blob: d06c1b5d33cb8f522b91d9ce71dd64f22f8321e0 [file] [log] [blame]
/*
* 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.wayang.java.operators.graph;
import java.util.HashMap;
import java.util.Map;
import java.util.function.BiFunction;
import org.apache.wayang.basic.data.Tuple2;
import org.apache.wayang.basic.operators.PageRankOperator;
import org.apache.wayang.core.optimizer.OptimizationContext;
import org.apache.wayang.core.plan.wayangplan.ExecutionOperator;
import org.apache.wayang.core.platform.ChannelDescriptor;
import org.apache.wayang.core.platform.ChannelInstance;
import org.apache.wayang.core.platform.lineage.ExecutionLineageNode;
import org.apache.wayang.core.util.Tuple;
import org.apache.wayang.java.channels.CollectionChannel;
import org.apache.wayang.java.channels.StreamChannel;
import org.apache.wayang.java.execution.JavaExecutor;
import org.apache.wayang.java.operators.JavaExecutionOperator;
import java.util.Collection;
import java.util.Collections;
import java.util.Iterator;
import java.util.List;
import java.util.Spliterators;
import java.util.stream.Stream;
import java.util.stream.StreamSupport;
/**
* Java implementation of the {@link PageRankOperator}.
*/
public class JavaPageRankOperator extends PageRankOperator implements JavaExecutionOperator {
public JavaPageRankOperator(int numIterations) {
super(numIterations);
}
public JavaPageRankOperator(PageRankOperator that) {
super(that);
}
@Override
public Tuple<Collection<ExecutionLineageNode>, Collection<ChannelInstance>> evaluate(
ChannelInstance[] inputs,
ChannelInstance[] outputs,
JavaExecutor javaExecutor,
OptimizationContext.OperatorContext operatorContext) {
CollectionChannel.Instance input = (CollectionChannel.Instance) inputs[0];
StreamChannel.Instance output = (StreamChannel.Instance) outputs[0];
final Collection<Tuple2<Long, Long>> edges = input.provideCollection();
final Map<Long, Float> pageRanks = this.pageRank(edges);
final Stream<Tuple2<Long, Float>> pageRankStream = pageRanks.entrySet().stream().map(entry -> new Tuple2<>(entry.getKey(), entry.getValue()));
output.accept(pageRankStream);
return ExecutionOperator.modelQuasiEagerExecution(inputs, outputs, operatorContext);
}
/**
* Execute the PageRank algorithm.
*
* @param edgeDataSet edges of a graph
* @return the page ranks
*/
//TODO: change for efficient map
private Map<Long, Float> pageRank(Collection<Tuple2<Long, Long>> edgeDataSet) {
// Get the degress of all vertices and make sure we collect *all* vertices.
//TODO: change for efficient map
HashMap<Long, Integer> degrees = new HashMap<>();
for (Tuple2<Long, Long> edge : edgeDataSet) {
this.adjustOrPutValue(degrees, edge.field0, 1, 1, Integer::sum);
this.adjustOrPutValue(degrees, edge.field0, 0, 0, Integer::sum);
}
int numVertices = degrees.size();
float initialRank = 1f / numVertices;
float dampingRank = (1 - this.dampingFactor) / numVertices;
// Initialize the rank map.
//TODO: change for efficient map
HashMap<Long, Float> initialRanks = new HashMap<>();
degrees.forEach( (k, v) -> {
initialRanks.putIfAbsent(k, initialRank);
});
HashMap<Long, Float> currentRanks = initialRanks;
for (int iteration = 0; iteration < this.getNumIterations(); iteration++) {
// Add the damping first.
//TODO: change for efficient map
HashMap<Long, Float> newRanks = new HashMap<Long, Float>(currentRanks.size());
degrees.forEach( (k, v) -> {
newRanks.putIfAbsent(k, dampingRank);
});
// Now add the other ranks.
for (Tuple2<Long, Long> edge : edgeDataSet) {
final long sourceVertex = edge.field0;
final long targetVertex = edge.field1;
final int degree = degrees.get(sourceVertex);
final float currentRank = currentRanks.get(sourceVertex);
final float partialRank = this.dampingFactor * currentRank / degree;
this.adjustOrPutValue(newRanks, targetVertex, partialRank, partialRank, Float::sum);
}
currentRanks = newRanks;
}
return currentRanks;
}
/**
* simulate the process on the Trove4j library
* @param key key to modify on the map
* @param default_value default value in the case of not key
* @param correction element to add the array in the case of the key exist
*/
private <T> void adjustOrPutValue(Map<Long, T> map, Long key, T default_value, T correction, BiFunction<T, T, T> update){
if(map.containsKey(key)){
T value = map.get(key);
map.replace(key, update.apply(value, correction) );
}else{
map.put(key, default_value);
}
}
@Override
public String getLoadProfileEstimatorConfigurationKey() {
return "wayang.java.pagerank.load";
}
@Override
public List<ChannelDescriptor> getSupportedInputChannels(int index) {
assert index == 0;
return Collections.singletonList(CollectionChannel.DESCRIPTOR);
}
@Override
public List<ChannelDescriptor> getSupportedOutputChannels(int index) {
assert index == 0;
return Collections.singletonList(StreamChannel.DESCRIPTOR);
}
}