X-Git-Url: http://juplo.de/gitweb/?a=blobdiff_plain;f=src%2Fmain%2Fjava%2Fde%2Fjuplo%2Fkafka%2FEndlessConsumer.java;h=01f9057a055aa360bb8d98cca1641073df19afe0;hb=a6a0a22a5fa34a01b0e8b2bc1e0e2b82d7b60f33;hp=2a3445c87a5d96837a77b6a2619a5ccede523f30;hpb=e53004f3133b737699e995a3b18fff28203a0e8c;p=demos%2Fkafka%2Ftraining

diff --git a/src/main/java/de/juplo/kafka/EndlessConsumer.java b/src/main/java/de/juplo/kafka/EndlessConsumer.java
index 2a3445c..01f9057 100644
--- a/src/main/java/de/juplo/kafka/EndlessConsumer.java
+++ b/src/main/java/de/juplo/kafka/EndlessConsumer.java
@@ -8,25 +8,32 @@ import org.apache.kafka.common.errors.RecordDeserializationException;
 import org.apache.kafka.common.errors.WakeupException;
 
 import javax.annotation.PreDestroy;
+import java.time.Clock;
 import java.time.Duration;
+import java.time.Instant;
 import java.util.*;
 import java.util.concurrent.ExecutionException;
 import java.util.concurrent.ExecutorService;
 import java.util.concurrent.locks.Condition;
 import java.util.concurrent.locks.Lock;
 import java.util.concurrent.locks.ReentrantLock;
+import java.util.regex.Pattern;
 
 
 @Slf4j
 @RequiredArgsConstructor
-public class EndlessConsumer<K, V> implements ConsumerRebalanceListener, Runnable
+public class EndlessConsumer implements ConsumerRebalanceListener, Runnable
 {
+  final static Pattern PATTERN = Pattern.compile("\\W+");
+
+
   private final ExecutorService executor;
   private final PartitionStatisticsRepository repository;
   private final String id;
   private final String topic;
-  private final Consumer<K, V> consumer;
-  private final java.util.function.Consumer<ConsumerRecord<K, V>> handler;
+  private final Clock clock;
+  private final Duration commitInterval;
+  private final Consumer<String, String> consumer;
 
   private final Lock lock = new ReentrantLock();
   private final Condition condition = lock.newCondition();
@@ -34,8 +41,7 @@ public class EndlessConsumer<K, V> implements ConsumerRebalanceListener, Runnabl
   private Exception exception;
   private long consumed = 0;
 
-  private final Map<Integer, Map<String, Long>> seen = new HashMap<>();
-  private final Map<Integer, Long> offsets = new HashMap<>();
+  private final Map<Integer, Map<String, Map<String, Long>>> seen = new HashMap<>();
 
 
   @Override
@@ -45,24 +51,12 @@ public class EndlessConsumer<K, V> implements ConsumerRebalanceListener, Runnabl
     {
       Integer partition = tp.partition();
       Long newOffset = consumer.position(tp);
-      Long oldOffset = offsets.remove(partition);
       log.info(
-          "{} - removing partition: {}, consumed {} records (offset {} -> {})",
+          "{} - removing partition: {}, offset of next message {})",
           id,
           partition,
-          newOffset - oldOffset,
-          oldOffset,
           newOffset);
-      Map<String, Long> removed = seen.remove(partition);
-      for (String key : removed.keySet())
-      {
-        log.info(
-            "{} - Seen {} messages for partition={}|key={}",
-            id,
-            removed.get(key),
-            partition,
-            key);
-      }
+      Map<String, Map<String, Long>> removed = seen.remove(partition);
       repository.save(new StatisticsDocument(partition, removed, consumer.position(tp)));
     });
   }
@@ -79,7 +73,12 @@ public class EndlessConsumer<K, V> implements ConsumerRebalanceListener, Runnabl
           repository
               .findById(Integer.toString(partition))
               .orElse(new StatisticsDocument(partition));
-      consumer.seek(tp, document.offset);
+      if (document.offset >= 0)
+      {
+        // Only seek, if a stored offset was found
+        // Otherwise: Use initial offset, generated by Kafka
+        consumer.seek(tp, document.offset);
+      }
       seen.put(partition, document.statistics);
     });
   }
@@ -93,14 +92,16 @@ public class EndlessConsumer<K, V> implements ConsumerRebalanceListener, Runnabl
       log.info("{} - Subscribing to topic {}", id, topic);
       consumer.subscribe(Arrays.asList(topic), this);
 
+      Instant lastCommit = clock.instant();
+
       while (true)
       {
-        ConsumerRecords<K, V> records =
+        ConsumerRecords<String, String> records =
             consumer.poll(Duration.ofSeconds(1));
 
         // Do something with the data...
         log.info("{} - Received {} messages", id, records.count());
-        for (ConsumerRecord<K, V> record : records)
+        for (ConsumerRecord<String, String> record : records)
         {
           log.info(
               "{} - {}: {}/{} - {}={}",
@@ -112,27 +113,44 @@ public class EndlessConsumer<K, V> implements ConsumerRebalanceListener, Runnabl
               record.value()
           );
 
-          handler.accept(record);
-
           consumed++;
 
           Integer partition = record.partition();
-          String key = record.key() == null ? "NULL" : record.key().toString();
-          Map<String, Long> byKey = seen.get(partition);
+          String user = record.key();
+          Map<String, Map<String, Long>> users = seen.get(partition);
 
-          if (!byKey.containsKey(key))
-            byKey.put(key, 0l);
+          Map<String, Long> words = users.get(user);
+          if (words == null)
+          {
+            words = new HashMap<>();
+            users.put(user, words);
+          }
 
-          long seenByKey = byKey.get(key);
-          seenByKey++;
-          byKey.put(key, seenByKey);
+          for (String word : PATTERN.split(record.value()))
+          {
+            Long num = words.get(word);
+            if (num == null)
+            {
+              num = 1l;
+            }
+            else
+            {
+              num++;
+            }
+            words.put(word, num);
+          }
         }
 
-        seen.forEach((partiton, statistics) -> repository.save(
-            new StatisticsDocument(
-                partiton,
-                statistics,
-                consumer.position(new TopicPartition(topic, partiton)))));
+        if (lastCommit.plus(commitInterval).isBefore(clock.instant()))
+        {
+          log.debug("Storing data and offsets, last commit: {}", lastCommit);
+          seen.forEach((partiton, statistics) -> repository.save(
+              new StatisticsDocument(
+                  partiton,
+                  statistics,
+                  consumer.position(new TopicPartition(topic, partiton)))));
+          lastCommit = clock.instant();
+        }
       }
     }
     catch(WakeupException e)
@@ -200,7 +218,7 @@ public class EndlessConsumer<K, V> implements ConsumerRebalanceListener, Runnabl
     }
   }
 
-  public Map<Integer, Map<String, Long>> getSeen()
+  public Map<Integer, Map<String, Map<String, Long>>> getSeen()
   {
     return seen;
   }
@@ -224,7 +242,7 @@ public class EndlessConsumer<K, V> implements ConsumerRebalanceListener, Runnabl
     }
   }
 
-  public synchronized void stop() throws ExecutionException, InterruptedException
+  public synchronized void stop() throws InterruptedException
   {
     lock.lock();
     try
@@ -247,22 +265,7 @@ public class EndlessConsumer<K, V> implements ConsumerRebalanceListener, Runnabl
   public void destroy() throws ExecutionException, InterruptedException
   {
     log.info("{} - Destroy!", id);
-    try
-    {
-      stop();
-    }
-    catch (IllegalStateException e)
-    {
-      log.info("{} - Was already stopped", id);
-    }
-    catch (Exception e)
-    {
-      log.error("{} - Unexpected exception while trying to stop the consumer", id, e);
-    }
-    finally
-    {
-      log.info("{}: Consumed {} messages in total, exiting!", id, consumed);
-    }
+    log.info("{}: Consumed {} messages in total, exiting!", id, consumed);
   }
 
   public boolean running()