com.google.cloud.dataflow.sdk.transforms.View Java Examples

The following examples show how to use com.google.cloud.dataflow.sdk.transforms.View. You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may check out the related API usage on the sidebar.
Example #1
Source File: SideInputITCase.java    From flink-dataflow with Apache License 2.0 6 votes vote down vote up 
@Override
protected void testProgram() throws Exception {


	Pipeline p = FlinkTestPipeline.createForBatch();


	final PCollectionView<String> sidesInput = p
			.apply(Create.of(expected))
			.apply(View.<String>asSingleton());

	p.apply(Create.of("bli"))
			.apply(ParDo.of(new DoFn<String, String>() {
				@Override
				public void processElement(ProcessContext c) throws Exception {
					String s = c.sideInput(sidesInput);
					c.output(s);
				}
			}).withSideInputs(sidesInput)).apply(TextIO.Write.to(resultPath));

	p.run();
}
Example #2
Source File: DesiredStateEnforcer.java    From policyscanner with Apache License 2.0 5 votes vote down vote up 
private PCollection<String> constructPipeline(Pipeline pipeline, String org,
    BoundedSource<KV<List<String>, String>> knownGoodSource) {
  // Read files from GCS.
  PCollection<KV<List<String>, String>> knownGoodFiles =
      pipeline.apply("Read known-good data", Read.from(knownGoodSource));
  // Convert files to GCPResourceState objects.
  PCollection<KV<GCPResource, GCPResourceState>> knownGoodStates =
      knownGoodFiles.apply(ParDo.named("Convert file data to Java Objects")
          .of(new FileToState()));
  // Tag the state objects to indicate they're from a checked-in repo and not live.
  PCollection<KV<GCPResource, KV<StateSource, GCPResourceState>>> taggedKnownGoodStates =
      knownGoodStates.apply(ParDo.named("Mark states as being known-good")
          .of(new TagStateWithSource(StateSource.DESIRED)));

  // Read projects from the CRM API.
  PCollection<GCPProject> allProjects =
      pipeline.apply("Read live projects", Read.from(new LiveProjectSource(org)));
  // Extract project states.
  PCollection<KV<GCPResource, GCPResourceState>> liveStates =
      allProjects
          .apply(ParDo.named("Extract project policies").of(new ExtractState()));
  // Tag the states to indicate they're live and not from a checked-in source.
  PCollection<KV<GCPResource, KV<StateSource, GCPResourceState>>> taggedLiveStates =
      liveStates.apply(ParDo.named("Mark states as being live")
          .of(new TagStateWithSource(StateSource.LIVE)));

  // Join the two known-good and the live halves.
  PCollectionView<Map<GCPResource, KV<StateSource, GCPResourceState>>> knownGoodStatesView =
      taggedKnownGoodStates.apply(View.<GCPResource, KV<StateSource, GCPResourceState>>asMap());
  PCollection<KV<GCPResource, Map<StateSource, GCPResourceState>>> mismatchedStates =
      taggedLiveStates.apply(ParDo.named("Find states that don't match")
          .withSideInputs(knownGoodStatesView)
          .of(new FilterOutMatchingState(knownGoodStatesView)));

  // Construct an alert message for all the discrepancies found and fix the discrepancies.
  return mismatchedStates
      .apply(ParDo.named("Fix discrepancies").of(discrepancyAutoFixMessenger));
}
Example #3
Source File: OnDemandLiveStateChecker.java    From policyscanner with Apache License 2.0 5 votes vote down vote up 
private PCollection<String> constructPipeline(Pipeline pipeline,
    BoundedSource<KV<List<String>, String>> knownGoodSource) {
  // Read files from GCS.
  PCollection<KV<List<String>, String>> knownGoodFiles =
      pipeline.apply("Read known-good data", Read.from(knownGoodSource));
  // Convert files to GCPResourceState objects.
  PCollection<KV<GCPResource, GCPResourceState>> knownGoodStates =
      knownGoodFiles.apply(ParDo.named("Convert file data to Java objects")
          .of(new FileToState()));
  // Tag the state objects to indicate they're from a checked-in repo and not live.
  PCollection<KV<GCPResource, KV<StateSource, GCPResourceState>>> taggedKnownGoodStates =
      knownGoodStates.apply(ParDo.named("Mark states as being known-good")
          .of(new TagStateWithSource(StateSource.DESIRED)));

  // Extract a list of checked-in projects from GCS.
  PCollection<List<String>> allFilePaths = knownGoodFiles
      .apply("Extract just the file paths", ParDo.of(new FilePathFromPair()));
  // Read the live version of the states of the checked-in projects.
  PCollection<KV<GCPResource, GCPResourceState>> liveStates =
      allFilePaths.apply(ParDo.named("Get live resource and states from file path")
          .of(new FilePathToLiveState()));
  // Tag the states to indicate they're live and not from a checked-in source.
  PCollection<KV<GCPResource, KV<StateSource, GCPResourceState>>> taggedLiveStates =
      liveStates.apply(ParDo.named("Mark states as being live")
          .of(new TagStateWithSource(StateSource.LIVE)));

  // Join the two known-good and the live halves.
  PCollectionView<Map<GCPResource, KV<StateSource, GCPResourceState>>> liveStatesView =
      taggedLiveStates.apply(View.<GCPResource, KV<StateSource, GCPResourceState>>asMap());
  PCollection<KV<GCPResource, Map<StateSource, GCPResourceState>>> mismatchedStates =
      taggedKnownGoodStates.apply(ParDo.named("Find states that don't match")
          .withSideInputs(liveStatesView)
          .of(new FilterOutMatchingState(liveStatesView)));
  // Construct an alert message for all the discrepancies found.
  return mismatchedStates.apply(ParDo
      .named("Generate notification messages")
      .of(new StateDiscrepancyMessenger()));
}
Example #4
Source File: FilterOutMatchingStateTest.java    From policyscanner with Apache License 2.0 5 votes vote down vote up 
@Test
public void testFilterStateNoMatchingResources() {
  GCPProject checkedProject = getSampleProject("_checked");
  GCPProject liveProject = getSampleProject("_live");
  GCPResourceState checkedPolicy = getSampleGCPResourcePolicy(checkedProject, 1);
  GCPResourceState livePolicy = getSampleGCPResourcePolicy(liveProject, 2);
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> sideInputList =
      Arrays.asList(KV.of((GCPResource) checkedProject, KV.of(StateSource.DESIRED, checkedPolicy)));
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> mainInputList =
      Arrays.asList(KV.of((GCPResource) liveProject, KV.of(StateSource.LIVE, livePolicy)));

  PCollection<KV<GCPResource, KV<StateSource, GCPResourceState>>> collection =
      pipeline.apply(Create.of(sideInputList)).setCoder(
          KvCoder.of(SerializableCoder.of(GCPResource.class),
              KvCoder.of(SerializableCoder.of(StateSource.class),
                  SerializableCoder.of(GCPResourceState.class))));
  PCollectionView<Map<GCPResource, KV<StateSource, GCPResourceState>>> view =
      View.<GCPResource, KV<StateSource, GCPResourceState>>asMap().apply(collection);

  FilterOutMatchingState function = new FilterOutMatchingState(view);
  DoFnTester<KV<GCPResource, KV<StateSource, GCPResourceState>>,
      KV<GCPResource, Map<StateSource, GCPResourceState>>> tester = DoFnTester.of(function);
  tester.setSideInputInGlobalWindow(view, sideInputList);

  List<KV<GCPResource, Map<StateSource, GCPResourceState>>> results =
      tester.processBatch(mainInputList);
  assertEquals(0, results.size());
}
Example #5
Source File: FilterOutMatchingStateTest.java    From policyscanner with Apache License 2.0 5 votes vote down vote up 
@Test
public void testFilterStateNoMismatches() {
  GCPProject project = getSampleProject("");
  GCPResourceState checkedPolicy = getSampleGCPResourcePolicy(project, 1);
  GCPResourceState livePolicy = checkedPolicy;
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> sideInputList =
      Arrays.asList(KV.of((GCPResource) project, KV.of(StateSource.DESIRED, checkedPolicy)));
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> mainInputList =
      Arrays.asList(KV.of((GCPResource) project, KV.of(StateSource.LIVE, livePolicy)));

  PCollection<KV<GCPResource, KV<StateSource, GCPResourceState>>> collection =
      pipeline.apply(Create.of(sideInputList)).setCoder(
          KvCoder.of(SerializableCoder.of(GCPResource.class),
              KvCoder.of(SerializableCoder.of(StateSource.class),
                  SerializableCoder.of(GCPResourceState.class))));
  PCollectionView<Map<GCPResource, KV<StateSource, GCPResourceState>>> view =
      View.<GCPResource, KV<StateSource, GCPResourceState>>asMap().apply(collection);

  FilterOutMatchingState function = new FilterOutMatchingState(view);
  DoFnTester<KV<GCPResource, KV<StateSource, GCPResourceState>>,
      KV<GCPResource, Map<StateSource, GCPResourceState>>> tester = DoFnTester.of(function);
  tester.setSideInputInGlobalWindow(view, sideInputList);

  List<KV<GCPResource, Map<StateSource, GCPResourceState>>> results =
      tester.processBatch(mainInputList);
  assertEquals(0, results.size());
}
Example #6
Source File: FilterOutMatchingStateTest.java    From policyscanner with Apache License 2.0 5 votes vote down vote up 
@Test
public void testFilterStateOneMismatch() {
  GCPProject project = getSampleProject("");
  GCPResourceState checkedPolicy = getSampleGCPResourcePolicy(project, 1);
  GCPResourceState livePolicy = getSampleGCPResourcePolicy(project, 2);
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> sideInputList =
      Arrays.asList(KV.of((GCPResource) project, KV.of(StateSource.DESIRED, checkedPolicy)));
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> mainInputList =
      Arrays.asList(KV.of((GCPResource) project, KV.of(StateSource.LIVE, livePolicy)));

  PCollection<KV<GCPResource, KV<StateSource, GCPResourceState>>> collection =
      pipeline.apply(Create.of(sideInputList)).setCoder(
          KvCoder.of(SerializableCoder.of(GCPResource.class),
              KvCoder.of(SerializableCoder.of(StateSource.class),
                  SerializableCoder.of(GCPResourceState.class))));
  PCollectionView<Map<GCPResource, KV<StateSource, GCPResourceState>>> view =
      View.<GCPResource, KV<StateSource, GCPResourceState>>asMap().apply(collection);

  FilterOutMatchingState function = new FilterOutMatchingState(view);
  DoFnTester<KV<GCPResource, KV<StateSource, GCPResourceState>>,
      KV<GCPResource, Map<StateSource, GCPResourceState>>> tester = DoFnTester.of(function);
  tester.setSideInputInGlobalWindow(view, sideInputList);

  Map<StateSource, GCPResourceState> outputMap = new HashMap<>(2);
  outputMap.put(StateSource.DESIRED, checkedPolicy);
  outputMap.put(StateSource.LIVE, livePolicy);
  List<KV<GCPResource, Map<StateSource, GCPResourceState>>> expectedOutput =
      Arrays.asList(KV.of((GCPResource) project, outputMap));
  List<KV<GCPResource, Map<StateSource, GCPResourceState>>> results =
      tester.processBatch(mainInputList);

  assertEquals(expectedOutput, results);
}
Example #7
Source File: JoinKnownGoodAndLiveStatesTest.java    From policyscanner with Apache License 2.0 5 votes vote down vote up 
@Test
public void testFilterStateNoMatchingResources() {
  GCPProject checkedProject = getSampleProject("_checked");
  GCPProject liveProject = getSampleProject("_live");
  GCPResourceState checkedPolicy = getSampleGCPResourcePolicy(checkedProject, 1);
  GCPResourceState livePolicy = getSampleGCPResourcePolicy(liveProject, 2);
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> sideInputList =
      Arrays.asList(KV.of((GCPResource) checkedProject, KV.of(StateSource.DESIRED, checkedPolicy)));
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> mainInputList =
      Arrays.asList(KV.of((GCPResource) liveProject, KV.of(StateSource.LIVE, livePolicy)));

  PCollection<KV<GCPResource, KV<StateSource, GCPResourceState>>> collection =
      pipeline.apply(Create.of(sideInputList)).setCoder(
          KvCoder.of(SerializableCoder.of(GCPResource.class),
              KvCoder.of(SerializableCoder.of(StateSource.class),
                  SerializableCoder.of(GCPResourceState.class))));
  PCollectionView<Map<GCPResource, KV<StateSource, GCPResourceState>>> view =
      View.<GCPResource, KV<StateSource, GCPResourceState>>asMap().apply(collection);

  JoinKnownGoodAndLiveStates function = new JoinKnownGoodAndLiveStates(view);
  DoFnTester<KV<GCPResource, KV<StateSource, GCPResourceState>>,
      KV<GCPResource, Map<StateSource, GCPResourceState>>> tester = DoFnTester.of(function);
  tester.setSideInputInGlobalWindow(view, sideInputList);

  List<KV<GCPResource, Map<StateSource, GCPResourceState>>> results =
      tester.processBatch(mainInputList);
  assertEquals(0, results.size());
}
Example #8
Source File: JoinKnownGoodAndLiveStatesTest.java    From policyscanner with Apache License 2.0 5 votes vote down vote up 
@Test
public void testFilterStateNoMismatches() {
  GCPProject project = getSampleProject("");
  GCPResourceState checkedPolicy = getSampleGCPResourcePolicy(project, 1);
  GCPResourceState livePolicy = checkedPolicy;
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> sideInputList =
      Arrays.asList(KV.of((GCPResource) project, KV.of(StateSource.DESIRED, checkedPolicy)));
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> mainInputList =
      Arrays.asList(KV.of((GCPResource) project, KV.of(StateSource.LIVE, livePolicy)));

  PCollection<KV<GCPResource, KV<StateSource, GCPResourceState>>> collection =
      pipeline.apply(Create.of(sideInputList)).setCoder(
          KvCoder.of(SerializableCoder.of(GCPResource.class),
              KvCoder.of(SerializableCoder.of(StateSource.class),
                  SerializableCoder.of(GCPResourceState.class))));
  PCollectionView<Map<GCPResource, KV<StateSource, GCPResourceState>>> view =
      View.<GCPResource, KV<StateSource, GCPResourceState>>asMap().apply(collection);

  JoinKnownGoodAndLiveStates function = new JoinKnownGoodAndLiveStates(view);
  DoFnTester<KV<GCPResource, KV<StateSource, GCPResourceState>>,
      KV<GCPResource, Map<StateSource, GCPResourceState>>> tester = DoFnTester.of(function);
  tester.setSideInputInGlobalWindow(view, sideInputList);

  List<KV<GCPResource, Map<StateSource, GCPResourceState>>> results =
      tester.processBatch(mainInputList);
  assertEquals(1, results.size());
}
Example #9
Source File: JoinKnownGoodAndLiveStatesTest.java    From policyscanner with Apache License 2.0 5 votes vote down vote up 
@Test
public void testFilterStateOneMismatch() {
  GCPProject project = getSampleProject("");
  GCPResourceState checkedPolicy = getSampleGCPResourcePolicy(project, 1);
  GCPResourceState livePolicy = getSampleGCPResourcePolicy(project, 2);
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> sideInputList =
      Arrays.asList(KV.of((GCPResource) project, KV.of(StateSource.DESIRED, checkedPolicy)));
  List<KV<GCPResource, KV<StateSource, GCPResourceState>>> mainInputList =
      Arrays.asList(KV.of((GCPResource) project, KV.of(StateSource.LIVE, livePolicy)));

  PCollection<KV<GCPResource, KV<StateSource, GCPResourceState>>> collection =
      pipeline.apply(Create.of(sideInputList)).setCoder(
          KvCoder.of(SerializableCoder.of(GCPResource.class),
              KvCoder.of(SerializableCoder.of(StateSource.class),
                  SerializableCoder.of(GCPResourceState.class))));
  PCollectionView<Map<GCPResource, KV<StateSource, GCPResourceState>>> view =
      View.<GCPResource, KV<StateSource, GCPResourceState>>asMap().apply(collection);

  JoinKnownGoodAndLiveStates function = new JoinKnownGoodAndLiveStates(view);
  DoFnTester<KV<GCPResource, KV<StateSource, GCPResourceState>>,
      KV<GCPResource, Map<StateSource, GCPResourceState>>> tester = DoFnTester.of(function);
  tester.setSideInputInGlobalWindow(view, sideInputList);

  Map<StateSource, GCPResourceState> outputMap = new HashMap<>(2);
  outputMap.put(StateSource.DESIRED, checkedPolicy);
  outputMap.put(StateSource.LIVE, livePolicy);
  List<KV<GCPResource, Map<StateSource, GCPResourceState>>> expectedOutput =
      Arrays.asList(KV.of((GCPResource) project, outputMap));
  List<KV<GCPResource, Map<StateSource, GCPResourceState>>> results =
      tester.processBatch(mainInputList);

  assertEquals(expectedOutput, results);
}
Example #10
Source File: FlinkBatchTransformTranslators.java    From flink-dataflow with Apache License 2.0 4 votes vote down vote up 
@Override
public void translateNode(View.CreatePCollectionView<R, T> transform, FlinkBatchTranslationContext context) {
	DataSet<T> inputDataSet = context.getInputDataSet(context.getInput(transform));
	PCollectionView<T> input = transform.apply(null);
	context.setSideInputDataSet(input, inputDataSet);
}