[PR] [WIP] [SPARK-53785][SS] Memory Source for RTM [spark]

["jerrypeng (via GitHub)" <gi...@apache.org>]

jerrypeng opened a new pull request, #52502:
URL: https://github.com/apache/spark/pull/52502

   <!--
   Thanks for sending a pull request!  Here are some tips for you:
     1. If this is your first time, please read our contributor guidelines: https://spark.apache.org/contributing.html
     2. Ensure you have added or run the appropriate tests for your PR: https://spark.apache.org/developer-tools.html
     3. If the PR is unfinished, add '[WIP]' in your PR title, e.g., '[WIP][SPARK-XXXX] Your PR title ...'.
     4. Be sure to keep the PR description updated to reflect all changes.
     5. Please write your PR title to summarize what this PR proposes.
     6. If possible, provide a concise example to reproduce the issue for a faster review.
     7. If you want to add a new configuration, please read the guideline first for naming configurations in
        'core/src/main/scala/org/apache/spark/internal/config/ConfigEntry.scala'.
     8. If you want to add or modify an error type or message, please read the guideline first in
        'common/utils/src/main/resources/error/README.md'.
   -->
   
   ### What changes were proposed in this pull request?
   <!--
   Please clarify what changes you are proposing. The purpose of this section is to outline the changes and how this PR fixes the issue. 
   If possible, please consider writing useful notes for better and faster reviews in your PR. See the examples below.
     1. If you refactor some codes with changing classes, showing the class hierarchy will help reviewers.
     2. If you fix some SQL features, you can provide some references of other DBMSes.
     3. If there is design documentation, please add the link.
     4. If there is a discussion in the mailing list, please add the link.
   -->
   
   
   ### Why are the changes needed?
   <!--
   Please clarify why the changes are needed. For instance,
     1. If you propose a new API, clarify the use case for a new API.
     2. If you fix a bug, you can clarify why it is a bug.
   -->
   
   
   ### Does this PR introduce _any_ user-facing change?
   <!--
   Note that it means *any* user-facing change including all aspects such as new features, bug fixes, or other behavior changes. Documentation-only updates are not considered user-facing changes.
   
   If yes, please clarify the previous behavior and the change this PR proposes - provide the console output, description and/or an example to show the behavior difference if possible.
   If possible, please also clarify if this is a user-facing change compared to the released Spark versions or within the unreleased branches such as master.
   If no, write 'No'.
   -->
   
   
   ### How was this patch tested?
   <!--
   If tests were added, say they were added here. Please make sure to add some test cases that check the changes thoroughly including negative and positive cases if possible.
   If it was tested in a way different from regular unit tests, please clarify how you tested step by step, ideally copy and paste-able, so that other reviewers can test and check, and descendants can verify in the future.
   If tests were not added, please describe why they were not added and/or why it was difficult to add.
   If benchmark tests were added, please run the benchmarks in GitHub Actions for the consistent environment, and the instructions could accord to: https://spark.apache.org/developer-tools.html#github-workflow-benchmarks.
   -->
   
   
   ### Was this patch authored or co-authored using generative AI tooling?
   <!--
   If generative AI tooling has been used in the process of authoring this patch, please include the
   phrase: 'Generated-by: ' followed by the name of the tool and its version.
   If no, write 'No'.
   Please refer to the [ASF Generative Tooling Guidance](https://www.apache.org/legal/generative-tooling.html) for details.
   -->
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433377244


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation
+
+  @GuardedBy("this")
+  private val records = Seq.fill(numPartitions)(new ListBuffer[UnsafeRow])
+
+  private val recordEndpoint = new ContinuousRecordEndpoint(records, this)
+  @volatile private var endpointRef: RpcEndpointRef = _
+
+  override def addData(data: IterableOnce[A]): Offset = synchronized {
+    // Distribute data evenly among partition lists.
+    data.iterator.to(Seq).zipWithIndex.map {
+      case (item, index) =>
+        records(index % numPartitions) += toRow(item).copy().asInstanceOf[UnsafeRow]
+    }
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  def addData(partitionId: Int, data: IterableOnce[A]): Offset = synchronized {
+    require(
+      partitionId >= 0 && partitionId < numPartitions,
+      s"Partition ID $partitionId is out of bounds for $numPartitions partitions."
+    )
+
+    // Add data to the specified partition.
+    records(partitionId) ++= data.iterator.map(item => toRow(item).copy().asInstanceOf[UnsafeRow])
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def initialOffset(): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, 0)).toMap)
+  }
+
+  override def latestOffset(startOffset: OffsetV2, limit: ReadLimit): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)

Review Comment:
   Do we also need to add `synchronized` to `latestOffset` as it also access `records`. Although it just reads, but the `size` might be inconsistent. 



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433337678


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation
+
+  @GuardedBy("this")
+  private val records = Seq.fill(numPartitions)(new ListBuffer[UnsafeRow])
+
+  private val recordEndpoint = new ContinuousRecordEndpoint(records, this)
+  @volatile private var endpointRef: RpcEndpointRef = _
+
+  override def addData(data: IterableOnce[A]): Offset = synchronized {
+    // Distribute data evenly among partition lists.
+    data.iterator.to(Seq).zipWithIndex.map {
+      case (item, index) =>
+        records(index % numPartitions) += toRow(item).copy().asInstanceOf[UnsafeRow]
+    }
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  def addData(partitionId: Int, data: IterableOnce[A]): Offset = synchronized {
+    require(
+      partitionId >= 0 && partitionId < numPartitions,
+      s"Partition ID $partitionId is out of bounds for $numPartitions partitions."
+    )
+
+    // Add data to the specified partition.
+    records(partitionId) ++= data.iterator.map(item => toRow(item).copy().asInstanceOf[UnsafeRow])
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def initialOffset(): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, 0)).toMap)
+  }
+
+  override def latestOffset(startOffset: OffsetV2, limit: ReadLimit): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def deserializeOffset(json: String): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(Serialization.read[Map[Int, Int]](json))
+  }
+
+  override def mergeOffsets(offsets: Array[PartitionOffset]): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(
+      offsets.map {
+        case ContinuousRecordPartitionOffset(part, num) => (part, num)
+      }.toMap
+    )
+  }
+
+  override def planInputPartitions(start: OffsetV2): Array[InputPartition] = {
+    val startOffset = start.asInstanceOf[LowLatencyMemoryStreamOffset]
+    synchronized {
+      val endpointName = s"ContinuousRecordEndpoint-${java.util.UUID.randomUUID()}-$id"
+      endpointRef = recordEndpoint.rpcEnv.setupEndpoint(endpointName, recordEndpoint)
+
+      startOffset.partitionNums.map {
+        case (part, index) =>
+          LowLatencyMemoryStreamInputPartition(
+            endpointName,
+            endpointRef.address,
+            part,
+            index,
+            Int.MaxValue
+          )
+      }.toArray
+    }
+  }
+
+  override def planInputPartitions(start: OffsetV2, end: OffsetV2): Array[InputPartition] = {
+    val startOffset = start.asInstanceOf[LowLatencyMemoryStreamOffset]
+    val endOffset = end.asInstanceOf[LowLatencyMemoryStreamOffset]
+    synchronized {
+      val endpointName = s"ContinuousRecordEndpoint-${java.util.UUID.randomUUID()}-$id"
+      endpointRef = recordEndpoint.rpcEnv.setupEndpoint(endpointName, recordEndpoint)
+
+      startOffset.partitionNums.map {
+        case (part, index) =>
+          LowLatencyMemoryStreamInputPartition(
+            endpointName,
+            endpointRef.address,
+            part,
+            index,
+            endOffset.partitionNums(part)
+          )
+      }.toArray
+    }
+  }
+
+  override def createReaderFactory(): PartitionReaderFactory = {
+    new LowLatencyMemoryStreamReaderFactory(clock)
+  }
+
+  override def stop(): Unit = {
+    if (endpointRef != null) recordEndpoint.rpcEnv.stop(endpointRef)
+  }
+
+  override def commit(end: OffsetV2): Unit = {}
+
+  override def reset(): Unit = {
+    super.reset()
+    records.foreach(_.clear())
+  }
+}
+
+object LowLatencyMemoryStream {
+  protected val memoryStreamId = new AtomicInteger(0)
+
+  def apply[A: Encoder](implicit sqlContext: SQLContext): LowLatencyMemoryStream[A] =
+    new LowLatencyMemoryStream[A](memoryStreamId.getAndIncrement(), sqlContext)
+
+  def apply[A: Encoder](numPartitions: Int)(
+      implicit
+      sqlContext: SQLContext): LowLatencyMemoryStream[A] =
+    new LowLatencyMemoryStream[A](
+      memoryStreamId.getAndIncrement(),
+      sqlContext,
+      numPartitions = numPartitions
+    )
+
+  def singlePartition[A: Encoder](implicit sqlContext: SQLContext): LowLatencyMemoryStream[A] =
+    new LowLatencyMemoryStream[A](memoryStreamId.getAndIncrement(), sqlContext, 1)
+}
+
+/**
+ * An input partition for LowLatency memory stream.
+ */
+case class LowLatencyMemoryStreamInputPartition(
+    driverEndpointName: String,
+    driverEndpointAddress: RpcAddress,
+    partition: Int,
+    startOffset: Int,
+    endOffset: Int)
+    extends InputPartition
+
+class LowLatencyMemoryStreamReaderFactory(clock: Clock) extends PartitionReaderFactory {
+  override def createReader(partition: InputPartition): PartitionReader[InternalRow] = {
+    val p = partition.asInstanceOf[LowLatencyMemoryStreamInputPartition]
+    new LowLatencyMemoryStreamPartitionReader(
+      p.driverEndpointName,
+      p.driverEndpointAddress,
+      p.partition,
+      p.startOffset,
+      p.endOffset,
+      clock
+    )
+  }
+}
+
+/**
+ * An input partition reader for LowLatency memory stream.
+ *
+ * Polls the driver endpoint for new records.
+ */
+class LowLatencyMemoryStreamPartitionReader(
+    driverEndpointName: String,
+    driverEndpointAddress: RpcAddress,
+    partition: Int,
+    startOffset: Int,
+    endOffset: Int,
+    clock: Clock)
+    extends SupportsRealTimeRead[InternalRow] {
+  // ES-1365239: Avoid tracking the ref, given that we create a new one for each partition reader

Review Comment:
   ```suggestion
     // Avoid tracking the ref, given that we create a new one for each partition reader
   ```



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["jerrypeng (via GitHub)" <gi...@apache.org>]

jerrypeng commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433500838


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation
+
+  @GuardedBy("this")
+  private val records = Seq.fill(numPartitions)(new ListBuffer[UnsafeRow])
+
+  private val recordEndpoint = new ContinuousRecordEndpoint(records, this)
+  @volatile private var endpointRef: RpcEndpointRef = _
+
+  override def addData(data: IterableOnce[A]): Offset = synchronized {
+    // Distribute data evenly among partition lists.
+    data.iterator.to(Seq).zipWithIndex.map {
+      case (item, index) =>
+        records(index % numPartitions) += toRow(item).copy().asInstanceOf[UnsafeRow]
+    }
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  def addData(partitionId: Int, data: IterableOnce[A]): Offset = synchronized {
+    require(
+      partitionId >= 0 && partitionId < numPartitions,
+      s"Partition ID $partitionId is out of bounds for $numPartitions partitions."
+    )
+
+    // Add data to the specified partition.
+    records(partitionId) ++= data.iterator.map(item => toRow(item).copy().asInstanceOf[UnsafeRow])
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def initialOffset(): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, 0)).toMap)
+  }
+
+  override def latestOffset(startOffset: OffsetV2, limit: ReadLimit): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def deserializeOffset(json: String): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(Serialization.read[Map[Int, Int]](json))
+  }
+
+  override def mergeOffsets(offsets: Array[PartitionOffset]): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(
+      offsets.map {
+        case ContinuousRecordPartitionOffset(part, num) => (part, num)
+      }.toMap
+    )
+  }
+
+  override def planInputPartitions(start: OffsetV2): Array[InputPartition] = {
+    val startOffset = start.asInstanceOf[LowLatencyMemoryStreamOffset]
+    synchronized {
+      val endpointName = s"ContinuousRecordEndpoint-${java.util.UUID.randomUUID()}-$id"
+      endpointRef = recordEndpoint.rpcEnv.setupEndpoint(endpointName, recordEndpoint)
+
+      startOffset.partitionNums.map {
+        case (part, index) =>
+          LowLatencyMemoryStreamInputPartition(
+            endpointName,
+            endpointRef.address,
+            part,
+            index,
+            Int.MaxValue
+          )
+      }.toArray
+    }
+  }
+
+  override def planInputPartitions(start: OffsetV2, end: OffsetV2): Array[InputPartition] = {
+    val startOffset = start.asInstanceOf[LowLatencyMemoryStreamOffset]
+    val endOffset = end.asInstanceOf[LowLatencyMemoryStreamOffset]
+    synchronized {
+      val endpointName = s"ContinuousRecordEndpoint-${java.util.UUID.randomUUID()}-$id"
+      endpointRef = recordEndpoint.rpcEnv.setupEndpoint(endpointName, recordEndpoint)
+
+      startOffset.partitionNums.map {
+        case (part, index) =>
+          LowLatencyMemoryStreamInputPartition(
+            endpointName,
+            endpointRef.address,
+            part,
+            index,
+            endOffset.partitionNums(part)
+          )
+      }.toArray
+    }
+  }
+
+  override def createReaderFactory(): PartitionReaderFactory = {
+    new LowLatencyMemoryStreamReaderFactory(clock)
+  }
+
+  override def stop(): Unit = {
+    if (endpointRef != null) recordEndpoint.rpcEnv.stop(endpointRef)
+  }
+
+  override def commit(end: OffsetV2): Unit = {}
+
+  override def reset(): Unit = {
+    super.reset()
+    records.foreach(_.clear())

Review Comment:
   will add



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["jerrypeng (via GitHub)" <gi...@apache.org>]

jerrypeng commented on PR #52502:
URL: https://github.com/apache/spark/pull/52502#issuecomment-3416372637

   @viirya thank you for the review and merging!


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433380105


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation
+
+  @GuardedBy("this")
+  private val records = Seq.fill(numPartitions)(new ListBuffer[UnsafeRow])
+
+  private val recordEndpoint = new ContinuousRecordEndpoint(records, this)
+  @volatile private var endpointRef: RpcEndpointRef = _
+
+  override def addData(data: IterableOnce[A]): Offset = synchronized {
+    // Distribute data evenly among partition lists.
+    data.iterator.to(Seq).zipWithIndex.map {
+      case (item, index) =>
+        records(index % numPartitions) += toRow(item).copy().asInstanceOf[UnsafeRow]
+    }
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  def addData(partitionId: Int, data: IterableOnce[A]): Offset = synchronized {
+    require(
+      partitionId >= 0 && partitionId < numPartitions,
+      s"Partition ID $partitionId is out of bounds for $numPartitions partitions."
+    )
+
+    // Add data to the specified partition.
+    records(partitionId) ++= data.iterator.map(item => toRow(item).copy().asInstanceOf[UnsafeRow])
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def initialOffset(): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, 0)).toMap)
+  }
+
+  override def latestOffset(startOffset: OffsetV2, limit: ReadLimit): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def deserializeOffset(json: String): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(Serialization.read[Map[Int, Int]](json))
+  }
+
+  override def mergeOffsets(offsets: Array[PartitionOffset]): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(
+      offsets.map {
+        case ContinuousRecordPartitionOffset(part, num) => (part, num)
+      }.toMap
+    )
+  }
+
+  override def planInputPartitions(start: OffsetV2): Array[InputPartition] = {
+    val startOffset = start.asInstanceOf[LowLatencyMemoryStreamOffset]
+    synchronized {
+      val endpointName = s"ContinuousRecordEndpoint-${java.util.UUID.randomUUID()}-$id"
+      endpointRef = recordEndpoint.rpcEnv.setupEndpoint(endpointName, recordEndpoint)
+
+      startOffset.partitionNums.map {
+        case (part, index) =>
+          LowLatencyMemoryStreamInputPartition(
+            endpointName,
+            endpointRef.address,
+            part,
+            index,
+            Int.MaxValue
+          )
+      }.toArray
+    }
+  }
+
+  override def planInputPartitions(start: OffsetV2, end: OffsetV2): Array[InputPartition] = {
+    val startOffset = start.asInstanceOf[LowLatencyMemoryStreamOffset]
+    val endOffset = end.asInstanceOf[LowLatencyMemoryStreamOffset]
+    synchronized {
+      val endpointName = s"ContinuousRecordEndpoint-${java.util.UUID.randomUUID()}-$id"
+      endpointRef = recordEndpoint.rpcEnv.setupEndpoint(endpointName, recordEndpoint)
+
+      startOffset.partitionNums.map {
+        case (part, index) =>
+          LowLatencyMemoryStreamInputPartition(
+            endpointName,
+            endpointRef.address,
+            part,
+            index,
+            endOffset.partitionNums(part)
+          )
+      }.toArray
+    }
+  }
+
+  override def createReaderFactory(): PartitionReaderFactory = {
+    new LowLatencyMemoryStreamReaderFactory(clock)
+  }
+
+  override def stop(): Unit = {
+    if (endpointRef != null) recordEndpoint.rpcEnv.stop(endpointRef)
+  }
+
+  override def commit(end: OffsetV2): Unit = {}
+
+  override def reset(): Unit = {
+    super.reset()
+    records.foreach(_.clear())

Review Comment:
   Need `synchronized` too?



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433363054


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation
+
+  @GuardedBy("this")
+  private val records = Seq.fill(numPartitions)(new ListBuffer[UnsafeRow])
+
+  private val recordEndpoint = new ContinuousRecordEndpoint(records, this)
+  @volatile private var endpointRef: RpcEndpointRef = _
+
+  override def addData(data: IterableOnce[A]): Offset = synchronized {
+    // Distribute data evenly among partition lists.
+    data.iterator.to(Seq).zipWithIndex.map {
+      case (item, index) =>
+        records(index % numPartitions) += toRow(item).copy().asInstanceOf[UnsafeRow]
+    }
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  def addData(partitionId: Int, data: IterableOnce[A]): Offset = synchronized {
+    require(
+      partitionId >= 0 && partitionId < numPartitions,
+      s"Partition ID $partitionId is out of bounds for $numPartitions partitions."
+    )
+
+    // Add data to the specified partition.
+    records(partitionId) ++= data.iterator.map(item => toRow(item).copy().asInstanceOf[UnsafeRow])
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def initialOffset(): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, 0)).toMap)
+  }
+
+  override def latestOffset(startOffset: OffsetV2, limit: ReadLimit): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def deserializeOffset(json: String): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(Serialization.read[Map[Int, Int]](json))
+  }
+
+  override def mergeOffsets(offsets: Array[PartitionOffset]): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(
+      offsets.map {
+        case ContinuousRecordPartitionOffset(part, num) => (part, num)
+      }.toMap
+    )
+  }
+
+  override def planInputPartitions(start: OffsetV2): Array[InputPartition] = {
+    val startOffset = start.asInstanceOf[LowLatencyMemoryStreamOffset]
+    synchronized {
+      val endpointName = s"ContinuousRecordEndpoint-${java.util.UUID.randomUUID()}-$id"

Review Comment:
   Hmm, oh, you are reusing `ContinuousRecordEndpoint`. I see.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["jerrypeng (via GitHub)" <gi...@apache.org>]

jerrypeng commented on PR #52502:
URL: https://github.com/apache/spark/pull/52502#issuecomment-3408414463

   @viirya thank you for the review.  I have addressed your comments.  PTAL!


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya closed pull request #52502:  [SPARK-53785][SS] Memory Source for RTM
URL: https://github.com/apache/spark/pull/52502


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["jerrypeng (via GitHub)" <gi...@apache.org>]

jerrypeng commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433346914


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation

Review Comment:
   will remove



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["jerrypeng (via GitHub)" <gi...@apache.org>]

jerrypeng commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2431306058


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/v2/RealTimeStreamScanExec.scala:
##########
@@ -0,0 +1,39 @@
+/*
+ * 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.sql.execution.datasources.v2
+
+import org.apache.spark.util.{Clock, SystemClock}
+
+/* The singleton object to control the time in testing */
+object LowLatencyClock {

Review Comment:
   This is the follow PR to provide more context:
   https://github.com/apache/spark/pull/52620



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2431344140


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/runtime/memory.scala:
##########
@@ -158,6 +158,20 @@ case class MemoryStream[A : Encoder](
     id: Int,
     sqlContext: SQLContext,
     numPartitions: Option[Int] = None)
+  extends MemoryStreamBaseClass[A](
+    id, sqlContext, numPartitions = numPartitions)
+
+/**
+ * A [[Source]] that produces value stored in memory as they are added by the user.  This [[Source]]
+ * is intended for use in unit tests as it can only replay data when the object is still
+ * available.
+ *
+ * If numPartitions is provided, the rows will be redistributed to the given number of partitions.
+ */

Review Comment:
   Hm, this doc is same and copied from `MemoryStream`. Do we need to update the doc of `MemoryStream` so it can be more specified or reflect the difference between it and LowLatencyMemoryStream?
   



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on PR #52502:
URL: https://github.com/apache/spark/pull/52502#issuecomment-3416119661

   Merged to master. Thanks @jerrypeng 


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["jerrypeng (via GitHub)" <gi...@apache.org>]

jerrypeng commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433390660


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/runtime/memory.scala:
##########
@@ -158,6 +158,20 @@ case class MemoryStream[A : Encoder](
     id: Int,
     sqlContext: SQLContext,
     numPartitions: Option[Int] = None)
+  extends MemoryStreamBaseClass[A](
+    id, sqlContext, numPartitions = numPartitions)
+
+/**
+ * A [[Source]] that produces value stored in memory as they are added by the user.  This [[Source]]
+ * is intended for use in unit tests as it can only replay data when the object is still
+ * available.
+ *
+ * If numPartitions is provided, the rows will be redistributed to the given number of partitions.
+ */

Review Comment:
   Let me remove this.  The difference is described in the class LowLatencyMemoryStream



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["jerrypeng (via GitHub)" <gi...@apache.org>]

jerrypeng commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2427848039


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/v2/RealTimeStreamScanExec.scala:
##########
@@ -0,0 +1,39 @@
+/*
+ * 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.sql.execution.datasources.v2
+
+import org.apache.spark.util.{Clock, SystemClock}
+
+/* The singleton object to control the time in testing */
+object LowLatencyClock {

Review Comment:
   Planning to add additional code to this file i.e. the actual implementation of RealTimeStreamScanExec.  Adding LowLatencyClock here since it is used by LowLatencyMemorySource.  Trying to keep the PRs small :) 



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433335809


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation
+
+  @GuardedBy("this")
+  private val records = Seq.fill(numPartitions)(new ListBuffer[UnsafeRow])
+
+  private val recordEndpoint = new ContinuousRecordEndpoint(records, this)
+  @volatile private var endpointRef: RpcEndpointRef = _
+
+  override def addData(data: IterableOnce[A]): Offset = synchronized {
+    // Distribute data evenly among partition lists.
+    data.iterator.to(Seq).zipWithIndex.map {
+      case (item, index) =>
+        records(index % numPartitions) += toRow(item).copy().asInstanceOf[UnsafeRow]
+    }
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  def addData(partitionId: Int, data: IterableOnce[A]): Offset = synchronized {
+    require(
+      partitionId >= 0 && partitionId < numPartitions,
+      s"Partition ID $partitionId is out of bounds for $numPartitions partitions."
+    )
+
+    // Add data to the specified partition.
+    records(partitionId) ++= data.iterator.map(item => toRow(item).copy().asInstanceOf[UnsafeRow])
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def initialOffset(): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, 0)).toMap)
+  }
+
+  override def latestOffset(startOffset: OffsetV2, limit: ReadLimit): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def deserializeOffset(json: String): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(Serialization.read[Map[Int, Int]](json))
+  }
+
+  override def mergeOffsets(offsets: Array[PartitionOffset]): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(
+      offsets.map {
+        case ContinuousRecordPartitionOffset(part, num) => (part, num)
+      }.toMap
+    )
+  }
+
+  override def planInputPartitions(start: OffsetV2): Array[InputPartition] = {
+    val startOffset = start.asInstanceOf[LowLatencyMemoryStreamOffset]
+    synchronized {
+      val endpointName = s"ContinuousRecordEndpoint-${java.util.UUID.randomUUID()}-$id"
+      endpointRef = recordEndpoint.rpcEnv.setupEndpoint(endpointName, recordEndpoint)
+
+      startOffset.partitionNums.map {
+        case (part, index) =>
+          LowLatencyMemoryStreamInputPartition(
+            endpointName,
+            endpointRef.address,
+            part,
+            index,
+            Int.MaxValue
+          )
+      }.toArray
+    }
+  }
+
+  override def planInputPartitions(start: OffsetV2, end: OffsetV2): Array[InputPartition] = {
+    val startOffset = start.asInstanceOf[LowLatencyMemoryStreamOffset]
+    val endOffset = end.asInstanceOf[LowLatencyMemoryStreamOffset]
+    synchronized {
+      val endpointName = s"ContinuousRecordEndpoint-${java.util.UUID.randomUUID()}-$id"

Review Comment:
   ditto



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433343385


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation

Review Comment:
   Is this comment supposed to be here? Seems unrelated to the code.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433377244


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation
+
+  @GuardedBy("this")
+  private val records = Seq.fill(numPartitions)(new ListBuffer[UnsafeRow])
+
+  private val recordEndpoint = new ContinuousRecordEndpoint(records, this)
+  @volatile private var endpointRef: RpcEndpointRef = _
+
+  override def addData(data: IterableOnce[A]): Offset = synchronized {
+    // Distribute data evenly among partition lists.
+    data.iterator.to(Seq).zipWithIndex.map {
+      case (item, index) =>
+        records(index % numPartitions) += toRow(item).copy().asInstanceOf[UnsafeRow]
+    }
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  def addData(partitionId: Int, data: IterableOnce[A]): Offset = synchronized {
+    require(
+      partitionId >= 0 && partitionId < numPartitions,
+      s"Partition ID $partitionId is out of bounds for $numPartitions partitions."
+    )
+
+    // Add data to the specified partition.
+    records(partitionId) ++= data.iterator.map(item => toRow(item).copy().asInstanceOf[UnsafeRow])
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def initialOffset(): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, 0)).toMap)
+  }
+
+  override def latestOffset(startOffset: OffsetV2, limit: ReadLimit): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)

Review Comment:
   Do we also need to add `synchronized` to `latestOffset` as it also access `records`? Although it just reads, but the `size` might be inconsistent. 



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["jerrypeng (via GitHub)" <gi...@apache.org>]

jerrypeng commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433499579


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation
+
+  @GuardedBy("this")
+  private val records = Seq.fill(numPartitions)(new ListBuffer[UnsafeRow])
+
+  private val recordEndpoint = new ContinuousRecordEndpoint(records, this)
+  @volatile private var endpointRef: RpcEndpointRef = _
+
+  override def addData(data: IterableOnce[A]): Offset = synchronized {
+    // Distribute data evenly among partition lists.
+    data.iterator.to(Seq).zipWithIndex.map {
+      case (item, index) =>
+        records(index % numPartitions) += toRow(item).copy().asInstanceOf[UnsafeRow]
+    }
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  def addData(partitionId: Int, data: IterableOnce[A]): Offset = synchronized {
+    require(
+      partitionId >= 0 && partitionId < numPartitions,
+      s"Partition ID $partitionId is out of bounds for $numPartitions partitions."
+    )
+
+    // Add data to the specified partition.
+    records(partitionId) ++= data.iterator.map(item => toRow(item).copy().asInstanceOf[UnsafeRow])
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def initialOffset(): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, 0)).toMap)
+  }
+
+  override def latestOffset(startOffset: OffsetV2, limit: ReadLimit): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)

Review Comment:
   This is an interesting point. For RTM, offset returned from latestOffset is actually not used.  The offset returned from latestOffset defines the end offset of a batch for non-rtm streaming queries.  In RTM, the end offset of a batch is calculated when the batch finishes.  However, this source also support non-RTM queries.  Though in streaming we typically use StreamTest framework that executes test actions and batches in synchronized steps so any race should not happen.  Though for best practices I will add synchronized to the method.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2431355580


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation
+
+  @GuardedBy("this")
+  private val records = Seq.fill(numPartitions)(new ListBuffer[UnsafeRow])
+
+  private val recordEndpoint = new ContinuousRecordEndpoint(records, this)
+  @volatile private var endpointRef: RpcEndpointRef = _
+
+  override def addData(data: IterableOnce[A]): Offset = synchronized {
+    // Distribute data evenly among partition lists.
+    data.iterator.to(Seq).zipWithIndex.map {
+      case (item, index) =>
+        records(index % numPartitions) += toRow(item).copy().asInstanceOf[UnsafeRow]
+    }
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  def addData(partitionId: Int, data: IterableOnce[A]): Offset = synchronized {
+    require(
+      partitionId >= 0 && partitionId < numPartitions,
+      s"Partition ID $partitionId is out of bounds for $numPartitions partitions."
+    )
+
+    // Add data to the specified partition.
+    records(partitionId) ++= data.iterator.map(item => toRow(item).copy().asInstanceOf[UnsafeRow])
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def initialOffset(): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, 0)).toMap)
+  }
+
+  override def latestOffset(startOffset: OffsetV2, limit: ReadLimit): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def deserializeOffset(json: String): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(Serialization.read[Map[Int, Int]](json))
+  }
+
+  override def mergeOffsets(offsets: Array[PartitionOffset]): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(
+      offsets.map {
+        case ContinuousRecordPartitionOffset(part, num) => (part, num)
+      }.toMap
+    )
+  }
+
+  override def planInputPartitions(start: OffsetV2): Array[InputPartition] = {
+    val startOffset = start.asInstanceOf[LowLatencyMemoryStreamOffset]
+    synchronized {
+      val endpointName = s"ContinuousRecordEndpoint-${java.util.UUID.randomUUID()}-$id"

Review Comment:
   RealTimeRecordEndpoint? Or LowLatencyRecordEndpoint? As this is not for ContinuousStream, ContinuousRecordEndpoint looks a bit confused.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["jerrypeng (via GitHub)" <gi...@apache.org>]

jerrypeng commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2433350667


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/sources/LowLatencyMemoryStream.scala:
##########
@@ -0,0 +1,300 @@
+/*
+ * 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.sql.execution.streaming
+
+import java.util.concurrent.atomic.AtomicInteger
+import javax.annotation.concurrent.GuardedBy
+
+import scala.collection.mutable.ListBuffer
+
+import org.json4s.{Formats, NoTypeHints}
+import org.json4s.jackson.Serialization
+
+import org.apache.spark.{SparkEnv, TaskContext}
+import org.apache.spark.rpc.{RpcAddress, RpcEndpointRef}
+import org.apache.spark.sql.{Encoder, SQLContext}
+import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.expressions.UnsafeRow
+import org.apache.spark.sql.connector.read.InputPartition
+import org.apache.spark.sql.connector.read.PartitionReader
+import org.apache.spark.sql.connector.read.PartitionReaderFactory
+import org.apache.spark.sql.connector.read.streaming.{
+  Offset => OffsetV2,
+  PartitionOffset,
+  ReadLimit,
+  SupportsRealTimeMode,
+  SupportsRealTimeRead
+}
+import org.apache.spark.sql.connector.read.streaming.SupportsRealTimeRead.RecordStatus
+import org.apache.spark.sql.execution.datasources.v2.LowLatencyClock
+import org.apache.spark.sql.execution.streaming.runtime._
+import org.apache.spark.util.{Clock, RpcUtils}
+
+/**
+ * A low latency memory source from memory, only for unit test purpose.
+ * This class is very similar to ContinuousMemoryStream, except that it implements the
+ * interface of SupportsRealTimeMode, rather than ContinuousStream
+ * The overall strategy here is:
+ *  * LowLatencyMemoryStream maintains a list of records for each partition. addData() will
+ *    distribute records evenly-ish across partitions.
+ *  * RecordEndpoint is set up as an endpoint for executor-side
+ *    LowLatencyMemoryStreamInputPartitionReader instances to poll. It returns the record at
+ *    the specified offset within the list, or null if that offset doesn't yet have a record.
+ */
+class LowLatencyMemoryStream[A: Encoder](
+    id: Int,
+    sqlContext: SQLContext,
+    numPartitions: Int = 2,
+    clock: Clock = LowLatencyClock.getClock)
+    extends MemoryStreamBaseClass[A](0, sqlContext)
+    with SupportsRealTimeMode {
+  private implicit val formats: Formats = Serialization.formats(NoTypeHints)
+
+  // LowLatencyReader implementation
+
+  @GuardedBy("this")
+  private val records = Seq.fill(numPartitions)(new ListBuffer[UnsafeRow])
+
+  private val recordEndpoint = new ContinuousRecordEndpoint(records, this)
+  @volatile private var endpointRef: RpcEndpointRef = _
+
+  override def addData(data: IterableOnce[A]): Offset = synchronized {
+    // Distribute data evenly among partition lists.
+    data.iterator.to(Seq).zipWithIndex.map {
+      case (item, index) =>
+        records(index % numPartitions) += toRow(item).copy().asInstanceOf[UnsafeRow]
+    }
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  def addData(partitionId: Int, data: IterableOnce[A]): Offset = synchronized {
+    require(
+      partitionId >= 0 && partitionId < numPartitions,
+      s"Partition ID $partitionId is out of bounds for $numPartitions partitions."
+    )
+
+    // Add data to the specified partition.
+    records(partitionId) ++= data.iterator.map(item => toRow(item).copy().asInstanceOf[UnsafeRow])
+
+    // The new target offset is the offset where all records in all partitions have been processed.
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def initialOffset(): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, 0)).toMap)
+  }
+
+  override def latestOffset(startOffset: OffsetV2, limit: ReadLimit): OffsetV2 = {
+    LowLatencyMemoryStreamOffset((0 until numPartitions).map(i => (i, records(i).size)).toMap)
+  }
+
+  override def deserializeOffset(json: String): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(Serialization.read[Map[Int, Int]](json))
+  }
+
+  override def mergeOffsets(offsets: Array[PartitionOffset]): LowLatencyMemoryStreamOffset = {
+    LowLatencyMemoryStreamOffset(
+      offsets.map {
+        case ContinuousRecordPartitionOffset(part, num) => (part, num)
+      }.toMap
+    )
+  }
+
+  override def planInputPartitions(start: OffsetV2): Array[InputPartition] = {
+    val startOffset = start.asInstanceOf[LowLatencyMemoryStreamOffset]
+    synchronized {
+      val endpointName = s"ContinuousRecordEndpoint-${java.util.UUID.randomUUID()}-$id"

Review Comment:
   I will rename but the reason is the ContinuousRecordEndpoint is borrowed from "ContinuousMemoryStream".



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2431305012


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/v2/RealTimeStreamScanExec.scala:
##########
@@ -0,0 +1,39 @@
+/*
+ * 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.sql.execution.datasources.v2
+
+import org.apache.spark.util.{Clock, SystemClock}
+
+/* The singleton object to control the time in testing */
+object LowLatencyClock {

Review Comment:
   Do you mean you will add the code of RealTimeStreamScanExec in other PR?



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

Re: [PR] [SPARK-53785][SS] Memory Source for RTM [spark]

["viirya (via GitHub)" <gi...@apache.org>]

viirya commented on code in PR #52502:
URL: https://github.com/apache/spark/pull/52502#discussion_r2431311480


##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/runtime/memory.scala:
##########
@@ -298,7 +312,6 @@ case class MemoryStream[A : Encoder](
   }
 }
 
-

Review Comment:
   unnecessary change.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org