Search in sources :

Example 51 with KeyedOneInputStreamOperatorTestHarness

use of org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness in project flink by apache.

the class WindowOperatorTest method testCleanupTimerWithEmptyFoldingStateForTumblingWindows.

@Test
public void testCleanupTimerWithEmptyFoldingStateForTumblingWindows() throws Exception {
    final int WINDOW_SIZE = 2;
    final long LATENESS = 1;
    TypeInformation<Tuple2<String, Integer>> inputType = TypeInfoParser.parse("Tuple2<String, Integer>");
    FoldingStateDescriptor<Tuple2<String, Integer>, Tuple2<String, Integer>> windowStateDesc = new FoldingStateDescriptor<>("window-contents", new Tuple2<>((String) null, 0), new FoldFunction<Tuple2<String, Integer>, Tuple2<String, Integer>>() {

        private static final long serialVersionUID = 1L;

        @Override
        public Tuple2<String, Integer> fold(Tuple2<String, Integer> accumulator, Tuple2<String, Integer> value) throws Exception {
            return new Tuple2<>(value.f0, accumulator.f1 + value.f1);
        }
    }, inputType);
    windowStateDesc.initializeSerializerUnlessSet(new ExecutionConfig());
    WindowOperator<String, Tuple2<String, Integer>, Tuple2<String, Integer>, Tuple2<String, Integer>, TimeWindow> operator = new WindowOperator<>(TumblingEventTimeWindows.of(Time.of(WINDOW_SIZE, TimeUnit.SECONDS)), new TimeWindow.Serializer(), new TupleKeySelector(), BasicTypeInfo.STRING_TYPE_INFO.createSerializer(new ExecutionConfig()), windowStateDesc, new InternalSingleValueWindowFunction<>(new PassThroughFunction()), EventTimeTrigger.create(), LATENESS, null);
    OneInputStreamOperatorTestHarness<Tuple2<String, Integer>, Tuple2<String, Integer>> testHarness = new KeyedOneInputStreamOperatorTestHarness<>(operator, new TupleKeySelector(), BasicTypeInfo.STRING_TYPE_INFO);
    testHarness.open();
    ConcurrentLinkedQueue<Object> expected = new ConcurrentLinkedQueue<>();
    // normal element
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), 1000));
    testHarness.processWatermark(new Watermark(1599));
    testHarness.processWatermark(new Watermark(1999));
    testHarness.processWatermark(new Watermark(2000));
    testHarness.processWatermark(new Watermark(5000));
    expected.add(new Watermark(1599));
    expected.add(new StreamRecord<>(new Tuple2<>("key2", 1), 1999));
    // here it fires and purges
    expected.add(new Watermark(1999));
    // here is the cleanup timer
    expected.add(new Watermark(2000));
    expected.add(new Watermark(5000));
    TestHarnessUtil.assertOutputEqualsSorted("Output was not correct.", expected, testHarness.getOutput(), new Tuple2ResultSortComparator());
    testHarness.close();
}
Also used : ExecutionConfig(org.apache.flink.api.common.ExecutionConfig) FoldingStateDescriptor(org.apache.flink.api.common.state.FoldingStateDescriptor) KeyedOneInputStreamOperatorTestHarness(org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness) TimeWindow(org.apache.flink.streaming.api.windowing.windows.TimeWindow) AtomicInteger(java.util.concurrent.atomic.AtomicInteger) Tuple2(org.apache.flink.api.java.tuple.Tuple2) ConcurrentLinkedQueue(java.util.concurrent.ConcurrentLinkedQueue) Watermark(org.apache.flink.streaming.api.watermark.Watermark) Test(org.junit.Test)

Example 52 with KeyedOneInputStreamOperatorTestHarness

use of org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness in project flink by apache.

the class WindowOperatorTest method testSessionWindowsWithContinuousEventTimeTrigger.

/**
	 * This tests whether merging works correctly with the ContinuousEventTimeTrigger.
	 * @throws Exception
	 */
@Test
@SuppressWarnings("unchecked")
public void testSessionWindowsWithContinuousEventTimeTrigger() throws Exception {
    closeCalled.set(0);
    final int SESSION_SIZE = 3;
    TypeInformation<Tuple2<String, Integer>> inputType = TypeInfoParser.parse("Tuple2<String, Integer>");
    ListStateDescriptor<Tuple2<String, Integer>> stateDesc = new ListStateDescriptor<>("window-contents", inputType.createSerializer(new ExecutionConfig()));
    WindowOperator<String, Tuple2<String, Integer>, Iterable<Tuple2<String, Integer>>, Tuple3<String, Long, Long>, TimeWindow> operator = new WindowOperator<>(EventTimeSessionWindows.withGap(Time.seconds(SESSION_SIZE)), new TimeWindow.Serializer(), new TupleKeySelector(), BasicTypeInfo.STRING_TYPE_INFO.createSerializer(new ExecutionConfig()), stateDesc, new InternalIterableWindowFunction<>(new SessionWindowFunction()), ContinuousEventTimeTrigger.of(Time.seconds(2)), 0, null);
    OneInputStreamOperatorTestHarness<Tuple2<String, Integer>, Tuple3<String, Long, Long>> testHarness = new KeyedOneInputStreamOperatorTestHarness<>(operator, new TupleKeySelector(), BasicTypeInfo.STRING_TYPE_INFO);
    ConcurrentLinkedQueue<Object> expectedOutput = new ConcurrentLinkedQueue<>();
    testHarness.open();
    // add elements out-of-order and first trigger time is 2000
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key1", 1), 1500));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), 0));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 3), 2500));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 2), 1000));
    // triggers emit and next trigger time is 4000
    testHarness.processWatermark(new Watermark(2500));
    expectedOutput.add(new StreamRecord<>(new Tuple3<>("key1-1", 1500L, 4500L), 4499));
    expectedOutput.add(new StreamRecord<>(new Tuple3<>("key2-6", 0L, 5500L), 5499));
    expectedOutput.add(new Watermark(2500));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 5), 4000));
    testHarness.processWatermark(new Watermark(3000));
    expectedOutput.add(new Watermark(3000));
    // do a snapshot, close and restore again
    OperatorStateHandles snapshot = testHarness.snapshot(0L, 0L);
    testHarness.close();
    testHarness.setup();
    testHarness.initializeState(snapshot);
    testHarness.open();
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key1", 2), 4000));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 4), 3500));
    // triggers emit and next trigger time is 6000
    testHarness.processWatermark(new Watermark(4000));
    expectedOutput.add(new StreamRecord<>(new Tuple3<>("key1-3", 1500L, 7000L), 6999));
    expectedOutput.add(new StreamRecord<>(new Tuple3<>("key2-15", 0L, 7000L), 6999));
    expectedOutput.add(new Watermark(4000));
    TestHarnessUtil.assertOutputEqualsSorted("Output was not correct.", expectedOutput, testHarness.getOutput(), new Tuple3ResultSortComparator());
    testHarness.close();
}
Also used : ListStateDescriptor(org.apache.flink.api.common.state.ListStateDescriptor) ExecutionConfig(org.apache.flink.api.common.ExecutionConfig) KeyedOneInputStreamOperatorTestHarness(org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness) OperatorStateHandles(org.apache.flink.streaming.runtime.tasks.OperatorStateHandles) TimeWindow(org.apache.flink.streaming.api.windowing.windows.TimeWindow) AtomicInteger(java.util.concurrent.atomic.AtomicInteger) Tuple2(org.apache.flink.api.java.tuple.Tuple2) Tuple3(org.apache.flink.api.java.tuple.Tuple3) ConcurrentLinkedQueue(java.util.concurrent.ConcurrentLinkedQueue) Watermark(org.apache.flink.streaming.api.watermark.Watermark) Test(org.junit.Test)

Example 53 with KeyedOneInputStreamOperatorTestHarness

use of org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness in project flink by apache.

the class WindowOperatorTest method testNotSideOutputDueToLatenessSessionWithLateness.

@Test
public void testNotSideOutputDueToLatenessSessionWithLateness() throws Exception {
    // same as testSideOutputDueToLatenessSessionWithLateness() but with an accumulating trigger, i.e.
    // one that does not return FIRE_AND_PURGE when firing but just FIRE. The expected
    // results are therefore slightly different.
    final int GAP_SIZE = 3;
    final long LATENESS = 10;
    TypeInformation<Tuple2<String, Integer>> inputType = TypeInfoParser.parse("Tuple2<String, Integer>");
    ReducingStateDescriptor<Tuple2<String, Integer>> stateDesc = new ReducingStateDescriptor<>("window-contents", new SumReducer(), inputType.createSerializer(new ExecutionConfig()));
    WindowOperator<String, Tuple2<String, Integer>, Tuple2<String, Integer>, Tuple3<String, Long, Long>, TimeWindow> operator = new WindowOperator<>(EventTimeSessionWindows.withGap(Time.seconds(GAP_SIZE)), new TimeWindow.Serializer(), new TupleKeySelector(), BasicTypeInfo.STRING_TYPE_INFO.createSerializer(new ExecutionConfig()), stateDesc, new InternalSingleValueWindowFunction<>(new ReducedSessionWindowFunction()), EventTimeTrigger.create(), LATENESS, null);
    OneInputStreamOperatorTestHarness<Tuple2<String, Integer>, Tuple3<String, Long, Long>> testHarness = new KeyedOneInputStreamOperatorTestHarness<>(operator, new TupleKeySelector(), BasicTypeInfo.STRING_TYPE_INFO);
    testHarness.open();
    ConcurrentLinkedQueue<Object> expected = new ConcurrentLinkedQueue<>();
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), 1000));
    testHarness.processWatermark(new Watermark(1999));
    expected.add(new Watermark(1999));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), 2000));
    testHarness.processWatermark(new Watermark(4998));
    expected.add(new Watermark(4998));
    // this will not be sideoutput because the session we're adding two has maxTimestamp
    // after the current watermark
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), 4500));
    // new session
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), 8500));
    testHarness.processWatermark(new Watermark(7400));
    expected.add(new Watermark(7400));
    // this will merge the two sessions into one
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), 7000));
    testHarness.processWatermark(new Watermark(11501));
    expected.add(new StreamRecord<>(new Tuple3<>("key2-5", 1000L, 11500L), 11499));
    expected.add(new Watermark(11501));
    // new session
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), 11600));
    testHarness.processWatermark(new Watermark(14600));
    expected.add(new StreamRecord<>(new Tuple3<>("key2-1", 11600L, 14600L), 14599));
    expected.add(new Watermark(14600));
    // because of the small allowed lateness and because the trigger is accumulating
    // this will be merged into the session (11600-14600) and therefore will not
    // be sideoutput as late
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), 10000));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), 14500));
    // adding ("key2", 1) extended the session to (10000-146000) for which
    // maxTimestamp <= currentWatermark. Therefore, we immediately get a firing
    // with the current version of EventTimeTrigger/EventTimeTriggerAccum
    expected.add(new StreamRecord<>(new Tuple3<>("key2-2", 10000L, 14600L), 14599));
    ConcurrentLinkedQueue<Object> actual = testHarness.getOutput();
    ConcurrentLinkedQueue<StreamRecord<Tuple2<String, Integer>>> sideActual = testHarness.getSideOutput(lateOutputTag);
    TestHarnessUtil.assertOutputEqualsSorted("Output was not correct.", expected, actual, new Tuple3ResultSortComparator());
    assertEquals(null, sideActual);
    testHarness.processWatermark(new Watermark(20000));
    expected.add(new StreamRecord<>(new Tuple3<>("key2-3", 10000L, 17500L), 17499));
    expected.add(new Watermark(20000));
    testHarness.processWatermark(new Watermark(100000));
    expected.add(new Watermark(100000));
    actual = testHarness.getOutput();
    sideActual = testHarness.getSideOutput(lateOutputTag);
    TestHarnessUtil.assertOutputEqualsSorted("Output was not correct.", expected, actual, new Tuple3ResultSortComparator());
    assertEquals(null, sideActual);
    testHarness.close();
}
Also used : ExecutionConfig(org.apache.flink.api.common.ExecutionConfig) KeyedOneInputStreamOperatorTestHarness(org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness) ReducingStateDescriptor(org.apache.flink.api.common.state.ReducingStateDescriptor) StreamRecord(org.apache.flink.streaming.runtime.streamrecord.StreamRecord) TimeWindow(org.apache.flink.streaming.api.windowing.windows.TimeWindow) AtomicInteger(java.util.concurrent.atomic.AtomicInteger) Tuple2(org.apache.flink.api.java.tuple.Tuple2) Tuple3(org.apache.flink.api.java.tuple.Tuple3) ConcurrentLinkedQueue(java.util.concurrent.ConcurrentLinkedQueue) Watermark(org.apache.flink.streaming.api.watermark.Watermark) Test(org.junit.Test)

Example 54 with KeyedOneInputStreamOperatorTestHarness

use of org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness in project flink by apache.

the class WindowTranslationTest method processElementAndEnsureOutput.

/**
	 * Ensure that we get some output from the given operator when pushing in an element and
	 * setting watermark and processing time to {@code Long.MAX_VALUE}.
	 */
private static <K, IN, OUT> void processElementAndEnsureOutput(OneInputStreamOperator<IN, OUT> operator, KeySelector<IN, K> keySelector, TypeInformation<K> keyType, IN element) throws Exception {
    KeyedOneInputStreamOperatorTestHarness<K, IN, OUT> testHarness = new KeyedOneInputStreamOperatorTestHarness<>(operator, keySelector, keyType);
    if (operator instanceof OutputTypeConfigurable) {
        // use a dummy type since window functions just need the ExecutionConfig
        // this is also only needed for Fold, which we're getting rid off soon.
        ((OutputTypeConfigurable) operator).setOutputType(BasicTypeInfo.STRING_TYPE_INFO, new ExecutionConfig());
    }
    testHarness.open();
    testHarness.setProcessingTime(0);
    testHarness.processWatermark(Long.MIN_VALUE);
    testHarness.processElement(new StreamRecord<>(element, 0));
    // provoke any processing-time/event-time triggers
    testHarness.setProcessingTime(Long.MAX_VALUE);
    testHarness.processWatermark(Long.MAX_VALUE);
    // we at least get the two watermarks and should also see an output element
    assertTrue(testHarness.getOutput().size() >= 3);
    testHarness.close();
}
Also used : OutputTypeConfigurable(org.apache.flink.streaming.api.operators.OutputTypeConfigurable) ExecutionConfig(org.apache.flink.api.common.ExecutionConfig) KeyedOneInputStreamOperatorTestHarness(org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness)

Example 55 with KeyedOneInputStreamOperatorTestHarness

use of org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness in project flink by apache.

the class EvictingWindowOperatorTest method testTimeEvictorEvictBefore.

/**
	 * Tests TimeEvictor evictBefore behavior
	 * @throws Exception
	 */
@Test
public void testTimeEvictorEvictBefore() throws Exception {
    AtomicInteger closeCalled = new AtomicInteger(0);
    final int TRIGGER_COUNT = 2;
    final int WINDOW_SIZE = 4;
    TypeInformation<Tuple2<String, Integer>> inputType = TypeInfoParser.parse("Tuple2<String, Integer>");
    @SuppressWarnings({ "unchecked", "rawtypes" }) TypeSerializer<StreamRecord<Tuple2<String, Integer>>> streamRecordSerializer = (TypeSerializer<StreamRecord<Tuple2<String, Integer>>>) new StreamElementSerializer(inputType.createSerializer(new ExecutionConfig()));
    ListStateDescriptor<StreamRecord<Tuple2<String, Integer>>> stateDesc = new ListStateDescriptor<>("window-contents", streamRecordSerializer);
    EvictingWindowOperator<String, Tuple2<String, Integer>, Tuple2<String, Integer>, TimeWindow> operator = new EvictingWindowOperator<>(TumblingEventTimeWindows.of(Time.of(WINDOW_SIZE, TimeUnit.SECONDS)), new TimeWindow.Serializer(), new TupleKeySelector(), BasicTypeInfo.STRING_TYPE_INFO.createSerializer(new ExecutionConfig()), stateDesc, new InternalIterableWindowFunction<>(new RichSumReducer<TimeWindow>(closeCalled)), CountTrigger.of(TRIGGER_COUNT), TimeEvictor.of(Time.seconds(2)), 0, null);
    OneInputStreamOperatorTestHarness<Tuple2<String, Integer>, Tuple2<String, Integer>> testHarness = new KeyedOneInputStreamOperatorTestHarness<>(operator, new TupleKeySelector(), BasicTypeInfo.STRING_TYPE_INFO);
    long initialTime = 0L;
    ConcurrentLinkedQueue<Object> expectedOutput = new ConcurrentLinkedQueue<>();
    testHarness.open();
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), initialTime + 1000));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), initialTime + 3999));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key1", 1), initialTime + 20));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key1", 1), initialTime));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key1", 1), initialTime + 999));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key1", 1), initialTime + 5999));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), initialTime + 3500));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), initialTime + 2001));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), initialTime + 1001));
    expectedOutput.add(new StreamRecord<>(new Tuple2<>("key2", 1), 3999));
    expectedOutput.add(new StreamRecord<>(new Tuple2<>("key1", 2), 3999));
    expectedOutput.add(new StreamRecord<>(new Tuple2<>("key2", 3), 3999));
    TestHarnessUtil.assertOutputEqualsSorted("Output was not correct.", expectedOutput, testHarness.getOutput(), new ResultSortComparator());
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key1", 1), initialTime + 6500));
    testHarness.processElement(new StreamRecord<>(new Tuple2<>("key2", 1), initialTime + 1002));
    expectedOutput.add(new StreamRecord<>(new Tuple2<>("key1", 2), 7999));
    expectedOutput.add(new StreamRecord<>(new Tuple2<>("key2", 3), 3999));
    TestHarnessUtil.assertOutputEqualsSorted("Output was not correct.", expectedOutput, testHarness.getOutput(), new ResultSortComparator());
    testHarness.close();
    Assert.assertEquals("Close was not called.", 1, closeCalled.get());
}
Also used : ListStateDescriptor(org.apache.flink.api.common.state.ListStateDescriptor) ExecutionConfig(org.apache.flink.api.common.ExecutionConfig) KeyedOneInputStreamOperatorTestHarness(org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness) TypeSerializer(org.apache.flink.api.common.typeutils.TypeSerializer) StreamElementSerializer(org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer) StreamRecord(org.apache.flink.streaming.runtime.streamrecord.StreamRecord) TimeWindow(org.apache.flink.streaming.api.windowing.windows.TimeWindow) AtomicInteger(java.util.concurrent.atomic.AtomicInteger) AtomicInteger(java.util.concurrent.atomic.AtomicInteger) Tuple2(org.apache.flink.api.java.tuple.Tuple2) ConcurrentLinkedQueue(java.util.concurrent.ConcurrentLinkedQueue) Test(org.junit.Test)

Aggregations

KeyedOneInputStreamOperatorTestHarness (org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness)55 Test (org.junit.Test)54 ConcurrentLinkedQueue (java.util.concurrent.ConcurrentLinkedQueue)43 Tuple2 (org.apache.flink.api.java.tuple.Tuple2)42 ExecutionConfig (org.apache.flink.api.common.ExecutionConfig)41 Watermark (org.apache.flink.streaming.api.watermark.Watermark)36 AtomicInteger (java.util.concurrent.atomic.AtomicInteger)31 TimeWindow (org.apache.flink.streaming.api.windowing.windows.TimeWindow)28 ListStateDescriptor (org.apache.flink.api.common.state.ListStateDescriptor)18 ReducingStateDescriptor (org.apache.flink.api.common.state.ReducingStateDescriptor)17 StreamRecord (org.apache.flink.streaming.runtime.streamrecord.StreamRecord)17 Tuple3 (org.apache.flink.api.java.tuple.Tuple3)14 OperatorStateHandles (org.apache.flink.streaming.runtime.tasks.OperatorStateHandles)14 TypeSerializer (org.apache.flink.api.common.typeutils.TypeSerializer)9 GlobalWindow (org.apache.flink.streaming.api.windowing.windows.GlobalWindow)9 StreamElementSerializer (org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer)9 PassThroughWindowFunction (org.apache.flink.streaming.api.functions.windowing.PassThroughWindowFunction)8 KeySelector (org.apache.flink.api.java.functions.KeySelector)4 PrepareForTest (org.powermock.core.classloader.annotations.PrepareForTest)4 Map (java.util.Map)3