/// <summary>
/// Return a new DStream by applying ReduceByKey to each RDD.
/// </summary>
/// <typeparam name="K"></typeparam>
/// <typeparam name="V"></typeparam>
/// <param name="self"></param>
/// <param name="reduceFunc"></param>
/// <param name="numPartitions"></param>
/// <returns></returns>
public static DStream <Tuple <K, V> > ReduceByKey <K, V>(this DStream <Tuple <K, V> > self, Func <V, V, V> reduceFunc, int numPartitions = 0)
{
var locallyCombined = self.MapPartitionsWithIndex(new GroupByMergeHelper <K, V>(reduceFunc).Execute, true);
var shuffled = locallyCombined.PartitionBy(numPartitions);
return(shuffled.MapPartitionsWithIndex(new GroupByMergeHelper <K, V>(reduceFunc).Execute, true));
}
/// <summary>
/// Create an input stream that directly pulls messages from a Kafka Broker and specific offset.
///
/// This is not a receiver based Kafka input stream, it directly pulls the message from Kafka
/// in each batch duration and processed without storing.
///
/// This does not use Zookeeper to store offsets. The consumed offsets are tracked
/// by the stream itself. For interoperability with Kafka monitoring tools that depend on
/// Zookeeper, you have to update Kafka/Zookeeper yourself from the streaming application.
/// You can access the offsets used in each batch from the generated RDDs (see
/// [[org.apache.spark.streaming.kafka.HasOffsetRanges]]).
/// To recover from driver failures, you have to enable checkpointing in the StreamingContext.
/// The information on consumed offset can be recovered from the checkpoint.
/// See the programming guide for details (constraints, etc.).
///
/// </summary>
/// <param name="ssc">Spark Streaming Context</param>
/// <param name="topics">list of topic_name to consume.</param>
/// <param name="kafkaParams">
/// Additional params for Kafka. Requires "metadata.broker.list" or "bootstrap.servers" to be set
/// with Kafka broker(s) (NOT zookeeper servers), specified in host1:port1,host2:port2 form.
/// </param>
/// <param name="fromOffsets">Per-topic/partition Kafka offsets defining the (inclusive) starting point of the stream.</param>
/// <param name="readFunc">user function to process the kafka data.</param>
/// <returns>A DStream object</returns>
public static DStream <T> CreateDirectStream <T>(StreamingContext ssc, List <string> topics, Dictionary <string, string> kafkaParams, Dictionary <string, long> fromOffsets, Func <int, IEnumerable <KeyValuePair <byte[], byte[]> >, IEnumerable <T> > readFunc)
{
int numPartitions = GetNumPartitionsFromConfig(ssc, topics, kafkaParams);
if (ssc.SparkContext.SparkConf.SparkConfProxy.GetInt("spark.mobius.streaming.kafka.numReceivers", 0) <= 0)
{
var dstream = new DStream <KeyValuePair <byte[], byte[]> >(ssc.streamingContextProxy.DirectKafkaStreamWithRepartition(topics, kafkaParams, fromOffsets, numPartitions, null, null), ssc, SerializedMode.Pair);
return(dstream.MapPartitionsWithIndex(readFunc, true));
}
var mapPartitionsWithIndexHelper = new MapPartitionsWithIndexHelper <KeyValuePair <byte[], byte[]>, T>(readFunc, true);
var transformHelper = new TransformHelper <KeyValuePair <byte[], byte[]>, T>(mapPartitionsWithIndexHelper.Execute);
var transformDynamicHelper = new TransformDynamicHelper <KeyValuePair <byte[], byte[]>, T>(transformHelper.Execute);
Func <double, RDD <dynamic>, RDD <dynamic> > func = transformDynamicHelper.Execute;
var formatter = new BinaryFormatter();
var stream = new MemoryStream();
formatter.Serialize(stream, func);
byte[] readFuncBytes = stream.ToArray();
string serializationMode = SerializedMode.Pair.ToString();
return(new DStream <T>(ssc.streamingContextProxy.DirectKafkaStreamWithRepartition(topics, kafkaParams, fromOffsets, numPartitions, readFuncBytes, serializationMode), ssc));
}
