use of org.apache.flink.api.common.aggregators.LongSumAggregator in project flink by apache.
the class DeltaIterationTranslationTest method testCorrectTranslation.
@Test
public void testCorrectTranslation() {
try {
final String JOB_NAME = "Test JobName";
final String ITERATION_NAME = "Test Name";
final String BEFORE_NEXT_WORKSET_MAP = "Some Mapper";
final String AGGREGATOR_NAME = "AggregatorName";
final int[] ITERATION_KEYS = new int[] { 2 };
final int NUM_ITERATIONS = 13;
final int DEFAULT_parallelism = 133;
final int ITERATION_parallelism = 77;
ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
// ------------ construct the test program ------------------
{
env.setParallelism(DEFAULT_parallelism);
@SuppressWarnings("unchecked") DataSet<Tuple3<Double, Long, String>> initialSolutionSet = env.fromElements(new Tuple3<Double, Long, String>(3.44, 5L, "abc"));
@SuppressWarnings("unchecked") DataSet<Tuple2<Double, String>> initialWorkSet = env.fromElements(new Tuple2<Double, String>(1.23, "abc"));
DeltaIteration<Tuple3<Double, Long, String>, Tuple2<Double, String>> iteration = initialSolutionSet.iterateDelta(initialWorkSet, NUM_ITERATIONS, ITERATION_KEYS);
iteration.name(ITERATION_NAME).parallelism(ITERATION_parallelism);
iteration.registerAggregator(AGGREGATOR_NAME, new LongSumAggregator());
// test that multiple workset consumers are supported
DataSet<Tuple2<Double, String>> worksetSelfJoin = iteration.getWorkset().map(new IdentityMapper<Tuple2<Double, String>>()).join(iteration.getWorkset()).where(1).equalTo(1).projectFirst(0, 1);
DataSet<Tuple3<Double, Long, String>> joined = worksetSelfJoin.join(iteration.getSolutionSet()).where(1).equalTo(2).with(new SolutionWorksetJoin());
DataSet<Tuple3<Double, Long, String>> result = iteration.closeWith(joined, joined.map(new NextWorksetMapper()).name(BEFORE_NEXT_WORKSET_MAP));
result.output(new DiscardingOutputFormat<Tuple3<Double, Long, String>>());
result.writeAsText("/dev/null");
}
Plan p = env.createProgramPlan(JOB_NAME);
// ------------- validate the plan ----------------
assertEquals(JOB_NAME, p.getJobName());
assertEquals(DEFAULT_parallelism, p.getDefaultParallelism());
// validate the iteration
GenericDataSinkBase<?> sink1, sink2;
{
Iterator<? extends GenericDataSinkBase<?>> sinks = p.getDataSinks().iterator();
sink1 = sinks.next();
sink2 = sinks.next();
}
DeltaIterationBase<?, ?> iteration = (DeltaIterationBase<?, ?>) sink1.getInput();
// check that multi consumer translation works for iterations
assertEquals(iteration, sink2.getInput());
// check the basic iteration properties
assertEquals(NUM_ITERATIONS, iteration.getMaximumNumberOfIterations());
assertArrayEquals(ITERATION_KEYS, iteration.getSolutionSetKeyFields());
assertEquals(ITERATION_parallelism, iteration.getParallelism());
assertEquals(ITERATION_NAME, iteration.getName());
MapOperatorBase<?, ?, ?> nextWorksetMapper = (MapOperatorBase<?, ?, ?>) iteration.getNextWorkset();
InnerJoinOperatorBase<?, ?, ?, ?> solutionSetJoin = (InnerJoinOperatorBase<?, ?, ?, ?>) iteration.getSolutionSetDelta();
InnerJoinOperatorBase<?, ?, ?, ?> worksetSelfJoin = (InnerJoinOperatorBase<?, ?, ?, ?>) solutionSetJoin.getFirstInput();
MapOperatorBase<?, ?, ?> worksetMapper = (MapOperatorBase<?, ?, ?>) worksetSelfJoin.getFirstInput();
assertEquals(IdentityMapper.class, worksetMapper.getUserCodeWrapper().getUserCodeClass());
assertEquals(NextWorksetMapper.class, nextWorksetMapper.getUserCodeWrapper().getUserCodeClass());
if (solutionSetJoin.getUserCodeWrapper().getUserCodeObject() instanceof WrappingFunction) {
WrappingFunction<?> wf = (WrappingFunction<?>) solutionSetJoin.getUserCodeWrapper().getUserCodeObject();
assertEquals(SolutionWorksetJoin.class, wf.getWrappedFunction().getClass());
} else {
assertEquals(SolutionWorksetJoin.class, solutionSetJoin.getUserCodeWrapper().getUserCodeClass());
}
assertEquals(BEFORE_NEXT_WORKSET_MAP, nextWorksetMapper.getName());
assertEquals(AGGREGATOR_NAME, iteration.getAggregators().getAllRegisteredAggregators().iterator().next().getName());
} catch (Exception e) {
System.err.println(e.getMessage());
e.printStackTrace();
fail(e.getMessage());
}
}
use of org.apache.flink.api.common.aggregators.LongSumAggregator in project flink by apache.
the class PregelTranslationTest method testTranslationPlainEdges.
@Test
public void testTranslationPlainEdges() {
try {
final String ITERATION_NAME = "Test Name";
final String AGGREGATOR_NAME = "AggregatorName";
final String BC_SET_NAME = "borat messages";
final int NUM_ITERATIONS = 13;
final int ITERATION_parallelism = 77;
ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
DataSet<Long> bcVar = env.fromElements(1L);
DataSet<Vertex<String, Double>> result;
// ------------ construct the test program ------------------
{
DataSet<Tuple2<String, Double>> initialVertices = env.fromElements(new Tuple2<>("abc", 3.44));
DataSet<Tuple2<String, String>> edges = env.fromElements(new Tuple2<>("a", "c"));
Graph<String, Double, NullValue> graph = Graph.fromTupleDataSet(initialVertices, edges.map(new MapFunction<Tuple2<String, String>, Tuple3<String, String, NullValue>>() {
public Tuple3<String, String, NullValue> map(Tuple2<String, String> edge) {
return new Tuple3<>(edge.f0, edge.f1, NullValue.getInstance());
}
}), env);
VertexCentricConfiguration parameters = new VertexCentricConfiguration();
parameters.addBroadcastSet(BC_SET_NAME, bcVar);
parameters.setName(ITERATION_NAME);
parameters.setParallelism(ITERATION_parallelism);
parameters.registerAggregator(AGGREGATOR_NAME, new LongSumAggregator());
result = graph.runVertexCentricIteration(new MyCompute(), null, NUM_ITERATIONS, parameters).getVertices();
result.output(new DiscardingOutputFormat<Vertex<String, Double>>());
}
// ------------- validate the java program ----------------
assertTrue(result instanceof DeltaIterationResultSet);
DeltaIterationResultSet<?, ?> resultSet = (DeltaIterationResultSet<?, ?>) result;
DeltaIteration<?, ?> iteration = resultSet.getIterationHead();
// check the basic iteration properties
assertEquals(NUM_ITERATIONS, resultSet.getMaxIterations());
assertArrayEquals(new int[] { 0 }, resultSet.getKeyPositions());
assertEquals(ITERATION_parallelism, iteration.getParallelism());
assertEquals(ITERATION_NAME, iteration.getName());
assertEquals(AGGREGATOR_NAME, iteration.getAggregators().getAllRegisteredAggregators().iterator().next().getName());
TwoInputUdfOperator<?, ?, ?, ?> computationCoGroup = (TwoInputUdfOperator<?, ?, ?, ?>) ((SingleInputUdfOperator<?, ?, ?>) resultSet.getNextWorkset()).getInput();
// validate that the broadcast sets are forwarded
assertEquals(bcVar, computationCoGroup.getBroadcastSets().get(BC_SET_NAME));
} catch (Exception e) {
System.err.println(e.getMessage());
e.printStackTrace();
fail(e.getMessage());
}
}
use of org.apache.flink.api.common.aggregators.LongSumAggregator in project flink by apache.
the class SpargelTranslationTest method testTranslationPlainEdgesWithForkedBroadcastVariable.
@Test
public void testTranslationPlainEdgesWithForkedBroadcastVariable() {
try {
final String ITERATION_NAME = "Test Name";
final String AGGREGATOR_NAME = "AggregatorName";
final String BC_SET_MESSAGES_NAME = "borat messages";
final String BC_SET_UPDATES_NAME = "borat updates";
final int NUM_ITERATIONS = 13;
final int ITERATION_parallelism = 77;
ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
DataSet<Long> bcVar = env.fromElements(1L);
DataSet<Vertex<String, Double>> result;
// ------------ construct the test program ------------------
{
DataSet<Tuple2<String, Double>> initialVertices = env.fromElements(new Tuple2<>("abc", 3.44));
DataSet<Tuple2<String, String>> edges = env.fromElements(new Tuple2<>("a", "c"));
Graph<String, Double, NullValue> graph = Graph.fromTupleDataSet(initialVertices, edges.map(new MapFunction<Tuple2<String, String>, Tuple3<String, String, NullValue>>() {
public Tuple3<String, String, NullValue> map(Tuple2<String, String> edge) {
return new Tuple3<>(edge.f0, edge.f1, NullValue.getInstance());
}
}), env);
ScatterGatherConfiguration parameters = new ScatterGatherConfiguration();
parameters.addBroadcastSetForScatterFunction(BC_SET_MESSAGES_NAME, bcVar);
parameters.addBroadcastSetForGatherFunction(BC_SET_UPDATES_NAME, bcVar);
parameters.setName(ITERATION_NAME);
parameters.setParallelism(ITERATION_parallelism);
parameters.registerAggregator(AGGREGATOR_NAME, new LongSumAggregator());
result = graph.runScatterGatherIteration(new MessageFunctionNoEdgeValue(), new UpdateFunction(), NUM_ITERATIONS, parameters).getVertices();
result.output(new DiscardingOutputFormat<Vertex<String, Double>>());
}
// ------------- validate the java program ----------------
assertTrue(result instanceof DeltaIterationResultSet);
DeltaIterationResultSet<?, ?> resultSet = (DeltaIterationResultSet<?, ?>) result;
DeltaIteration<?, ?> iteration = resultSet.getIterationHead();
// check the basic iteration properties
assertEquals(NUM_ITERATIONS, resultSet.getMaxIterations());
assertArrayEquals(new int[] { 0 }, resultSet.getKeyPositions());
assertEquals(ITERATION_parallelism, iteration.getParallelism());
assertEquals(ITERATION_NAME, iteration.getName());
assertEquals(AGGREGATOR_NAME, iteration.getAggregators().getAllRegisteredAggregators().iterator().next().getName());
// validate that the semantic properties are set as they should
TwoInputUdfOperator<?, ?, ?, ?> solutionSetJoin = (TwoInputUdfOperator<?, ?, ?, ?>) resultSet.getNextWorkset();
assertTrue(solutionSetJoin.getSemanticProperties().getForwardingTargetFields(0, 0).contains(0));
assertTrue(solutionSetJoin.getSemanticProperties().getForwardingTargetFields(1, 0).contains(0));
TwoInputUdfOperator<?, ?, ?, ?> edgesJoin = (TwoInputUdfOperator<?, ?, ?, ?>) solutionSetJoin.getInput1();
// validate that the broadcast sets are forwarded
assertEquals(bcVar, solutionSetJoin.getBroadcastSets().get(BC_SET_UPDATES_NAME));
assertEquals(bcVar, edgesJoin.getBroadcastSets().get(BC_SET_MESSAGES_NAME));
} catch (Exception e) {
System.err.println(e.getMessage());
e.printStackTrace();
fail(e.getMessage());
}
}
use of org.apache.flink.api.common.aggregators.LongSumAggregator in project flink by apache.
the class GSATranslationTest method testTranslation.
@Test
public void testTranslation() {
try {
final String ITERATION_NAME = "Test Name";
final String AGGREGATOR_NAME = "AggregatorName";
final String BC_SET_GATHER_NAME = "gather messages";
final String BC_SET_SUM_NAME = "sum updates";
final String BC_SET_APLLY_NAME = "apply updates";
final int NUM_ITERATIONS = 13;
final int ITERATION_parallelism = 77;
ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
DataSet<Long> bcGather = env.fromElements(1L);
DataSet<Long> bcSum = env.fromElements(1L);
DataSet<Long> bcApply = env.fromElements(1L);
DataSet<Vertex<Long, Long>> result;
// ------------ construct the test program ------------------
{
DataSet<Edge<Long, NullValue>> edges = env.fromElements(new Tuple3<>(1L, 2L, NullValue.getInstance())).map(new Tuple3ToEdgeMap<Long, NullValue>());
Graph<Long, Long, NullValue> graph = Graph.fromDataSet(edges, new InitVertices(), env);
GSAConfiguration parameters = new GSAConfiguration();
parameters.registerAggregator(AGGREGATOR_NAME, new LongSumAggregator());
parameters.setName(ITERATION_NAME);
parameters.setParallelism(ITERATION_parallelism);
parameters.addBroadcastSetForGatherFunction(BC_SET_GATHER_NAME, bcGather);
parameters.addBroadcastSetForSumFunction(BC_SET_SUM_NAME, bcSum);
parameters.addBroadcastSetForApplyFunction(BC_SET_APLLY_NAME, bcApply);
result = graph.runGatherSumApplyIteration(new GatherNeighborIds(), new SelectMinId(), new UpdateComponentId(), NUM_ITERATIONS, parameters).getVertices();
result.output(new DiscardingOutputFormat<Vertex<Long, Long>>());
}
// ------------- validate the java program ----------------
assertTrue(result instanceof DeltaIterationResultSet);
DeltaIterationResultSet<?, ?> resultSet = (DeltaIterationResultSet<?, ?>) result;
DeltaIteration<?, ?> iteration = resultSet.getIterationHead();
// check the basic iteration properties
assertEquals(NUM_ITERATIONS, resultSet.getMaxIterations());
assertArrayEquals(new int[] { 0 }, resultSet.getKeyPositions());
assertEquals(ITERATION_parallelism, iteration.getParallelism());
assertEquals(ITERATION_NAME, iteration.getName());
assertEquals(AGGREGATOR_NAME, iteration.getAggregators().getAllRegisteredAggregators().iterator().next().getName());
// validate that the semantic properties are set as they should
TwoInputUdfOperator<?, ?, ?, ?> solutionSetJoin = (TwoInputUdfOperator<?, ?, ?, ?>) resultSet.getNextWorkset();
assertTrue(solutionSetJoin.getSemanticProperties().getForwardingTargetFields(0, 0).contains(0));
assertTrue(solutionSetJoin.getSemanticProperties().getForwardingTargetFields(1, 0).contains(0));
SingleInputUdfOperator<?, ?, ?> sumReduce = (SingleInputUdfOperator<?, ?, ?>) solutionSetJoin.getInput1();
SingleInputUdfOperator<?, ?, ?> gatherMap = (SingleInputUdfOperator<?, ?, ?>) sumReduce.getInput();
// validate that the broadcast sets are forwarded
assertEquals(bcGather, gatherMap.getBroadcastSets().get(BC_SET_GATHER_NAME));
assertEquals(bcSum, sumReduce.getBroadcastSets().get(BC_SET_SUM_NAME));
assertEquals(bcApply, solutionSetJoin.getBroadcastSets().get(BC_SET_APLLY_NAME));
} catch (Exception e) {
System.err.println(e.getMessage());
e.printStackTrace();
fail(e.getMessage());
}
}
use of org.apache.flink.api.common.aggregators.LongSumAggregator in project flink by apache.
the class ScatterGatherConfigurationITCase method testRunWithConfiguration.
@Test
public void testRunWithConfiguration() throws Exception {
/*
* Test Graph's runScatterGatherIteration when configuration parameters are provided
*/
final ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
Graph<Long, Long, Long> graph = Graph.fromCollection(TestGraphUtils.getLongLongVertices(), TestGraphUtils.getLongLongEdges(), env).mapVertices(new AssignOneMapper());
// create the configuration object
ScatterGatherConfiguration parameters = new ScatterGatherConfiguration();
parameters.addBroadcastSetForScatterFunction("messagingBcastSet", env.fromElements(4, 5, 6));
parameters.addBroadcastSetForGatherFunction("updateBcastSet", env.fromElements(1, 2, 3));
parameters.registerAggregator("superstepAggregator", new LongSumAggregator());
parameters.setOptNumVertices(true);
Graph<Long, Long, Long> res = graph.runScatterGatherIteration(new MessageFunction(), new UpdateFunction(), 10, parameters);
DataSet<Vertex<Long, Long>> data = res.getVertices();
List<Vertex<Long, Long>> result = data.collect();
expectedResult = "1,11\n" + "2,11\n" + "3,11\n" + "4,11\n" + "5,11";
compareResultAsTuples(result, expectedResult);
}
Aggregations