Github user tdas commented on a diff in the pull request: https://github.com/apache/spark/pull/2538#discussion_r18194443 --- Diff: streaming/src/main/scala/org/apache/spark/streaming/api/python/PythonDStream.scala --- @@ -0,0 +1,261 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.spark.streaming.api.python + +import java.util.{ArrayList => JArrayList, List => JList} +import scala.collection.JavaConversions._ +import scala.collection.JavaConverters._ +import scala.collection.mutable + +import org.apache.spark.api.java._ +import org.apache.spark.api.python._ +import org.apache.spark.rdd.RDD +import org.apache.spark.storage.StorageLevel +import org.apache.spark.streaming.{Interval, Duration, Time} +import org.apache.spark.streaming.dstream._ +import org.apache.spark.streaming.api.java._ + + +/** + * Interface for Python callback function with three arguments + */ +trait PythonRDDFunction { + def call(time: Long, rdds: JList[_]): JavaRDD[Array[Byte]] +} + +/** + * Wrapper for PythonRDDFunction + */ +private[python] class RDDFunction(pfunc: PythonRDDFunction) + extends function.Function2[JList[JavaRDD[_]], Time, JavaRDD[Array[Byte]]] with Serializable { + + def wrapRDD(rdd: Option[RDD[_]]): JavaRDD[_] = { + if (rdd.isDefined) { + JavaRDD.fromRDD(rdd.get) + } else { + null + } + } + + def some(jrdd: JavaRDD[Array[Byte]]): Option[RDD[Array[Byte]]] = { + if (jrdd != null) { + Some(jrdd.rdd) + } else { + None + } + } + + def apply(rdd: Option[RDD[_]], time: Time): Option[RDD[Array[Byte]]] = { + some(pfunc.call(time.milliseconds, List(wrapRDD(rdd)).asJava)) + } + + def apply(rdd: Option[RDD[_]], rdd2: Option[RDD[_]], time: Time): Option[RDD[Array[Byte]]] = { + some(pfunc.call(time.milliseconds, List(wrapRDD(rdd), wrapRDD(rdd2)).asJava)) + } + + // for JFunction2 + def call(rdds: JList[JavaRDD[_]], time: Time): JavaRDD[Array[Byte]] = { + pfunc.call(time.milliseconds, rdds) + } +} + +private[python] +abstract class PythonDStream(parent: DStream[_]) extends DStream[Array[Byte]] (parent.ssc) { + + override def dependencies = List(parent) + + override def slideDuration: Duration = parent.slideDuration + + val asJavaDStream = JavaDStream.fromDStream(this) +} + +private[spark] object PythonDStream { + + // helper function for DStream.foreachRDD(), + // cannot be `foreachRDD`, it will confusing py4j + def callForeachRDD(jdstream: JavaDStream[Array[Byte]], pyfunc: PythonRDDFunction){ + val func = new RDDFunction(pyfunc) + jdstream.dstream.foreachRDD((rdd, time) => func(Some(rdd), time)) + } + + // helper function for ssc.transform() + def callTransform(ssc: JavaStreamingContext, jdsteams: JList[JavaDStream[_]], + pyfunc: PythonRDDFunction) + :JavaDStream[Array[Byte]] = { + val func = new RDDFunction(pyfunc) + ssc.transform(jdsteams, func) + } + + // convert list of RDD into queue of RDDs, for ssc.queueStream() + def toRDDQueue(rdds: JArrayList[JavaRDD[Array[Byte]]]): java.util.Queue[JavaRDD[Array[Byte]]] = { + val queue = new java.util.LinkedList[JavaRDD[Array[Byte]]] + rdds.forall(queue.add(_)) + queue + } +} + +/** + * Transformed DStream in Python. + * + * If the result RDD is PythonRDD, then it will cache it as an template for future use, + * this can reduce the Python callbacks. + */ +private[spark] +class PythonTransformedDStream (parent: DStream[_], pfunc: PythonRDDFunction, + var reuse: Boolean = false) + extends PythonDStream(parent) { + + val func = new RDDFunction(pfunc) + var lastResult: PythonRDD = _ + + override def compute(validTime: Time): Option[RDD[Array[Byte]]] = { + val rdd1 = parent.getOrCompute(validTime) --- End diff -- Please add some comments on what this method does, especially regarding reuse and all. Also, how is `lastResult` used?
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org