Examples of BulkIteration

Examples of eu.stratosphere.api.java.record.operators.BulkIteration

74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
 
  private static Plan getPlan(int numSubTasks, String input, String output) {
    FileDataSource initialInput = new FileDataSource(new PointInFormat(), input, "Input");
    initialInput.setDegreeOfParallelism(1);
   
    BulkIteration iteration = new BulkIteration("Loop");
    iteration.setInput(initialInput);
    iteration.setMaximumNumberOfIterations(2);

    @SuppressWarnings("unchecked")
    MapOperator map2 = MapOperator.builder(new IdentityMapper()).input(iteration.getPartialSolution(), iteration.getPartialSolution()).name("map").build();
   
    iteration.setNextPartialSolution(map2);

    FileDataSink finalResult = new FileDataSink(new PointOutFormat(), output, iteration, "Output");

    Plan plan = new Plan(finalResult, "Iteration with union test");
    plan.setDefaultParallelism(numSubTasks);
View Full Code Here

Examples of eu.stratosphere.api.java.record.operators.BulkIteration

48
49
50
51
52
53
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
    // 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 eu.stratosphere.api.java.record.operators.BulkIteration

69
70
71
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
   
    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 eu.stratosphere.api.java.record.operators.BulkIteration

85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
 
  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(NUM_ITERATIONS);

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

    @SuppressWarnings("unchecked")
    FileDataSink finalResult = new FileDataSink(new CsvOutputFormat("\n",  " ", StringValue.class), output, iteration, "Output");

    Plan plan = new Plan(finalResult, "Iteration with AllReducer (keyless Reducer)");
View Full Code Here

Examples of eu.stratosphere.api.java.record.operators.BulkIteration

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
90
   
    FileDataSource pageWithRankInput = new FileDataSource(new DanglingPageRankInputFormat(),
      pageWithRankInputPath, "DanglingPageWithRankInput");
    pageWithRankInput.getParameters().setLong(DanglingPageRankInputFormat.NUM_VERTICES_PARAMETER, numVertices);
   
    BulkIteration iteration = new BulkIteration("Page Rank Loop");
    iteration.setInput(pageWithRankInput);
   
    FileDataSource adjacencyListInput = new FileDataSource(new ImprovedAdjacencyListInputFormat(),
      adjacencyListInputPath, "AdjancencyListInput");
   
    JoinOperator join = JoinOperator.builder(new DotProductMatch(), LongValue.class, 0, 0)
        .input1(iteration.getPartialSolution())
        .input2(adjacencyListInput)
        .name("Join with Edges")
        .build();
   
    CoGroupOperator rankAggregation = CoGroupOperator.builder(new DotProductCoGroup(), LongValue.class, 0, 0)
        .input1(iteration.getPartialSolution())
        .input2(join)
        .name("Rank Aggregation")
        .build();
    rankAggregation.getParameters().setLong(DotProductCoGroup.NUM_VERTICES_PARAMETER, numVertices);
    rankAggregation.getParameters().setLong(DotProductCoGroup.NUM_DANGLING_VERTICES_PARAMETER, numDanglingVertices);
   
    iteration.setNextPartialSolution(rankAggregation);
    iteration.setMaximumNumberOfIterations(numIterations);
    iteration.getAggregators().registerAggregationConvergenceCriterion(DotProductCoGroup.AGGREGATOR_NAME, PageRankStatsAggregator.class, DiffL1NormConvergenceCriterion.class);
   
    FileDataSink out = new FileDataSink(new PageWithRankOutFormat(), outputPath, iteration, "Final Ranks");

    Plan p = new Plan(out, "Dangling PageRank");
    p.setDefaultParallelism(dop);
View Full Code Here

Examples of eu.stratosphere.api.java.record.operators.BulkIteration

145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
   
    FileDataSource pageWithRankInput = new FileDataSource(new DanglingPageRankInputFormat(),
      pageWithRankInputPath, "PageWithRank Input");
    pageWithRankInput.getParameters().setLong(NUM_VERTICES_CONFIG_PARAM, numVertices);
   
    BulkIteration iteration = new BulkIteration("Page Rank Loop");
    iteration.setInput(pageWithRankInput);
   
    FileDataSource adjacencyListInput = new FileDataSource(new ImprovedAdjacencyListInputFormat(),
      adjacencyListInputPath, "AdjancencyListInput");
   
    JoinOperator join = JoinOperator.builder(new JoinVerexWithEdgesMatch(), LongValue.class, 0, 0)
        .input1(iteration.getPartialSolution())
        .input2(adjacencyListInput)
        .name("Join with Edges")
        .build();
   
    ReduceOperator rankAggregation = ReduceOperator.builder(new AggregatingReduce(), LongValue.class, 0)
        .input(join)
        .name("Rank Aggregation")
        .build();
   
    iteration.setNextPartialSolution(rankAggregation);
    iteration.setMaximumNumberOfIterations(numIterations);
   
    JoinOperator termination = JoinOperator.builder(new JoinOldAndNew(), LongValue.class, 0, 0)
        .input1(iteration.getPartialSolution())
        .input2(rankAggregation)
        .name("Join Old and New")
        .build();
   
    iteration.setTerminationCriterion(termination);
   
    FileDataSink out = new FileDataSink(new PageWithRankOutFormat(), outputPath, iteration, "Final Ranks");

    Plan p = new Plan(out, "Simple PageRank");
    p.setDefaultParallelism(dop);
View Full Code Here

Examples of eu.stratosphere.api.java.record.operators.BulkIteration

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
 
  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 SumReducer())
        .input(iteration.getPartialSolution())
        .name("Compute sum (Reduce)")
        .build();
   
    iteration.setNextPartialSolution(sumReduce);
   
    MapOperator terminationMapper = MapOperator.builder(new TerminationMapper())
        .input(sumReduce)
        .name("Compute termination criterion (Map)")
        .build();
   
    iteration.setTerminationCriterion(terminationMapper);

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

Examples of eu.stratosphere.api.java.record.operators.BulkIteration

62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
 
  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 eu.stratosphere.api.java.record.operators.BulkIteration

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
 
  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 SumReducer())
        .input(iteration.getPartialSolution())
        .name("Compute sum (Reduce)")
        .build();
   
    iteration.setNextPartialSolution(sumReduce);
   
    MapOperator terminationMapper = MapOperator.builder(new TerminationMapper())
        .input(iteration.getPartialSolution())
        .name("Compute termination criterion (Map)")
        .build();
   
    iteration.setTerminationCriterion(terminationMapper);

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

Examples of eu.stratosphere.api.java.record.operators.BulkIteration

643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
 
  @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
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.