Examples with Edge org.apache.flink.graph.Edge used on opensource projects

Example 1 with Edge

use of org.apache.flink.graph.Edge in project flink by apache.

the class SimplifyTest method setup.

@Before
public void setup() {
    ExecutionEnvironment env = ExecutionEnvironment.createCollectionsEnvironment();
    Object[][] edges = new Object[][] { new Object[] { 0, 0 }, new Object[] { 0, 1 }, new Object[] { 0, 1 }, new Object[] { 0, 2 }, new Object[] { 0, 2 }, new Object[] { 1, 0 }, new Object[] { 2, 2 } };
    List<Edge<IntValue, NullValue>> edgeList = new LinkedList<>();
    for (Object[] edge : edges) {
        edgeList.add(new Edge<>(new IntValue((int) edge[0]), new IntValue((int) edge[1]), NullValue.getInstance()));
    }
    graph = Graph.fromCollection(edgeList, env);
}

Example 2 with Edge

use of org.apache.flink.graph.Edge in project flink by apache.

the class ConnectedComponents method main.

@SuppressWarnings("serial")
public static void main(String[] args) throws Exception {
    if (!parseParameters(args)) {
        return;
    }
    ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
    DataSet<Edge<Long, NullValue>> edges = getEdgesDataSet(env);
    Graph<Long, Long, NullValue> graph = Graph.fromDataSet(edges, new MapFunction<Long, Long>() {

        @Override
        public Long map(Long value) throws Exception {
            return value;
        }
    }, env);
    DataSet<Vertex<Long, Long>> verticesWithMinIds = graph.run(new GSAConnectedComponents<Long, Long, NullValue>(maxIterations));
    // emit result
    if (fileOutput) {
        verticesWithMinIds.writeAsCsv(outputPath, "\n", ",");
        // since file sinks are lazy, we trigger the execution explicitly
        env.execute("Connected Components Example");
    } else {
        verticesWithMinIds.print();
    }
}

Example 3 with Edge

use of org.apache.flink.graph.Edge in project flink by apache.

the class CompleteGraph method generate.

@Override
public Graph<LongValue, NullValue, NullValue> generate() {
    // Vertices
    DataSet<Vertex<LongValue, NullValue>> vertices = GraphGeneratorUtils.vertexSequence(env, parallelism, vertexCount);
    // Edges
    LongValueSequenceIterator iterator = new LongValueSequenceIterator(0, this.vertexCount - 1);
    DataSet<Edge<LongValue, NullValue>> edges = env.fromParallelCollection(iterator, LongValue.class).setParallelism(parallelism).name("Edge iterators").flatMap(new LinkVertexToAll(vertexCount)).setParallelism(parallelism).name("Complete graph edges");
    // Graph
    return Graph.fromDataSet(vertices, edges, env);
}

Example 4 with Edge

use of org.apache.flink.graph.Edge in project flink by apache.

the class PregelCompilerTest method testPregelCompilerWithBroadcastVariable.

@SuppressWarnings("serial")
@Test
public void testPregelCompilerWithBroadcastVariable() {
    final String broadcastSetName = "broadcast";
    ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
    env.setParallelism(DEFAULT_PARALLELISM);
    // compose test program
    {
        DataSet<Long> bcVar = env.fromElements(1L);
        DataSet<Vertex<Long, Long>> initialVertices = env.fromElements(new Tuple2<>(1L, 1L), new Tuple2<>(2L, 2L)).map(new Tuple2ToVertexMap<>());
        DataSet<Edge<Long, NullValue>> edges = env.fromElements(new Tuple2<>(1L, 2L)).map(new MapFunction<Tuple2<Long, Long>, Edge<Long, NullValue>>() {

            public Edge<Long, NullValue> map(Tuple2<Long, Long> edge) {
                return new Edge<>(edge.f0, edge.f1, NullValue.getInstance());
            }
        });
        Graph<Long, Long, NullValue> graph = Graph.fromDataSet(initialVertices, edges, env);
        VertexCentricConfiguration parameters = new VertexCentricConfiguration();
        parameters.addBroadcastSet(broadcastSetName, bcVar);
        DataSet<Vertex<Long, Long>> result = graph.runVertexCentricIteration(new CCCompute(), null, 100, parameters).getVertices();
        result.output(new DiscardingOutputFormat<>());
    }
    Plan p = env.createProgramPlan("Pregel Connected Components");
    OptimizedPlan op = compileNoStats(p);
    // check the sink
    SinkPlanNode sink = op.getDataSinks().iterator().next();
    assertEquals(ShipStrategyType.FORWARD, sink.getInput().getShipStrategy());
    assertEquals(DEFAULT_PARALLELISM, sink.getParallelism());
    // check the iteration
    WorksetIterationPlanNode iteration = (WorksetIterationPlanNode) sink.getInput().getSource();
    assertEquals(DEFAULT_PARALLELISM, iteration.getParallelism());
    // check the solution set delta
    PlanNode ssDelta = iteration.getSolutionSetDeltaPlanNode();
    assertTrue(ssDelta instanceof SingleInputPlanNode);
    SingleInputPlanNode ssFlatMap = (SingleInputPlanNode) ((SingleInputPlanNode) (ssDelta)).getInput().getSource();
    assertEquals(DEFAULT_PARALLELISM, ssFlatMap.getParallelism());
    assertEquals(ShipStrategyType.FORWARD, ssFlatMap.getInput().getShipStrategy());
    // check the computation coGroup
    DualInputPlanNode computationCoGroup = (DualInputPlanNode) (ssFlatMap.getInput().getSource());
    assertEquals(DEFAULT_PARALLELISM, computationCoGroup.getParallelism());
    assertEquals(ShipStrategyType.FORWARD, computationCoGroup.getInput1().getShipStrategy());
    assertEquals(ShipStrategyType.PARTITION_HASH, computationCoGroup.getInput2().getShipStrategy());
    assertTrue(computationCoGroup.getInput2().getTempMode().isCached());
    assertEquals(new FieldList(0), computationCoGroup.getInput2().getShipStrategyKeys());
    // check that the initial partitioning is pushed out of the loop
    assertEquals(ShipStrategyType.PARTITION_HASH, iteration.getInput1().getShipStrategy());
    assertEquals(new FieldList(0), iteration.getInput1().getShipStrategyKeys());
}

Example 5 with Edge

use of org.apache.flink.graph.Edge in project flink by apache.

the class SpargelCompilerTest method testSpargelCompilerWithBroadcastVariable.

@SuppressWarnings("serial")
@Test
public void testSpargelCompilerWithBroadcastVariable() {
    final String broadcastVariableName = "broadcast variable";
    ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
    env.setParallelism(DEFAULT_PARALLELISM);
    // compose test program
    DataSet<Long> bcVar = env.fromElements(1L);
    DataSet<Vertex<Long, Long>> initialVertices = env.fromElements(new Tuple2<>(1L, 1L), new Tuple2<>(2L, 2L)).map(new Tuple2ToVertexMap<>());
    DataSet<Edge<Long, NullValue>> edges = env.fromElements(new Tuple2<>(1L, 2L)).map(new MapFunction<Tuple2<Long, Long>, Edge<Long, NullValue>>() {

        public Edge<Long, NullValue> map(Tuple2<Long, Long> edge) {
            return new Edge<>(edge.f0, edge.f1, NullValue.getInstance());
        }
    });
    Graph<Long, Long, NullValue> graph = Graph.fromDataSet(initialVertices, edges, env);
    ScatterGatherConfiguration parameters = new ScatterGatherConfiguration();
    parameters.addBroadcastSetForScatterFunction(broadcastVariableName, bcVar);
    parameters.addBroadcastSetForGatherFunction(broadcastVariableName, bcVar);
    DataSet<Vertex<Long, Long>> result = graph.runScatterGatherIteration(new ConnectedComponents.CCMessenger<>(BasicTypeInfo.LONG_TYPE_INFO), new ConnectedComponents.CCUpdater<>(), 100).getVertices();
    result.output(new DiscardingOutputFormat<>());
    Plan p = env.createProgramPlan("Spargel Connected Components");
    OptimizedPlan op = compileNoStats(p);
    // check the sink
    SinkPlanNode sink = op.getDataSinks().iterator().next();
    assertEquals(ShipStrategyType.FORWARD, sink.getInput().getShipStrategy());
    assertEquals(DEFAULT_PARALLELISM, sink.getParallelism());
    // check the iteration
    WorksetIterationPlanNode iteration = (WorksetIterationPlanNode) sink.getInput().getSource();
    assertEquals(DEFAULT_PARALLELISM, iteration.getParallelism());
    // check the solution set join and the delta
    PlanNode ssDelta = iteration.getSolutionSetDeltaPlanNode();
    assertTrue(ssDelta instanceof // this is only true if the update functions preserves
    DualInputPlanNode);
    // the partitioning
    DualInputPlanNode ssJoin = (DualInputPlanNode) ssDelta;
    assertEquals(DEFAULT_PARALLELISM, ssJoin.getParallelism());
    assertEquals(ShipStrategyType.PARTITION_HASH, ssJoin.getInput1().getShipStrategy());
    assertEquals(new FieldList(0), ssJoin.getInput1().getShipStrategyKeys());
    // check the workset set join
    DualInputPlanNode edgeJoin = (DualInputPlanNode) ssJoin.getInput1().getSource();
    assertEquals(DEFAULT_PARALLELISM, edgeJoin.getParallelism());
    assertEquals(ShipStrategyType.PARTITION_HASH, edgeJoin.getInput1().getShipStrategy());
    assertEquals(ShipStrategyType.FORWARD, edgeJoin.getInput2().getShipStrategy());
    assertTrue(edgeJoin.getInput1().getTempMode().isCached());
    assertEquals(new FieldList(0), edgeJoin.getInput1().getShipStrategyKeys());
    // check that the initial partitioning is pushed out of the loop
    assertEquals(ShipStrategyType.PARTITION_HASH, iteration.getInput1().getShipStrategy());
    assertEquals(ShipStrategyType.PARTITION_HASH, iteration.getInput2().getShipStrategy());
    assertEquals(new FieldList(0), iteration.getInput1().getShipStrategyKeys());
    assertEquals(new FieldList(0), iteration.getInput2().getShipStrategyKeys());
}