Examples of BulkIteration

Examples of org.apache.flink.api.java.record.operators.BulkIteration

68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
 
  private static Plan getTestPlanPlan(int numSubTasks, String input, String output) {

    FileDataSource initialInput = new FileDataSource(TextInputFormat.class, input, "input");
   
    BulkIteration iteration = new BulkIteration("Loop");
    iteration.setInput(initialInput);
    iteration.setMaximumNumberOfIterations(5);
   
    Assert.assertTrue(iteration.getMaximumNumberOfIterations() > 1);

    ReduceOperator sumReduce = ReduceOperator.builder(new PickOneReducer())
        .input(iteration.getPartialSolution())
        .name("Compute sum (Reduce)")
        .build();
   
    iteration.setNextPartialSolution(sumReduce);

    FileDataSink finalResult = new FileDataSink(CsvOutputFormat.class, output, iteration, "Output");
    CsvOutputFormat.configureRecordFormat(finalResult)
      .recordDelimiter('\n')
      .fieldDelimiter(' ')
View Full Code Here

Examples of org.apache.flink.api.java.record.operators.BulkIteration

54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
    // create DataSourceContract for cluster center input
    FileDataSource initialClusterPoints = new FileDataSource(new PointInFormat(), clusterInput, "Centers");
    initialClusterPoints.setDegreeOfParallelism(1);
   
    BulkIteration iteration = new BulkIteration("K-Means Loop");
    iteration.setInput(initialClusterPoints);
    iteration.setMaximumNumberOfIterations(numIterations);
   
    // create DataSourceContract for data point input
    FileDataSource dataPoints = new FileDataSource(new PointInFormat(), dataPointInput, "Data Points");

    // create CrossOperator for distance computation
    CrossOperator computeDistance = CrossOperator.builder(new ComputeDistance())
        .input1(dataPoints)
        .input2(iteration.getPartialSolution())
        .name("Compute Distances")
        .build();

    // create ReduceOperator for finding the nearest cluster centers
    ReduceOperator findNearestClusterCenters = ReduceOperator.builder(new FindNearestCenter(), IntValue.class, 0)
        .input(computeDistance)
        .name("Find Nearest Centers")
        .build();

    // create ReduceOperator for computing new cluster positions
    ReduceOperator recomputeClusterCenter = ReduceOperator.builder(new RecomputeClusterCenter(), IntValue.class, 0)
        .input(findNearestClusterCenters)
        .name("Recompute Center Positions")
        .build();
    iteration.setNextPartialSolution(recomputeClusterCenter);
   
    // create DataSourceContract for data point input
    FileDataSource dataPoints2 = new FileDataSource(new PointInFormat(), dataPointInput, "Data Points 2");
   
    // compute distance of points to final clusters
View Full Code Here

Examples of org.apache.flink.api.java.record.operators.BulkIteration

72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
   
    MapOperator clusterPoints = MapOperator.builder(new PointBuilder()).name("Build cluster points").input(clustersSource).build();
   
    // ---------------------- Begin K-Means Loop ---------------------
   
    BulkIteration iter = new BulkIteration("k-means loop");
    iter.setInput(clusterPoints);
    iter.setMaximumNumberOfIterations(numIterations);

    // compute the distances and select the closest center
    MapOperator findNearestClusterCenters = MapOperator.builder(new SelectNearestCenter())
      .setBroadcastVariable("centers", iter.getPartialSolution())
      .input(dataPoints)
      .name("Find Nearest Centers")
      .build();

    // computing the new cluster positions
    ReduceOperator recomputeClusterCenter = ReduceOperator.builder(new RecomputeClusterCenter(), IntValue.class, 0)
      .input(findNearestClusterCenters)
      .name("Recompute Center Positions")
      .build();
   
    iter.setNextPartialSolution(recomputeClusterCenter);

    // ---------------------- End K-Means Loop ---------------------
   
    // create DataSinkContract for writing the new cluster positions
    FileDataSink newClusterPoints = new FileDataSink(new PointOutFormat(), output, iter, "New Center Positions");
View Full Code Here

Examples of org.apache.flink.api.java.record.operators.BulkIteration

646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
 
  @Test
  public void testBranchAfterIteration() {
    FileDataSource sourceA = new FileDataSource(DummyInputFormat.class, IN_FILE, "Source 2");
   
    BulkIteration iteration = new BulkIteration("Loop");
    iteration.setInput(sourceA);
    iteration.setMaximumNumberOfIterations(10);
   
    MapOperator mapper = MapOperator.builder(IdentityMap.class).name("Mapper").input(iteration.getPartialSolution()).build();
    iteration.setNextPartialSolution(mapper);
   
    FileDataSink sink1 = new FileDataSink(DummyOutputFormat.class, OUT_FILE, iteration, "Sink 1");
   
    MapOperator postMap = MapOperator.builder(IdentityMap.class).name("Post Iteration Mapper")
        .input(iteration).build();
View Full Code Here

Examples of org.apache.flink.api.java.record.operators.BulkIteration

680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
  @Test
  public void testBranchBeforeIteration() {
    FileDataSource source1 = new FileDataSource(DummyInputFormat.class, IN_FILE, "Source 1");
    FileDataSource source2 = new FileDataSource(DummyInputFormat.class, IN_FILE, "Source 2");
   
    BulkIteration iteration = new BulkIteration("Loop");
    iteration.setInput(source2);
    iteration.setMaximumNumberOfIterations(10);
   
    MapOperator inMap = MapOperator.builder(new IdentityMap())
                               .input(source1)
                               .name("In Iteration Map")
                               .setBroadcastVariable("BC", iteration.getPartialSolution())
                               .build();
   
    iteration.setNextPartialSolution(inMap);
   
    MapOperator postMap = MapOperator.builder(new IdentityMap())
                     .input(source1)
                     .name("Post Iteration Map")
                     .setBroadcastVariable("BC", iteration)
View Full Code Here

Examples of org.apache.flink.api.java.record.operators.BulkIteration

731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
    FileDataSource sourceB = new FileDataSource(DummyInputFormat.class, IN_FILE, "Source 2");

    FileDataSink sink1 = new FileDataSink(DummyOutputFormat.class, OUT_FILE, sourceA, "Sink 1");
    FileDataSink sink2 = new FileDataSink(DummyOutputFormat.class, OUT_FILE, sourceB, "Sink 2");

    BulkIteration iteration = new BulkIteration("Loop");
    iteration.setInput(sourceA);
    iteration.setMaximumNumberOfIterations(10);

    CrossOperator stepFunction = CrossOperator.builder(DummyCrossStub.class).name("StepFunction").
        input1(iteration.getPartialSolution()).
        input2(sourceB).
        build();

    iteration.setNextPartialSolution(stepFunction);

    FileDataSink sink3 = new FileDataSink(DummyOutputFormat.class, OUT_FILE, iteration, "Sink 3");

    List<FileDataSink> sinks = new ArrayList<FileDataSink>();
    sinks.add(sink1);
View Full Code Here

Examples of org.apache.flink.api.java.record.operators.BulkIteration

909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
    ReduceOperator reducedSource = ReduceOperator.builder(IdentityReduce.class).
        input(source).
        name("Identity reduce source").
        build();

    BulkIteration iteration = new BulkIteration("Loop");
    iteration.setInput(mappedSource);
    iteration.setMaximumNumberOfIterations(10);

    JoinOperator nextPartialSolution = JoinOperator.builder(DummyMatchStub.class, IntValue.class, 0,0).
        input1(iteration.getPartialSolution()).
        input2(reducedSource).
        name("Next partial solution").
        build();

    iteration.setNextPartialSolution(nextPartialSolution);

    FileDataSink sink = new FileDataSink(DummyOutputFormat.class, OUT_FILE, iteration, "Iteration sink");
    List<FileDataSink> sinks = new ArrayList<FileDataSink>();
    sinks.add(sink);
View Full Code Here
TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.