentication details.

Customerkey

• CustomerSecret
• accessToken
• AccessTookenSecret
Storm provides a twitter spout,TwitterSampleSpout. We will use this to retrieve tweets. The email requires OAuth authentication details and at least one keyword. This spout will emit real-time tweets based on keywords. The complete program code is as follows.

Encoding: TwitterSampleSpout.java

import java.util.Map;
import java.util.concurrent.LinkedBlockingQueue;

import twitter4j.FilterQuery;
import twitter4j.StallWarning;
import twitter4j.Status;
import twitter4j.StatusDeletionNotice;
import twitter4j.StatusListener;

import twitter4j.TwitterStream;
import twitter4j.TwitterStreamFactory;
import twitter4j.auth.AccessToken;
import twitter4j.conf.ConfigurationBuilder;

import backtype.storm.Config;
import backtype.storm.spout.SpoutOutputCollector;

import backtype.storm.task.TopologyContext;
import backtype.storm.topology.OutputFieldsDeclarer;
import backtype.storm.topology.base.BaseRichSpout;
import backtype.storm.tuple.Fields;
import backtype.storm.tuple.Values;

import backtype.storm.utils.Utils;

@SuppressWarnings("serial")
public class TwitterSampleSpout extends BaseRichSpout {
   SpoutOutputCollector _collector;
   LinkedBlockingQueue<Status> queue = null;
   TwitterStream _twitterStream;
		
   String consumerKey;
   String consumerSecret;
   String accessToken;
   String accessTokenSecret;
   String[] keyWords;
		
   public TwitterSampleSpout(String consumerKey, String consumerSecret,
      String accessToken, String accessTokenSecret, String[] keyWords) {
         this.consumerKey = consumerKey;
         this.consumerSecret = consumerSecret;
         this.accessToken = accessToken;
         this.accessTokenSecret = accessTokenSecret;
         this.keyWords = keyWords;
   }
		
   public TwitterSampleSpout() {
      // TODO Auto-generated constructor stub
   }
		
   @Override
   public void open(Map conf, TopologyContext context,
      SpoutOutputCollector collector) {
         queue = new LinkedBlockingQueue<Status>(1000);
         _collector = collector;
         StatusListener listener = new StatusListener() {
            @Override
            public void onStatus(Status status) {
               queue.offer(status);
            }
					
            @Override
            public void onDeletionNotice(StatusDeletionNotice sdn) {}
					
            @Override
            public void onTrackLimitationNotice(int i) {}
					
            @Override
            public void onScrubGeo(long l, long l1) {}
					
            @Override
            public void onException(Exception ex) {}
					
            @Override
            public void onStallWarning(StallWarning arg0) {
               // TODO Auto-generated method stub
            }
         };
				
         ConfigurationBuilder cb = new ConfigurationBuilder();
				
         cb.setDebugEnabled(true)
            .setOAuthConsumerKey(consumerKey)
            .setOAuthConsumerSecret(consumerSecret)
            .setOAuthAccessToken(accessToken)
            .setOAuthAccessTokenSecret(accessTokenSecret);
					
         _twitterStream = new TwitterStreamFactory(cb.build()).getInstance();
         _twitterStream.addListener(listener);
				
         if (keyWords.length == 0) {
            _twitterStream.sample();
         }else {
            FilterQuery query = new FilterQuery().track(keyWords);
            _twitterStream.filter(query);
         }
   }
			
   @Override
   public void nextTuple() {
      Status ret = queue.poll();
				
      if (ret == null) {
         Utils.sleep(50);
      } else {
         _collector.emit(new Values(ret));
      }
   }
			
   @Override
   public void close() {
      _twitterStream.shutdown();
   }
			
   @Override
   public Map<String, Object> getComponentConfiguration() {
      Config ret = new Config();
      ret.setMaxTaskParallelism(1);
      return ret;
   }
			
   @Override
   public void ack(Object id) {}
			
   @Override
   public void fail(Object id) {}
			
   @Override
   public void declareOutputFields(OutputFieldsDeclarer declarer) {
      declarer.declare(new Fields("tweet"));
   }
}

Hashtag Reader spout

Tweets sent by the spout will be forwarded to

HashtagReaderBolt

, which will handle the tweet and emit all available hashtags. HashtagReaderBolt uses the getHashTagEntities method provided by twitter4j. getHashTagEntities reads tweets and returns a list of hashtags. The complete program code is as follows -
Encoding: HashtagReaderBolt.java

import java.util.HashMap;
import java.util.Map;

import twitter4j.*;
import twitter4j.conf.*;

import backtype.storm.tuple.Fields;
import backtype.storm.tuple.Values;

import backtype.storm.task.OutputCollector;
import backtype.storm.task.TopologyContext;
import backtype.storm.topology.IRichBolt;
import backtype.storm.topology.OutputFieldsDeclarer;
import backtype.storm.tuple.Tuple;

public class HashtagReaderBolt implements IRichBolt {
   private OutputCollector collector;

   @Override
   public void prepare(Map conf, TopologyContext context, OutputCollector collector) {
      this.collector = collector;
   }

   @Override
   public void execute(Tuple tuple) {
      Status tweet = (Status) tuple.getValueByField("tweet");
      for(HashtagEntity hashtage : tweet.getHashtagEntities()) {
         System.out.println("Hashtag: " + hashtage.getText());
         this.collector.emit(new Values(hashtage.getText()));
      }
   }

   @Override
   public void cleanup() {}

   @Override
   public void declareOutputFields(OutputFieldsDeclarer declarer) {
      declarer.declare(new Fields("hashtag"));
   }
	
   @Override
   public Map<String, Object> getComponentConfiguration() {
      return null;
   }
	
}

Hashtag counter spout

The hashtag issued will be forwarded to HashtagCounterBolt. This bolt will process all the hashtags and save each hashtag and its count in memory using a Java Map object. The complete program code is as follows.

Encoding: HashtagCounterBolt.java

import java.util.HashMap;
import java.util.Map;

import backtype.storm.tuple.Fields;
import backtype.storm.tuple.Values;

import backtype.storm.task.OutputCollector;
import backtype.storm.task.TopologyContext;

import backtype.storm.topology.IRichBolt;
import backtype.storm.topology.OutputFieldsDeclarer;
import backtype.storm.tuple.Tuple;

public class HashtagCounterBolt implements IRichBolt {
   Map<String, Integer> counterMap;
   private OutputCollector collector;

   @Override
   public void prepare(Map conf, TopologyContext context, OutputCollector collector) {
      this.counterMap = new HashMap<String, Integer>();
      this.collector = collector;
   }

   @Override
   public void execute(Tuple tuple) {
      String key = tuple.getString(0);

      if(!counterMap.containsKey(key)){
         counterMap.put(key, 1);
      }else{
         Integer c = counterMap.get(key) + 1;
         counterMap.put(key, c);
      }
		
      collector.ack(tuple);
   }

   @Override
   public void cleanup() {
      for(Map.Entry<String, Integer> entry:counterMap.entrySet()){
         System.out.println("Result: " + entry.getKey()+" : " + entry.getValue());
      }
   }

   @Override
   public void declareOutputFields(OutputFieldsDeclarer declarer) {
      declarer.declare(new Fields("hashtag"));
   }
	
   @Override
   public Map<String, Object> getComponentConfiguration() {
      return null;
   }
	
}

Submit Topology

Submit Topology is the main application. The Twitter topology consists of TwitterSampleSpout, HashtagReaderBolt and HashtagCounterBolt. The following program code shows how to submit a topology.

Coding: TwitterHashtagStorm.java

import java.util.*;

import backtype.storm.tuple.Fields;
import backtype.storm.tuple.Values;
import backtype.storm.Config;
import backtype.storm.LocalCluster;
import backtype.storm.topology.TopologyBuilder;

public class TwitterHashtagStorm {
   public static void main(String[] args) throws Exception{
      String consumerKey = args[0];
      String consumerSecret = args[1];
		
      String accessToken = args[2];
      String accessTokenSecret = args[3];
		
      String[] arguments = args.clone();
      String[] keyWords = Arrays.copyOfRange(arguments, 4, arguments.length);
		
      Config config = new Config();
      config.setDebug(true);
		
      TopologyBuilder builder = new TopologyBuilder();
      builder.setSpout("twitter-spout", new TwitterSampleSpout(consumerKey,
         consumerSecret, accessToken, accessTokenSecret, keyWords));

      builder.setBolt("twitter-hashtag-reader-bolt", new HashtagReaderBolt())
         .shuffleGrouping("twitter-spout");

      builder.setBolt("twitter-hashtag-counter-bolt", new HashtagCounterBolt())
         .fieldsGrouping("twitter-hashtag-reader-bolt", new Fields("hashtag"));
			
      LocalCluster cluster = new LocalCluster();
      cluster.submitTopology("TwitterHashtagStorm", config,
         builder.createTopology());
      Thread.sleep(10000);
      cluster.shutdown();
   }
}

Building and running the application

The complete application has four Java codes. They are as follows -

TwitterSampleSpout.java

• HashtagReaderBolt.java
• HashtagCounterBolt.java
• TwitterHashtagStorm.java
You can compile the application using the following command -

javac -cp “/path/to/storm/apache-storm-0.9.5/lib/*”:”/path/to/twitter4j/lib/*” *.java

Execute the application using the following command -

javac -cp “/path/to/storm/apache-storm-0.9.5/lib/*”:”/path/to/twitter4j/lib/*”:.
TwitterHashtagStorm <customerkey> <customersecret> <accesstoken> <accesstokensecret>
<keyword1> <keyword2> … <keywordN>

Output

The application will print the currently available hashtags and their count. The output should resemble the following -

Result: jazztastic : 1
Result: foodie : 1
Result: Redskins : 1
Result: Recipe : 1
Result: cook : 1
Result: android : 1
Result: food : 2
Result: NoToxicHorseMeat : 1
Result: Purrs4Peace : 1
Result: livemusic : 1
Result: VIPremium : 1
Result: Frome : 1
Result: SundayRoast : 1
Result: Millennials : 1
Result: HealthWithKier : 1
Result: LPs30DaysofGratitude : 1
Result: cooking : 1
Result: gameinsight : 1
Result: Countryfile : 1
Result: androidgames : 1