Examples with KvCoder org.apache.beam.sdk.coders.KvCoder used on opensource projects

Example 26 with KvCoder

use of org.apache.beam.sdk.coders.KvCoder in project beam by apache.

the class IntrinsicMapTaskExecutorFactory method createPartialGroupByKeyOperation.

<K> OperationNode createPartialGroupByKeyOperation(Network<Node, Edge> network, ParallelInstructionNode node, PipelineOptions options, DataflowExecutionContext<?> executionContext, DataflowOperationContext operationContext) throws Exception {
    ParallelInstruction instruction = node.getParallelInstruction();
    PartialGroupByKeyInstruction pgbk = instruction.getPartialGroupByKey();
    OutputReceiver[] receivers = getOutputReceivers(network, node);
    Coder<?> windowedCoder = CloudObjects.coderFromCloudObject(CloudObject.fromSpec(pgbk.getInputElementCodec()));
    if (!(windowedCoder instanceof WindowedValueCoder)) {
        throw new IllegalArgumentException(String.format("unexpected kind of input coder for PartialGroupByKeyOperation: %s", windowedCoder));
    }
    Coder<?> elemCoder = ((WindowedValueCoder<?>) windowedCoder).getValueCoder();
    if (!(elemCoder instanceof KvCoder)) {
        throw new IllegalArgumentException(String.format("unexpected kind of input element coder for PartialGroupByKeyOperation: %s", elemCoder));
    }
    @SuppressWarnings("unchecked") KvCoder<K, ?> keyedElementCoder = (KvCoder<K, ?>) elemCoder;
    CloudObject cloudUserFn = pgbk.getValueCombiningFn() != null ? CloudObject.fromSpec(pgbk.getValueCombiningFn()) : null;
    ParDoFn fn = PartialGroupByKeyParDoFns.create(options, keyedElementCoder, cloudUserFn, pgbk.getSideInputs(), Arrays.<Receiver>asList(receivers), executionContext, operationContext);
    return OperationNode.create(new ParDoOperation(fn, receivers, operationContext));
}

Example 27 with KvCoder

use of org.apache.beam.sdk.coders.KvCoder in project beam by apache.

the class GroupAlsoByWindowParDoFnFactory method create.

@Override
public ParDoFn create(PipelineOptions options, CloudObject cloudUserFn, @Nullable List<SideInputInfo> sideInputInfos, TupleTag<?> mainOutputTag, Map<TupleTag<?>, Integer> outputTupleTagsToReceiverIndices, final DataflowExecutionContext<?> executionContext, DataflowOperationContext operationContext) throws Exception {
    Map.Entry<TupleTag<?>, Integer> entry = Iterables.getOnlyElement(outputTupleTagsToReceiverIndices.entrySet());
    checkArgument(entry.getKey().equals(mainOutputTag), "Output tags should reference only the main output tag: %s vs %s", entry.getKey(), mainOutputTag);
    checkArgument(entry.getValue() == 0, "There should be a single receiver, but using receiver index %s", entry.getValue());
    byte[] encodedWindowingStrategy = getBytes(cloudUserFn, PropertyNames.SERIALIZED_FN);
    WindowingStrategy windowingStrategy;
    try {
        windowingStrategy = deserializeWindowingStrategy(encodedWindowingStrategy);
    } catch (Exception e) {
        // TODO: Catch block disappears, becoming an error once Python SDK is compliant.
        if (DataflowRunner.hasExperiment(options.as(DataflowPipelineDebugOptions.class), "beam_fn_api")) {
            LOG.info("FnAPI: Unable to deserialize windowing strategy, assuming default", e);
            windowingStrategy = WindowingStrategy.globalDefault();
        } else {
            throw e;
        }
    }
    byte[] serializedCombineFn = getBytes(cloudUserFn, WorkerPropertyNames.COMBINE_FN, null);
    AppliedCombineFn<?, ?, ?, ?> combineFn = null;
    if (serializedCombineFn != null) {
        Object combineFnObj = SerializableUtils.deserializeFromByteArray(serializedCombineFn, "serialized combine fn");
        checkArgument(combineFnObj instanceof AppliedCombineFn, "unexpected kind of AppliedCombineFn: " + combineFnObj.getClass().getName());
        combineFn = (AppliedCombineFn<?, ?, ?, ?>) combineFnObj;
    }
    Map<String, Object> inputCoderObject = getObject(cloudUserFn, WorkerPropertyNames.INPUT_CODER);
    Coder<?> inputCoder = CloudObjects.coderFromCloudObject(CloudObject.fromSpec(inputCoderObject));
    checkArgument(inputCoder instanceof WindowedValueCoder, "Expected WindowedValueCoder for inputCoder, got: " + inputCoder.getClass().getName());
    @SuppressWarnings("unchecked") WindowedValueCoder<?> windowedValueCoder = (WindowedValueCoder<?>) inputCoder;
    Coder<?> elemCoder = windowedValueCoder.getValueCoder();
    checkArgument(elemCoder instanceof KvCoder, "Expected KvCoder for inputCoder, got: " + elemCoder.getClass().getName());
    @SuppressWarnings("unchecked") KvCoder<?, ?> kvCoder = (KvCoder<?, ?>) elemCoder;
    boolean isStreamingPipeline = options.as(StreamingOptions.class).isStreaming();
    SideInputReader sideInputReader = NullSideInputReader.empty();
    @Nullable AppliedCombineFn<?, ?, ?, ?> maybeMergingCombineFn = null;
    if (combineFn != null) {
        sideInputReader = executionContext.getSideInputReader(sideInputInfos, combineFn.getSideInputViews(), operationContext);
        String phase = getString(cloudUserFn, WorkerPropertyNames.PHASE, CombinePhase.ALL);
        checkArgument(phase.equals(CombinePhase.ALL) || phase.equals(CombinePhase.MERGE), "Unexpected phase: %s", phase);
        if (phase.equals(CombinePhase.MERGE)) {
            maybeMergingCombineFn = makeAppliedMergingFunction(combineFn);
        } else {
            maybeMergingCombineFn = combineFn;
        }
    }
    StateInternalsFactory<?> stateInternalsFactory = key -> executionContext.getStepContext(operationContext).stateInternals();
    // This will be a GABW Fn for either batch or streaming, with combiner in it or not
    GroupAlsoByWindowFn<?, ?> fn;
    // This will be a FakeKeyedWorkItemCoder for streaming or null for batch
    Coder<?> gabwInputCoder;
    // TODO: do not do this with mess of "if"
    if (isStreamingPipeline) {
        if (maybeMergingCombineFn == null) {
            fn = StreamingGroupAlsoByWindowsDoFns.createForIterable(windowingStrategy, stateInternalsFactory, ((KvCoder) kvCoder).getValueCoder());
            gabwInputCoder = WindmillKeyedWorkItem.FakeKeyedWorkItemCoder.of(kvCoder);
        } else {
            fn = StreamingGroupAlsoByWindowsDoFns.create(windowingStrategy, stateInternalsFactory, (AppliedCombineFn) maybeMergingCombineFn, ((KvCoder) kvCoder).getKeyCoder());
            gabwInputCoder = WindmillKeyedWorkItem.FakeKeyedWorkItemCoder.of(((AppliedCombineFn) maybeMergingCombineFn).getKvCoder());
        }
    } else {
        if (maybeMergingCombineFn == null) {
            fn = BatchGroupAlsoByWindowsDoFns.createForIterable(windowingStrategy, stateInternalsFactory, ((KvCoder) kvCoder).getValueCoder());
            gabwInputCoder = null;
        } else {
            fn = BatchGroupAlsoByWindowsDoFns.create(windowingStrategy, (AppliedCombineFn) maybeMergingCombineFn);
            gabwInputCoder = null;
        }
    }
    // TODO: or anyhow related to it, do not do this with mess of "if"
    if (maybeMergingCombineFn != null) {
        return new GroupAlsoByWindowsParDoFn(options, fn, windowingStrategy, ((AppliedCombineFn) maybeMergingCombineFn).getSideInputViews(), gabwInputCoder, sideInputReader, mainOutputTag, executionContext.getStepContext(operationContext));
    } else {
        return new GroupAlsoByWindowsParDoFn(options, fn, windowingStrategy, null, gabwInputCoder, sideInputReader, mainOutputTag, executionContext.getStepContext(operationContext));
    }
}

Example 28 with KvCoder

use of org.apache.beam.sdk.coders.KvCoder in project beam by apache.

the class PartialGroupByKeyParDoFnsTest method testCreateWithCombinerAndStreaming.

@Test
public void testCreateWithCombinerAndStreaming() throws Exception {
    StreamingOptions options = PipelineOptionsFactory.as(StreamingOptions.class);
    options.setStreaming(true);
    Coder keyCoder = StringUtf8Coder.of();
    Coder valueCoder = BigEndianIntegerCoder.of();
    KvCoder<String, Integer> kvCoder = KvCoder.of(keyCoder, valueCoder);
    TestOutputReceiver receiver = new TestOutputReceiver(new ElementByteSizeObservableCoder(WindowedValue.getValueOnlyCoder(kvCoder)), counterSet, NameContextsForTests.nameContextForTest());
    ParDoFn pgbk = PartialGroupByKeyParDoFns.create(options, kvCoder, AppliedCombineFn.withInputCoder(Sum.ofIntegers(), CoderRegistry.createDefault(), kvCoder), NullSideInputReader.empty(), receiver, null);
    assertTrue(pgbk instanceof SimplePartialGroupByKeyParDoFn);
}

Example 29 with KvCoder

use of org.apache.beam.sdk.coders.KvCoder in project beam by apache.

the class PartialGroupByKeyParDoFnsTest method testCreateWithCombinerAndBatchSideInputs.

@Test
public void testCreateWithCombinerAndBatchSideInputs() throws Exception {
    PipelineOptions options = PipelineOptionsFactory.create();
    Coder keyCoder = StringUtf8Coder.of();
    Coder valueCoder = BigEndianIntegerCoder.of();
    KvCoder<String, Integer> kvCoder = KvCoder.of(keyCoder, valueCoder);
    TestOutputReceiver receiver = new TestOutputReceiver(new ElementByteSizeObservableCoder(WindowedValue.getValueOnlyCoder(kvCoder)), counterSet, NameContextsForTests.nameContextForTest());
    StepContext stepContext = BatchModeExecutionContext.forTesting(options, "testStage").getStepContext(TestOperationContext.create(counterSet));
    when(mockSideInputReader.isEmpty()).thenReturn(false);
    ParDoFn pgbk = PartialGroupByKeyParDoFns.create(options, kvCoder, AppliedCombineFn.withInputCoder(Sum.ofIntegers(), CoderRegistry.createDefault(), kvCoder, ImmutableList.<PCollectionView<?>>of(), WindowingStrategy.globalDefault()), mockSideInputReader, receiver, stepContext);
    assertTrue(pgbk instanceof BatchSideInputPGBKParDoFn);
}

Example 30 with KvCoder

use of org.apache.beam.sdk.coders.KvCoder in project beam by apache.

the class BatchSideInputHandlerFactory method forMultimapSideInput.

@Override
public <K, V, W extends BoundedWindow> MultimapSideInputHandler<K, V, W> forMultimapSideInput(String transformId, String sideInputId, KvCoder<K, V> elementCoder, Coder<W> windowCoder) {
    PCollectionNode collectionNode = sideInputToCollection.get(SideInputId.newBuilder().setTransformId(transformId).setLocalName(sideInputId).build());
    checkArgument(collectionNode != null, "No side input for %s/%s", transformId, sideInputId);
    Coder<K> keyCoder = elementCoder.getKeyCoder();
    Map<Object, Map<Object, KV<K, List<V>>>> /* structural key */
    data = new HashMap<>();
    List<WindowedValue<KV<K, V>>> broadcastVariable = sideInputGetter.getSideInput(collectionNode.getId());
    for (WindowedValue<KV<K, V>> windowedValue : broadcastVariable) {
        K key = windowedValue.getValue().getKey();
        V value = windowedValue.getValue().getValue();
        for (BoundedWindow boundedWindow : windowedValue.getWindows()) {
            @SuppressWarnings("unchecked") W window = (W) boundedWindow;
            Object structuralW = windowCoder.structuralValue(window);
            Object structuralK = keyCoder.structuralValue(key);
            KV<K, List<V>> records = data.computeIfAbsent(structuralW, o -> new HashMap<>()).computeIfAbsent(structuralK, o -> KV.of(key, new ArrayList<>()));
            records.getValue().add(value);
        }
    }
    return new MultimapSideInputHandler<K, V, W>() {

        @Override
        public Iterable<V> get(K key, W window) {
            KV<K, List<V>> records = data.getOrDefault(windowCoder.structuralValue(window), Collections.emptyMap()).get(keyCoder.structuralValue(key));
            if (records == null) {
                return Collections.emptyList();
            }
            return Collections.unmodifiableList(records.getValue());
        }

        @Override
        public Coder<V> valueCoder() {
            return elementCoder.getValueCoder();
        }

        @Override
        public Iterable<K> get(W window) {
            Map<Object, KV<K, List<V>>> records = data.getOrDefault(windowCoder.structuralValue(window), Collections.emptyMap());
            return Iterables.unmodifiableIterable(FluentIterable.concat(records.values()).transform(kListKV -> kListKV.getKey()));
        }

        @Override
        public Coder<K> keyCoder() {
            return elementCoder.getKeyCoder();
        }
    };
}