public ISparkContextProxy CreateSparkContext(ISparkConfProxy conf)
{
JvmObjectReference jvmSparkContextReference = JvmBridge.CallConstructor("org.apache.spark.SparkContext", (conf as SparkConfIpcProxy).JvmSparkConfReference);
JvmObjectReference jvmJavaContextReference = JvmBridge.CallConstructor("org.apache.spark.api.java.JavaSparkContext", new object[] { jvmSparkContextReference });
sparkContextProxy = new SparkContextIpcProxy(jvmSparkContextReference, jvmJavaContextReference);
return sparkContextProxy;
}
public IDStreamProxy SocketTextStream(string hostname, int port, StorageLevelType storageLevelType)
{
JvmObjectReference jlevel = SparkContextIpcProxy.GetJavaStorageLevel(storageLevelType);
var jstream = new JvmObjectReference(SparkCLRIpcProxy.JvmBridge.CallNonStaticJavaMethod(jvmJavaStreamingReference, "socketTextStream", hostname, port, jlevel).ToString());
return(new DStreamIpcProxy(jstream));
}
public void Persist(StorageLevelType storageLevelType)
{
var jstorageLevel = SparkContextIpcProxy.GetJavaStorageLevel(storageLevelType);
SparkCLRIpcProxy.JvmBridge.CallNonStaticJavaMethod(
jvmDataFrameReference, "persist", new object[] { jstorageLevel });
}
public ISparkContextProxy CreateSparkContext(ISparkConfProxy conf)
{
JvmObjectReference jvmSparkContextReference = JvmBridge.CallConstructor("org.apache.spark.SparkContext", (conf as SparkConfIpcProxy).JvmSparkConfReference);
JvmObjectReference jvmJavaContextReference = JvmBridge.CallConstructor("org.apache.spark.api.java.JavaSparkContext", new object[] { jvmSparkContextReference });
sparkContextProxy = new SparkContextIpcProxy(jvmSparkContextReference, jvmJavaContextReference);
return(sparkContextProxy);
}
public IDStreamProxy KafkaStream(Dictionary <string, int> topics, Dictionary <string, string> kafkaParams, StorageLevelType storageLevelType)
{
JvmObjectReference jtopics = SparkContextIpcProxy.GetJavaMap <string, int>(topics);
JvmObjectReference jkafkaParams = SparkContextIpcProxy.GetJavaMap <string, string>(kafkaParams);
JvmObjectReference jlevel = SparkContextIpcProxy.GetJavaStorageLevel(storageLevelType);
JvmObjectReference jhelper = SparkCLRIpcProxy.JvmBridge.CallConstructor("org.apache.spark.streaming.kafka.KafkaUtilsPythonHelper", new object[] { });
var jstream = new JvmObjectReference(SparkCLRIpcProxy.JvmBridge.CallNonStaticJavaMethod(jhelper, "createStream", new object[] { jvmJavaStreamingReference, jkafkaParams, jtopics, jlevel }).ToString());
return(new DStreamIpcProxy(jstream));
}
public IDStreamProxy KafkaStream(IEnumerable <Tuple <string, int> > topics, IEnumerable <Tuple <string, string> > kafkaParams, StorageLevelType storageLevelType)
{
JvmObjectReference jtopics = JvmBridgeUtils.GetJavaMap <string, int>(topics);
JvmObjectReference jkafkaParams = JvmBridgeUtils.GetJavaMap <string, string>(kafkaParams);
JvmObjectReference jlevel = SparkContextIpcProxy.GetJavaStorageLevel(storageLevelType);
// KafkaUtilsPythonHelper: external/kafka/src/main/scala/org/apache/spark/streaming/kafka/KafkaUtils.scala
JvmObjectReference jhelper = SparkCLRIpcProxy.JvmBridge.CallConstructor("org.apache.spark.streaming.kafka.KafkaUtilsPythonHelper", new object[] { });
var jstream = new JvmObjectReference(SparkCLRIpcProxy.JvmBridge.CallNonStaticJavaMethod(jhelper, "createStream", new object[] { jvmJavaStreamingReference, jkafkaParams, jtopics, jlevel }).ToString());
return(new DStreamIpcProxy(jstream));
}
public IDStreamProxy Union(IDStreamProxy firstDStream, IDStreamProxy[] otherDStreams)
{
return(new DStreamIpcProxy(
new JvmObjectReference(
(string)SparkCLRIpcProxy.JvmBridge.CallNonStaticJavaMethod(jvmJavaStreamingReference, "union",
new object[]
{
(firstDStream as DStreamIpcProxy).javaDStreamReference,
SparkContextIpcProxy.GetJavaList <JvmObjectReference>(otherDStreams.Select(x => (x as DStreamIpcProxy).javaDStreamReference))
}
))));
}
public IDStreamProxy EventHubsUnionStream(Dictionary <string, string> eventHubsParams, StorageLevelType storageLevelType)
{
JvmObjectReference eventHubsParamsReference = JvmBridgeUtils.GetScalaMutableMap <string, string>(eventHubsParams);
JvmObjectReference storageLevelTypeReference = SparkContextIpcProxy.GetJavaStorageLevel(storageLevelType);
return
(new DStreamIpcProxy(
new JvmObjectReference(
SparkCLRIpcProxy.JvmBridge.CallStaticJavaMethod(
"org.apache.spark.streaming.api.csharp.EventHubsUtils", "createUnionStream",
new object[] { jvmJavaStreamingReference, eventHubsParamsReference, storageLevelTypeReference })
.ToString())));
}
public IDStreamProxy DirectKafkaStream(List <string> topics, Dictionary <string, string> kafkaParams, Dictionary <string, long> fromOffsets)
{
JvmObjectReference jtopics = SparkContextIpcProxy.GetJavaSet <string>(topics);
JvmObjectReference jkafkaParams = SparkContextIpcProxy.GetJavaMap <string, string>(kafkaParams);
var jTopicAndPartitions = fromOffsets.Select(x =>
new KeyValuePair <JvmObjectReference, long>
(
SparkCLRIpcProxy.JvmBridge.CallConstructor("kafka.common.TopicAndPartition", new object[] { x.Key.Split(':')[0], int.Parse(x.Key.Split(':')[1]) }),
x.Value
)
);
JvmObjectReference jfromOffsets = SparkContextIpcProxy.GetJavaMap <JvmObjectReference, long>(jTopicAndPartitions);
JvmObjectReference jhelper = SparkCLRIpcProxy.JvmBridge.CallConstructor("org.apache.spark.streaming.kafka.KafkaUtilsPythonHelper", new object[] { });
var jstream = new JvmObjectReference(SparkCLRIpcProxy.JvmBridge.CallNonStaticJavaMethod(jhelper, "createDirectStream", new object[] { jvmJavaStreamingReference, jkafkaParams, jtopics, jfromOffsets }).ToString());
return(new DStreamIpcProxy(jstream));
}
public BroadcastIpcProxy(JvmObjectReference jvmBroadcastReference, SparkContextIpcProxy sparkContextIpcProxy)
{
this.jvmBroadcastReference = jvmBroadcastReference;
this.sparkContextIpcProxy = sparkContextIpcProxy;
}
public BroadcastIpcProxy(JvmObjectReference jvmBroadcastReference, SparkContextIpcProxy sparkContextIpcProxy)
{
this.jvmBroadcastReference = jvmBroadcastReference;
this.sparkContextIpcProxy = sparkContextIpcProxy;
}
public void SaveAsHadoopFile(string path, string outputFormatClass, string keyClass, string valueClass, IEnumerable <KeyValuePair <string, string> > conf, string compressionCodecClass)
{
var jconf = SparkContextIpcProxy.GetJavaMap <string, string>(conf);
SparkCLRIpcProxy.JvmBridge.CallStaticJavaMethod("org.apache.spark.api.python.PythonRDD", "saveAsHadoopFile", new object[] { jvmRddReference, false, path, outputFormatClass, keyClass, valueClass, null, null, jconf, compressionCodecClass });
}
public void SaveAsHadoopDataset(IEnumerable <KeyValuePair <string, string> > conf)
{
var jconf = SparkContextIpcProxy.GetJavaMap <string, string>(conf);
SparkCLRIpcProxy.JvmBridge.CallStaticJavaMethod("org.apache.spark.api.python.PythonRDD", "saveAsHadoopDataset", new object[] { jvmRddReference, false, jconf, null, null, false });
}
public IRDDProxy SampleByKey(bool withReplacement, Dictionary <string, double> fractions, long seed)
{
var jfractions = SparkContextIpcProxy.GetJavaMap(fractions) as JvmObjectReference;
return(new RDDIpcProxy(new JvmObjectReference((string)SparkCLRIpcProxy.JvmBridge.CallNonStaticJavaMethod(jvmRddReference, "sampleByKey", new object[] { withReplacement, jfractions, seed }))));
}
