Search in sources :

Example 1 with BagUserStateHandlerFactory

use of org.apache.beam.runners.fnexecution.state.StateRequestHandlers.BagUserStateHandlerFactory in project beam by apache.

the class RemoteExecutionTest method testExecutionWithUserStateCaching.

@Test
public void testExecutionWithUserStateCaching() throws Exception {
    Pipeline p = Pipeline.create();
    launchSdkHarness(p.getOptions());
    final String stateId = "foo";
    final String stateId2 = "bar";
    p.apply("impulse", Impulse.create()).apply("create", ParDo.of(new DoFn<byte[], KV<String, String>>() {

        @ProcessElement
        public void process(ProcessContext ctxt) {
        }
    })).setCoder(KvCoder.of(StringUtf8Coder.of(), StringUtf8Coder.of())).apply("userState", ParDo.of(new DoFn<KV<String, String>, KV<String, String>>() {

        @StateId(stateId)
        private final StateSpec<BagState<String>> bufferState = StateSpecs.bag(StringUtf8Coder.of());

        @StateId(stateId2)
        private final StateSpec<BagState<String>> bufferState2 = StateSpecs.bag(StringUtf8Coder.of());

        @ProcessElement
        public void processElement(@Element KV<String, String> element, @StateId(stateId) BagState<String> state, @StateId(stateId2) BagState<String> state2, OutputReceiver<KV<String, String>> r) {
            for (String value : state.read()) {
                r.output(KV.of(element.getKey(), value));
            }
            ReadableState<Boolean> isEmpty = state2.isEmpty();
            if (isEmpty.read()) {
                r.output(KV.of(element.getKey(), "Empty"));
            } else {
                state2.clear();
            }
        }
    })).apply("gbk", GroupByKey.create());
    RunnerApi.Pipeline pipelineProto = PipelineTranslation.toProto(p);
    FusedPipeline fused = GreedyPipelineFuser.fuse(pipelineProto);
    Optional<ExecutableStage> optionalStage = Iterables.tryFind(fused.getFusedStages(), (ExecutableStage stage) -> !stage.getUserStates().isEmpty());
    checkState(optionalStage.isPresent(), "Expected a stage with user state.");
    ExecutableStage stage = optionalStage.get();
    ExecutableProcessBundleDescriptor descriptor = ProcessBundleDescriptors.fromExecutableStage("test_stage", stage, dataServer.getApiServiceDescriptor(), stateServer.getApiServiceDescriptor());
    BundleProcessor processor = controlClient.getProcessor(descriptor.getProcessBundleDescriptor(), descriptor.getRemoteInputDestinations(), stateDelegator);
    Map<String, Coder> remoteOutputCoders = descriptor.getRemoteOutputCoders();
    Map<String, Collection<WindowedValue<?>>> outputValues = new HashMap<>();
    Map<String, RemoteOutputReceiver<?>> outputReceivers = new HashMap<>();
    for (Entry<String, Coder> remoteOutputCoder : remoteOutputCoders.entrySet()) {
        List<WindowedValue<?>> outputContents = Collections.synchronizedList(new ArrayList<>());
        outputValues.put(remoteOutputCoder.getKey(), outputContents);
        outputReceivers.put(remoteOutputCoder.getKey(), RemoteOutputReceiver.of((Coder<WindowedValue<?>>) remoteOutputCoder.getValue(), outputContents::add));
    }
    Map<String, List<ByteString>> userStateData = ImmutableMap.of(stateId, new ArrayList(Arrays.asList(ByteString.copyFrom(CoderUtils.encodeToByteArray(StringUtf8Coder.of(), "A", Coder.Context.NESTED)), ByteString.copyFrom(CoderUtils.encodeToByteArray(StringUtf8Coder.of(), "B", Coder.Context.NESTED)), ByteString.copyFrom(CoderUtils.encodeToByteArray(StringUtf8Coder.of(), "C", Coder.Context.NESTED)))), stateId2, new ArrayList(Arrays.asList(ByteString.copyFrom(CoderUtils.encodeToByteArray(StringUtf8Coder.of(), "D", Coder.Context.NESTED)))));
    StoringStateRequestHandler stateRequestHandler = new StoringStateRequestHandler(StateRequestHandlers.forBagUserStateHandlerFactory(descriptor, new BagUserStateHandlerFactory<ByteString, Object, BoundedWindow>() {

        @Override
        public BagUserStateHandler<ByteString, Object, BoundedWindow> forUserState(String pTransformId, String userStateId, Coder<ByteString> keyCoder, Coder<Object> valueCoder, Coder<BoundedWindow> windowCoder) {
            return new BagUserStateHandler<ByteString, Object, BoundedWindow>() {

                @Override
                public Iterable<Object> get(ByteString key, BoundedWindow window) {
                    return (Iterable) userStateData.get(userStateId);
                }

                @Override
                public void append(ByteString key, BoundedWindow window, Iterator<Object> values) {
                    Iterators.addAll(userStateData.get(userStateId), (Iterator) values);
                }

                @Override
                public void clear(ByteString key, BoundedWindow window) {
                    userStateData.get(userStateId).clear();
                }
            };
        }
    }));
    try (RemoteBundle bundle = processor.newBundle(outputReceivers, stateRequestHandler, BundleProgressHandler.ignored())) {
        Iterables.getOnlyElement(bundle.getInputReceivers().values()).accept(valueInGlobalWindow(KV.of("X", "Y")));
    }
    try (RemoteBundle bundle2 = processor.newBundle(outputReceivers, stateRequestHandler, BundleProgressHandler.ignored())) {
        Iterables.getOnlyElement(bundle2.getInputReceivers().values()).accept(valueInGlobalWindow(KV.of("X", "Z")));
    }
    for (Collection<WindowedValue<?>> windowedValues : outputValues.values()) {
        assertThat(windowedValues, containsInAnyOrder(valueInGlobalWindow(KV.of("X", "A")), valueInGlobalWindow(KV.of("X", "B")), valueInGlobalWindow(KV.of("X", "C")), valueInGlobalWindow(KV.of("X", "A")), valueInGlobalWindow(KV.of("X", "B")), valueInGlobalWindow(KV.of("X", "C")), valueInGlobalWindow(KV.of("X", "Empty"))));
    }
    assertThat(userStateData.get(stateId), IsIterableContainingInOrder.contains(ByteString.copyFrom(CoderUtils.encodeToByteArray(StringUtf8Coder.of(), "A", Coder.Context.NESTED)), ByteString.copyFrom(CoderUtils.encodeToByteArray(StringUtf8Coder.of(), "B", Coder.Context.NESTED)), ByteString.copyFrom(CoderUtils.encodeToByteArray(StringUtf8Coder.of(), "C", Coder.Context.NESTED))));
    assertThat(userStateData.get(stateId2), IsEmptyIterable.emptyIterable());
    // 3 Requests expected: state read, state2 read, and state2 clear
    assertEquals(3, stateRequestHandler.getRequestCount());
    ByteString.Output out = ByteString.newOutput();
    StringUtf8Coder.of().encode("X", out);
    assertEquals(stateId, stateRequestHandler.receivedRequests.get(0).getStateKey().getBagUserState().getUserStateId());
    assertEquals(stateRequestHandler.receivedRequests.get(0).getStateKey().getBagUserState().getKey(), out.toByteString());
    assertTrue(stateRequestHandler.receivedRequests.get(0).hasGet());
    assertEquals(stateId2, stateRequestHandler.receivedRequests.get(1).getStateKey().getBagUserState().getUserStateId());
    assertEquals(stateRequestHandler.receivedRequests.get(1).getStateKey().getBagUserState().getKey(), out.toByteString());
    assertTrue(stateRequestHandler.receivedRequests.get(1).hasGet());
    assertEquals(stateId2, stateRequestHandler.receivedRequests.get(2).getStateKey().getBagUserState().getUserStateId());
    assertEquals(stateRequestHandler.receivedRequests.get(2).getStateKey().getBagUserState().getKey(), out.toByteString());
    assertTrue(stateRequestHandler.receivedRequests.get(2).hasClear());
}
Also used : IsEmptyIterable(org.hamcrest.collection.IsEmptyIterable) ConcurrentHashMap(java.util.concurrent.ConcurrentHashMap) HashMap(java.util.HashMap) ByteString(org.apache.beam.vendor.grpc.v1p43p2.com.google.protobuf.ByteString) ArrayList(java.util.ArrayList) ByteString(org.apache.beam.vendor.grpc.v1p43p2.com.google.protobuf.ByteString) RunnerApi(org.apache.beam.model.pipeline.v1.RunnerApi) BundleProcessor(org.apache.beam.runners.fnexecution.control.SdkHarnessClient.BundleProcessor) WindowedValue(org.apache.beam.sdk.util.WindowedValue) Iterator(java.util.Iterator) BoundedWindow(org.apache.beam.sdk.transforms.windowing.BoundedWindow) ArrayList(java.util.ArrayList) PCollectionList(org.apache.beam.sdk.values.PCollectionList) List(java.util.List) ExecutableStage(org.apache.beam.runners.core.construction.graph.ExecutableStage) BagState(org.apache.beam.sdk.state.BagState) ExecutableProcessBundleDescriptor(org.apache.beam.runners.fnexecution.control.ProcessBundleDescriptors.ExecutableProcessBundleDescriptor) KvCoder(org.apache.beam.sdk.coders.KvCoder) Coder(org.apache.beam.sdk.coders.Coder) StringUtf8Coder(org.apache.beam.sdk.coders.StringUtf8Coder) BigEndianLongCoder(org.apache.beam.sdk.coders.BigEndianLongCoder) BagUserStateHandlerFactory(org.apache.beam.runners.fnexecution.state.StateRequestHandlers.BagUserStateHandlerFactory) BagUserStateHandler(org.apache.beam.runners.fnexecution.state.StateRequestHandlers.BagUserStateHandler) FusedPipeline(org.apache.beam.runners.core.construction.graph.FusedPipeline) ReadableState(org.apache.beam.sdk.state.ReadableState) KV(org.apache.beam.sdk.values.KV) Pipeline(org.apache.beam.sdk.Pipeline) FusedPipeline(org.apache.beam.runners.core.construction.graph.FusedPipeline) Collection(java.util.Collection) PCollection(org.apache.beam.sdk.values.PCollection) Test(org.junit.Test)

Aggregations

ArrayList (java.util.ArrayList)1 Collection (java.util.Collection)1 HashMap (java.util.HashMap)1 Iterator (java.util.Iterator)1 List (java.util.List)1 ConcurrentHashMap (java.util.concurrent.ConcurrentHashMap)1 RunnerApi (org.apache.beam.model.pipeline.v1.RunnerApi)1 ExecutableStage (org.apache.beam.runners.core.construction.graph.ExecutableStage)1 FusedPipeline (org.apache.beam.runners.core.construction.graph.FusedPipeline)1 ExecutableProcessBundleDescriptor (org.apache.beam.runners.fnexecution.control.ProcessBundleDescriptors.ExecutableProcessBundleDescriptor)1 BundleProcessor (org.apache.beam.runners.fnexecution.control.SdkHarnessClient.BundleProcessor)1 BagUserStateHandler (org.apache.beam.runners.fnexecution.state.StateRequestHandlers.BagUserStateHandler)1 BagUserStateHandlerFactory (org.apache.beam.runners.fnexecution.state.StateRequestHandlers.BagUserStateHandlerFactory)1 Pipeline (org.apache.beam.sdk.Pipeline)1 BigEndianLongCoder (org.apache.beam.sdk.coders.BigEndianLongCoder)1 Coder (org.apache.beam.sdk.coders.Coder)1 KvCoder (org.apache.beam.sdk.coders.KvCoder)1 StringUtf8Coder (org.apache.beam.sdk.coders.StringUtf8Coder)1 BagState (org.apache.beam.sdk.state.BagState)1 ReadableState (org.apache.beam.sdk.state.ReadableState)1