Added map-based connector.
git-svn-id: https://hyracks.googlecode.com/svn/branches/hyracks_dev_next@1148 123451ca-8445-de46-9d55-352943316053
diff --git a/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/GlobalHashingLocalityMap.java b/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/GlobalHashingLocalityMap.java
new file mode 100644
index 0000000..f3086f0
--- /dev/null
+++ b/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/GlobalHashingLocalityMap.java
@@ -0,0 +1,50 @@
+/*
+ * Copyright 2009-2010 by The Regents of the University of California
+ * Licensed 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 from
+ *
+ * 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 edu.uci.ics.hyracks.dataflow.std.connectors;
+
+public class GlobalHashingLocalityMap implements ILocalityMap {
+
+ private static final long serialVersionUID = 1L;
+
+ private int[] consumers;
+
+ /* (non-Javadoc)
+ * @see edu.uci.ics.hyracks.examples.text.client.aggregation.helpers.ILocalityMap#getConsumers(int)
+ */
+ @Override
+ public int[] getConsumers(int senderID, int nConsumerPartitions) {
+ if (consumers == null) {
+ consumers = new int[nConsumerPartitions];
+ for (int i = 0; i < consumers.length; i++) {
+ consumers[i] = i;
+ }
+ }
+ return consumers;
+ }
+
+ /* (non-Javadoc)
+ * @see edu.uci.ics.hyracks.examples.text.client.aggregation.helpers.ILocalityMap#getConsumerPartitionCount()
+ */
+ @Override
+ public int getConsumerPartitionCount(int nConsumerPartitions) {
+ return nConsumerPartitions;
+ }
+
+ @Override
+ public boolean isConnected(int senderID, int receiverID, int nConsumerPartitions) {
+ return true;
+ }
+
+}
diff --git a/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/HashtableLocalityMap.java b/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/HashtableLocalityMap.java
new file mode 100644
index 0000000..a187a59
--- /dev/null
+++ b/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/HashtableLocalityMap.java
@@ -0,0 +1,72 @@
+/*
+ * Copyright 2009-2010 by The Regents of the University of California
+ * Licensed 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 from
+ *
+ * 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 edu.uci.ics.hyracks.dataflow.std.connectors;
+
+import java.util.BitSet;
+
+public class HashtableLocalityMap implements ILocalityMap {
+
+ private static final long serialVersionUID = 1L;
+
+ private final BitSet nodeMap;
+
+ public HashtableLocalityMap(BitSet nodeMap) {
+ this.nodeMap = nodeMap;
+ }
+
+ /*
+ * (non-Javadoc)
+ *
+ * @see
+ * edu.uci.ics.hyracks.examples.text.client.aggregation.helpers.ILocalityMap
+ * #getConsumers(int, int)
+ */
+ @Override
+ public int[] getConsumers(int senderID, int nConsumerPartitions) {
+ int consumersForSender = 0;
+ // Get the count of consumers
+ for (int i = senderID * nConsumerPartitions; i < (senderID + 1) * nConsumerPartitions; i++) {
+ if (nodeMap.get(i))
+ consumersForSender++;
+ }
+ int[] cons = new int[consumersForSender];
+ int consIdx = 0;
+ for (int i = senderID * nConsumerPartitions; i < (senderID + 1) * nConsumerPartitions; i++) {
+ if (nodeMap.get(i)) {
+ cons[consIdx] = i - senderID * nConsumerPartitions;
+ consIdx++;
+ }
+ }
+ return cons;
+ }
+
+ /*
+ * (non-Javadoc)
+ *
+ * @see
+ * edu.uci.ics.hyracks.examples.text.client.aggregation.helpers.ILocalityMap
+ * #getConsumerPartitionCount(int)
+ */
+ @Override
+ public int getConsumerPartitionCount(int nConsumerPartitions) {
+ return nConsumerPartitions;
+ }
+
+ @Override
+ public boolean isConnected(int senderID, int receiverID, int nConsumerPartitions) {
+ return nodeMap.get(senderID * nConsumerPartitions + receiverID);
+ }
+
+}
diff --git a/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/ILocalityMap.java b/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/ILocalityMap.java
new file mode 100644
index 0000000..b30971d
--- /dev/null
+++ b/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/ILocalityMap.java
@@ -0,0 +1,27 @@
+/*
+ * Copyright 2009-2010 by The Regents of the University of California
+ * Licensed 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 from
+ *
+ * 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 edu.uci.ics.hyracks.dataflow.std.connectors;
+
+import java.io.Serializable;
+
+public interface ILocalityMap extends Serializable {
+
+ public int[] getConsumers(int senderID, int nConsumerPartitions);
+
+ public boolean isConnected(int senderID, int receiverID, int nConsumerPartitions);
+
+ public int getConsumerPartitionCount(int nConsumerPartitions);
+
+}
diff --git a/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/LocalityAwareMToNPartitioningConnectorDescriptor.java b/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/LocalityAwareMToNPartitioningConnectorDescriptor.java
new file mode 100644
index 0000000..7d92150
--- /dev/null
+++ b/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/LocalityAwareMToNPartitioningConnectorDescriptor.java
@@ -0,0 +1,82 @@
+/*
+ * Copyright 2009-2010 by The Regents of the University of California
+ * Licensed 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 from
+ *
+ * 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 edu.uci.ics.hyracks.dataflow.std.connectors;
+
+import java.util.BitSet;
+
+import edu.uci.ics.hyracks.api.comm.IFrameWriter;
+import edu.uci.ics.hyracks.api.comm.IPartitionCollector;
+import edu.uci.ics.hyracks.api.comm.IPartitionWriterFactory;
+import edu.uci.ics.hyracks.api.context.IHyracksTaskContext;
+import edu.uci.ics.hyracks.api.dataflow.value.ITuplePartitionComputerFactory;
+import edu.uci.ics.hyracks.api.dataflow.value.RecordDescriptor;
+import edu.uci.ics.hyracks.api.exceptions.HyracksDataException;
+import edu.uci.ics.hyracks.api.job.JobSpecification;
+import edu.uci.ics.hyracks.dataflow.std.base.AbstractMToNConnectorDescriptor;
+import edu.uci.ics.hyracks.dataflow.std.collectors.NonDeterministicPartitionCollector;
+
+public class LocalityAwareMToNPartitioningConnectorDescriptor extends AbstractMToNConnectorDescriptor {
+
+ private static final long serialVersionUID = 1L;
+
+ private ILocalityMap localityMap;
+
+ private ITuplePartitionComputerFactory tpcf;
+
+ public LocalityAwareMToNPartitioningConnectorDescriptor(JobSpecification spec, ITuplePartitionComputerFactory tpcf,
+ ILocalityMap localityMap) {
+ super(spec);
+ this.localityMap = localityMap;
+ this.tpcf = tpcf;
+ }
+
+ /*
+ * (non-Javadoc)
+ *
+ * @see
+ * edu.uci.ics.hyracks.api.dataflow.IConnectorDescriptor#createPartitioner
+ * (edu.uci.ics.hyracks.api.context.IHyracksTaskContext,
+ * edu.uci.ics.hyracks.api.dataflow.value.RecordDescriptor,
+ * edu.uci.ics.hyracks.api.comm.IPartitionWriterFactory, int, int, int)
+ */
+ @Override
+ public IFrameWriter createPartitioner(IHyracksTaskContext ctx, RecordDescriptor recordDesc,
+ IPartitionWriterFactory edwFactory, int index, int nProducerPartitions, int nConsumerPartitions)
+ throws HyracksDataException {
+ return new LocalityAwarePartitionDataWriter(ctx, edwFactory, recordDesc, tpcf.createPartitioner(),
+ nConsumerPartitions, localityMap, index);
+ }
+
+ /*
+ * (non-Javadoc)
+ *
+ * @see edu.uci.ics.hyracks.api.dataflow.IConnectorDescriptor#
+ * createPartitionCollector
+ * (edu.uci.ics.hyracks.api.context.IHyracksTaskContext,
+ * edu.uci.ics.hyracks.api.dataflow.value.RecordDescriptor, int, int, int)
+ */
+ @Override
+ public IPartitionCollector createPartitionCollector(IHyracksTaskContext ctx, RecordDescriptor recordDesc,
+ int receiverIndex, int nProducerPartitions, int nConsumerPartitions) throws HyracksDataException {
+ BitSet expectedPartitions = new BitSet(nProducerPartitions);
+ for (int i = 0; i < nProducerPartitions; i++) {
+ if (localityMap.isConnected(i, receiverIndex, nConsumerPartitions))
+ expectedPartitions.set(i);
+ }
+ return new NonDeterministicPartitionCollector(ctx, getConnectorId(), receiverIndex, nProducerPartitions,
+ expectedPartitions);
+ }
+
+}
diff --git a/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/LocalityAwarePartitionDataWriter.java b/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/LocalityAwarePartitionDataWriter.java
new file mode 100644
index 0000000..6ec9013
--- /dev/null
+++ b/hyracks-dataflow-std/src/main/java/edu/uci/ics/hyracks/dataflow/std/connectors/LocalityAwarePartitionDataWriter.java
@@ -0,0 +1,125 @@
+/*
+ * Copyright 2009-2010 by The Regents of the University of California
+ * Licensed 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 from
+ *
+ * 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 edu.uci.ics.hyracks.dataflow.std.connectors;
+
+import java.io.IOException;
+import java.nio.ByteBuffer;
+
+import edu.uci.ics.hyracks.api.comm.IFrameWriter;
+import edu.uci.ics.hyracks.api.comm.IPartitionWriterFactory;
+import edu.uci.ics.hyracks.api.context.IHyracksTaskContext;
+import edu.uci.ics.hyracks.api.dataflow.value.ITuplePartitionComputer;
+import edu.uci.ics.hyracks.api.dataflow.value.RecordDescriptor;
+import edu.uci.ics.hyracks.api.exceptions.HyracksDataException;
+import edu.uci.ics.hyracks.dataflow.common.comm.io.FrameTupleAccessor;
+import edu.uci.ics.hyracks.dataflow.common.comm.io.FrameTupleAppender;
+
+public class LocalityAwarePartitionDataWriter implements IFrameWriter {
+
+ private final IFrameWriter[] pWriters;
+ private final FrameTupleAppender[] appenders;
+ private final FrameTupleAccessor tupleAccessor;
+ private final ITuplePartitionComputer tpc;
+
+ public LocalityAwarePartitionDataWriter(IHyracksTaskContext ctx, IPartitionWriterFactory pwFactory,
+ RecordDescriptor recordDescriptor, ITuplePartitionComputer tpc, int nConsumerPartitions,
+ ILocalityMap localityMap, int senderIndex) throws HyracksDataException {
+ int[] consumerPartitions = localityMap.getConsumers(senderIndex, nConsumerPartitions);
+ pWriters = new IFrameWriter[consumerPartitions.length];
+ appenders = new FrameTupleAppender[consumerPartitions.length];
+ for (int i = 0; i < consumerPartitions.length; ++i) {
+ try {
+ pWriters[i] = pwFactory.createFrameWriter(consumerPartitions[i]);
+ appenders[i] = new FrameTupleAppender(ctx.getFrameSize());
+ appenders[i].reset(ctx.allocateFrame(), true);
+ } catch (IOException e) {
+ throw new HyracksDataException(e);
+ }
+ }
+ tupleAccessor = new FrameTupleAccessor(ctx.getFrameSize(), recordDescriptor);
+ this.tpc = tpc;
+ }
+
+ /*
+ * (non-Javadoc)
+ *
+ * @see edu.uci.ics.hyracks.api.comm.IFrameWriter#open()
+ */
+ @Override
+ public void open() throws HyracksDataException {
+ for (int i = 0; i < pWriters.length; ++i) {
+ pWriters[i].open();
+ appenders[i].reset(appenders[i].getBuffer(), true);
+ }
+ }
+
+ /*
+ * (non-Javadoc)
+ *
+ * @see
+ * edu.uci.ics.hyracks.api.comm.IFrameWriter#nextFrame(java.nio.ByteBuffer)
+ */
+ @Override
+ public void nextFrame(ByteBuffer buffer) throws HyracksDataException {
+ tupleAccessor.reset(buffer);
+ int tupleCount = tupleAccessor.getTupleCount();
+ for (int i = 0; i < tupleCount; ++i) {
+ int h = pWriters.length == 1 ? 0 : tpc.partition(tupleAccessor, i, pWriters.length);
+ FrameTupleAppender appender = appenders[h];
+ if (!appender.append(tupleAccessor, i)) {
+ ByteBuffer appenderBuffer = appender.getBuffer();
+ flushFrame(appenderBuffer, pWriters[h]);
+ appender.reset(appenderBuffer, true);
+ if (!appender.append(tupleAccessor, i)) {
+ throw new IllegalStateException();
+ }
+ }
+ }
+ }
+
+ /*
+ * (non-Javadoc)
+ *
+ * @see edu.uci.ics.hyracks.api.comm.IFrameWriter#fail()
+ */
+ @Override
+ public void fail() throws HyracksDataException {
+ for (int i = 0; i < appenders.length; ++i) {
+ pWriters[i].fail();
+ }
+ }
+
+ private void flushFrame(ByteBuffer buffer, IFrameWriter frameWriter) throws HyracksDataException {
+ buffer.position(0);
+ buffer.limit(buffer.capacity());
+ frameWriter.nextFrame(buffer);
+ }
+
+ /*
+ * (non-Javadoc)
+ *
+ * @see edu.uci.ics.hyracks.api.comm.IFrameWriter#close()
+ */
+ @Override
+ public void close() throws HyracksDataException {
+ for (int i = 0; i < pWriters.length; ++i) {
+ if (appenders[i].getTupleCount() > 0) {
+ flushFrame(appenders[i].getBuffer(), pWriters[i]);
+ }
+ pWriters[i].close();
+ }
+ }
+
+}