Examples with ETLConfig io.cdap.cdap.etl.proto.v2.ETLConfig used on opensource projects

Example 41 with ETLConfig

use of io.cdap.cdap.etl.proto.v2.ETLConfig in project cdap by caskdata.

the class DataPipelineTest method testServiceUrl.

private void testServiceUrl(Engine engine) throws Exception {
    // Deploy the ServiceApp application
    ApplicationManager appManager = deployApplication(ServiceApp.class);
    // Start Greeting service and use it
    ServiceManager serviceManager = appManager.getServiceManager(ServiceApp.Name.SERVICE_NAME).start();
    // Wait service startup
    serviceManager.waitForRun(ProgramRunStatus.RUNNING, 10, TimeUnit.SECONDS);
    URL url = new URL(serviceManager.getServiceURL(), "name");
    HttpRequest httpRequest = HttpRequest.post(url).withBody("bob").build();
    HttpResponse httpResponse = HttpRequests.execute(httpRequest, new DefaultHttpRequestConfig(false));
    Assert.assertEquals(HttpURLConnection.HTTP_OK, httpResponse.getResponseCode());
    url = new URL(serviceManager.getServiceURL(), "name/bob");
    HttpURLConnection connection = (HttpURLConnection) url.openConnection();
    Assert.assertEquals(HttpURLConnection.HTTP_OK, connection.getResponseCode());
    String response;
    try {
        response = new String(ByteStreams.toByteArray(connection.getInputStream()), Charsets.UTF_8);
    } finally {
        connection.disconnect();
    }
    Assert.assertEquals("bob", response);
    String sourceName = "ServiceUrlInput-" + engine.name();
    String sinkName = "ServiceUrlOutput-" + engine.name();
    /*
     * source --> filter --> sink
     */
    ETLBatchConfig etlConfig = ETLBatchConfig.builder().setEngine(engine).addStage(new ETLStage("source", MockSource.getPlugin(sourceName))).addStage(new ETLStage("filter", FilterTransform.getPlugin("name"))).addStage(new ETLStage("sink", MockSink.getPlugin(sinkName))).addConnection("source", "filter").addConnection("filter", "sink").build();
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app("ServiceUrl-" + engine);
    appManager = deployApplication(appId, appRequest);
    Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
    StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
    StructuredRecord recordBob = StructuredRecord.builder(schema).set("name", "bob").build();
    StructuredRecord recordJane = StructuredRecord.builder(schema).set("name", "jane").build();
    // write one record to each source
    DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(sourceName));
    MockSource.writeInput(inputManager, ImmutableList.of(recordSamuel, recordBob, recordJane));
    WorkflowManager workflowManager = appManager.getWorkflowManager(SmartWorkflow.NAME);
    workflowManager.start();
    workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
    // check output
    DataSetManager<Table> sinkManager = getDataset(sinkName);
    Set<StructuredRecord> expected = ImmutableSet.of(recordBob);
    Set<StructuredRecord> actual = Sets.newHashSet(MockSink.readOutput(sinkManager));
    Assert.assertEquals(expected, actual);
    serviceManager.stop();
    serviceManager.waitForRun(ProgramRunStatus.KILLED, 180, TimeUnit.SECONDS);
}

Example 42 with ETLConfig

use of io.cdap.cdap.etl.proto.v2.ETLConfig in project cdap by caskdata.

the class DataPipelineTest method testPipelineWithAllActions.

@Test
public void testPipelineWithAllActions() throws Exception {
    String actionTable = "actionTable";
    String action1RowKey = "action1.row";
    String action1ColumnKey = "action1.column";
    String action1Value = "action1.value";
    String action2RowKey = "action2.row";
    String action2ColumnKey = "action2.column";
    String action2Value = "action2.value";
    String action3RowKey = "action3.row";
    String action3ColumnKey = "action3.column";
    String action3Value = "action3.value";
    ETLBatchConfig etlConfig = ETLBatchConfig.builder().addStage(new ETLStage("action1", MockAction.getPlugin(actionTable, action1RowKey, action1ColumnKey, action1Value))).addStage(new ETLStage("action2", MockAction.getPlugin(actionTable, action2RowKey, action2ColumnKey, action2Value))).addStage(new ETLStage("action3", MockAction.getPlugin(actionTable, action3RowKey, action3ColumnKey, action3Value))).addConnection("action1", "action2").addConnection("action1", "action3").setEngine(Engine.MAPREDUCE).build();
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app("MyActionOnlyApp");
    ApplicationManager appManager = deployApplication(appId, appRequest);
    WorkflowManager workflowManager = appManager.getWorkflowManager(SmartWorkflow.NAME);
    workflowManager.start();
    workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
    DataSetManager<Table> actionTableDS = getDataset(actionTable);
    Assert.assertEquals(action1Value, MockAction.readOutput(actionTableDS, action1RowKey, action1ColumnKey));
    Assert.assertEquals(action2Value, MockAction.readOutput(actionTableDS, action2RowKey, action2ColumnKey));
    Assert.assertEquals(action3Value, MockAction.readOutput(actionTableDS, action3RowKey, action3ColumnKey));
    List<RunRecord> history = workflowManager.getHistory(ProgramRunStatus.COMPLETED);
    Assert.assertEquals(1, history.size());
    String runId = history.get(0).getPid();
    WorkflowTokenDetail tokenDetail = workflowManager.getToken(runId, WorkflowToken.Scope.USER, action1RowKey + action1ColumnKey);
    validateToken(tokenDetail, action1RowKey + action1ColumnKey, action1Value);
    tokenDetail = workflowManager.getToken(runId, WorkflowToken.Scope.USER, action2RowKey + action2ColumnKey);
    validateToken(tokenDetail, action2RowKey + action2ColumnKey, action2Value);
    tokenDetail = workflowManager.getToken(runId, WorkflowToken.Scope.USER, action3RowKey + action3ColumnKey);
    validateToken(tokenDetail, action3RowKey + action3ColumnKey, action3Value);
}

Example 43 with ETLConfig

use of io.cdap.cdap.etl.proto.v2.ETLConfig in project cdap by caskdata.

the class DataPipelineTest method testParallelAggregators.

private void testParallelAggregators(Engine engine) throws Exception {
    String source1Name = "pAggInput1-" + engine.name();
    String source2Name = "pAggInput2-" + engine.name();
    String sink1Name = "pAggOutput1-" + engine.name();
    String sink2Name = "pAggOutput2-" + engine.name();
    Schema inputSchema = Schema.recordOf("testRecord", Schema.Field.of("user", Schema.of(Schema.Type.STRING)), Schema.Field.of("item", Schema.of(Schema.Type.LONG)));
    /*
       source1 --|--> agg1 --> sink1
                 |
       source2 --|--> agg2 --> sink2
     */
    ETLBatchConfig etlConfig = ETLBatchConfig.builder().setEngine(engine).addStage(new ETLStage("source1", MockSource.getPlugin(source1Name, inputSchema))).addStage(new ETLStage("source2", MockSource.getPlugin(source2Name, inputSchema))).addStage(new ETLStage("sink1", MockSink.getPlugin(sink1Name))).addStage(new ETLStage("sink2", MockSink.getPlugin(sink2Name))).addStage(new ETLStage("agg1", FieldCountAggregator.getPlugin("user", "string"))).addStage(new ETLStage("agg2", FieldCountAggregator.getPlugin("item", "long"))).addConnection("source1", "agg1").addConnection("source1", "agg2").addConnection("source2", "agg1").addConnection("source2", "agg2").addConnection("agg1", "sink1").addConnection("agg2", "sink2").build();
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app("ParallelAggApp-" + engine);
    ApplicationManager appManager = deployApplication(appId, appRequest);
    // write few records to each source
    DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(source1Name));
    MockSource.writeInput(inputManager, ImmutableList.of(StructuredRecord.builder(inputSchema).set("user", "samuel").set("item", 1L).build(), StructuredRecord.builder(inputSchema).set("user", "samuel").set("item", 2L).build()));
    inputManager = getDataset(NamespaceId.DEFAULT.dataset(source2Name));
    MockSource.writeInput(inputManager, ImmutableList.of(StructuredRecord.builder(inputSchema).set("user", "samuel").set("item", 3L).build(), StructuredRecord.builder(inputSchema).set("user", "john").set("item", 4L).build(), StructuredRecord.builder(inputSchema).set("user", "john").set("item", 3L).build()));
    WorkflowManager workflowManager = appManager.getWorkflowManager(SmartWorkflow.NAME);
    workflowManager.start();
    workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
    Schema outputSchema1 = Schema.recordOf("user.count", Schema.Field.of("user", Schema.of(Schema.Type.STRING)), Schema.Field.of("ct", Schema.of(Schema.Type.LONG)));
    Schema outputSchema2 = Schema.recordOf("item.count", Schema.Field.of("item", Schema.of(Schema.Type.LONG)), Schema.Field.of("ct", Schema.of(Schema.Type.LONG)));
    // check output
    DataSetManager<Table> sinkManager = getDataset(sink1Name);
    Set<StructuredRecord> expected = ImmutableSet.of(StructuredRecord.builder(outputSchema1).set("user", "all").set("ct", 5L).build(), StructuredRecord.builder(outputSchema1).set("user", "samuel").set("ct", 3L).build(), StructuredRecord.builder(outputSchema1).set("user", "john").set("ct", 2L).build());
    Set<StructuredRecord> actual = Sets.newHashSet(MockSink.readOutput(sinkManager));
    Assert.assertEquals(expected, actual);
    sinkManager = getDataset(sink2Name);
    expected = ImmutableSet.of(StructuredRecord.builder(outputSchema2).set("item", 0L).set("ct", 5L).build(), StructuredRecord.builder(outputSchema2).set("item", 1L).set("ct", 1L).build(), StructuredRecord.builder(outputSchema2).set("item", 2L).set("ct", 1L).build(), StructuredRecord.builder(outputSchema2).set("item", 3L).set("ct", 2L).build(), StructuredRecord.builder(outputSchema2).set("item", 4L).set("ct", 1L).build());
    actual = Sets.newHashSet(MockSink.readOutput(sinkManager));
    Assert.assertEquals(expected, actual);
    validateMetric(2, appId, "source1.records.out");
    validateMetric(3, appId, "source2.records.out");
    validateMetric(5, appId, "agg1.records.in");
    // 2 users, but FieldCountAggregator always emits an 'all' group
    validateMetric(3, appId, "agg1.aggregator.groups");
    validateMetric(3, appId, "agg1.records.out");
    validateMetric(5, appId, "agg2.records.in");
    // 4 items, but FieldCountAggregator always emits an 'all' group
    validateMetric(5, appId, "agg2.aggregator.groups");
    validateMetric(5, appId, "agg2.records.out");
    validateMetric(3, appId, "sink1.records.in");
    validateMetric(5, appId, "sink2.records.in");
}

Example 44 with ETLConfig

use of io.cdap.cdap.etl.proto.v2.ETLConfig in project cdap by caskdata.

the class DataPipelineTest method testExternalDatasetTracking.

private void testExternalDatasetTracking(Engine engine, boolean backwardsCompatible) throws Exception {
    String suffix = engine.name() + (backwardsCompatible ? "-bc" : "");
    // Define input/output datasets
    String expectedExternalDatasetInput = "fileInput-" + suffix;
    String expectedExternalDatasetOutput = "fileOutput-" + suffix;
    // Define input/output directories
    File inputDir = TMP_FOLDER.newFolder("input-" + suffix);
    String inputFile = "input-file1.txt";
    File outputDir = TMP_FOLDER.newFolder("output-" + suffix);
    File outputSubDir1 = new File(outputDir, "subdir1");
    File outputSubDir2 = new File(outputDir, "subdir2");
    if (!backwardsCompatible) {
        // Assert that there are no external datasets
        Assert.assertNull(getDataset(NamespaceId.DEFAULT.dataset(expectedExternalDatasetInput)).get());
        Assert.assertNull(getDataset(NamespaceId.DEFAULT.dataset(expectedExternalDatasetOutput)).get());
    }
    String name = backwardsCompatible ? null : expectedExternalDatasetOutput;
    ETLBatchConfig.Builder builder = ETLBatchConfig.builder();
    ETLBatchConfig etlConfig = builder.setEngine(engine).addStage(new ETLStage("source", MockExternalSource.getPlugin(expectedExternalDatasetInput, inputDir.getAbsolutePath()))).addStage(new ETLStage("sink", MockExternalSink.getPlugin(name, "dir1", outputSubDir1.getAbsolutePath(), name, "dir2", outputSubDir2.getAbsolutePath()))).addConnection("source", "sink").build();
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app("ExternalDatasetApp-" + suffix);
    ApplicationManager appManager = deployApplication(appId, appRequest);
    Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
    StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
    StructuredRecord recordBob = StructuredRecord.builder(schema).set("name", "bob").build();
    StructuredRecord recordJane = StructuredRecord.builder(schema).set("name", "jane").build();
    ImmutableList<StructuredRecord> allInput = ImmutableList.of(recordSamuel, recordBob, recordJane);
    // Create input files
    MockExternalSource.writeInput(new File(inputDir, inputFile).getAbsolutePath(), allInput);
    Map<String, String> args = Collections.singletonMap(io.cdap.cdap.etl.common.Constants.CONSOLIDATE_STAGES, "true");
    WorkflowManager workflowManager = appManager.getWorkflowManager(SmartWorkflow.NAME);
    workflowManager.startAndWaitForGoodRun(args, ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
    List<RunRecord> history = workflowManager.getHistory();
    // there should be only one completed run
    Assert.assertEquals(1, history.size());
    Assert.assertEquals(ProgramRunStatus.COMPLETED, history.get(0).getStatus());
    // Assert output
    Assert.assertEquals(allInput, MockExternalSink.readOutput(outputSubDir1.getAbsolutePath(), schema));
    Assert.assertEquals(allInput, MockExternalSink.readOutput(outputSubDir2.getAbsolutePath(), schema));
    if (!backwardsCompatible) {
        // Assert that external datasets got created
        Assert.assertNotNull(getDataset(NamespaceId.DEFAULT.dataset(expectedExternalDatasetInput)).get());
        Assert.assertNotNull(getDataset(NamespaceId.DEFAULT.dataset(expectedExternalDatasetOutput)).get());
    }
}

Example 45 with ETLConfig

use of io.cdap.cdap.etl.proto.v2.ETLConfig in project cdap by caskdata.

the class DataPipelineTest method testMacroEvaluationActionPipeline.

private void testMacroEvaluationActionPipeline(Engine engine) throws Exception {
    ETLStage action1 = new ETLStage("action1", MockAction.getPlugin("actionTable", "action1.row", "action1.column", "${value}"));
    ETLBatchConfig etlConfig = ETLBatchConfig.builder().addStage(action1).setEngine(engine).build();
    // set runtime arguments for macro substitution
    Map<String, String> runtimeArguments = ImmutableMap.of("value", "macroValue");
    AppRequest<io.cdap.cdap.etl.proto.v2.ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app("macroActionTest-" + engine);
    ApplicationManager appManager = deployApplication(appId, appRequest);
    WorkflowManager manager = appManager.getWorkflowManager(SmartWorkflow.NAME);
    manager.setRuntimeArgs(runtimeArguments);
    manager.start(ImmutableMap.of("logical.start.time", "0"));
    manager.waitForRun(ProgramRunStatus.COMPLETED, 3, TimeUnit.MINUTES);
    DataSetManager<Table> actionTableDS = getDataset("actionTable");
    Assert.assertEquals("macroValue", MockAction.readOutput(actionTableDS, "action1.row", "action1.column"));
}