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

Example 11 with ETLStage

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

the class DataPipelineTest method testMultipleOrderedInputActions.

@Test
public void testMultipleOrderedInputActions() throws Exception {
    Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
    /*
     * action1--->action2---|
     *                      |--> condition --> file ---> trueSink
     * action3--->action4---|      |
     *                             |--->file----> falseSink
     *
     */
    String appName = "MultipleOrderedInputActions";
    String trueSource = "true" + appName + "Source";
    String falseSource = "false" + appName + "Source";
    String trueSink = "true" + appName + "Sink";
    String falseSink = "false" + appName + "Sink";
    String actionTable = "actionTable" + appName;
    ETLBatchConfig etlConfig = ETLBatchConfig.builder("* * * * *").addStage(new ETLStage("trueSource", MockSource.getPlugin(trueSource, schema))).addStage(new ETLStage("falseSource", MockSource.getPlugin(falseSource, schema))).addStage(new ETLStage("trueSink", MockSink.getPlugin(trueSink))).addStage(new ETLStage("falseSink", MockSink.getPlugin(falseSink))).addStage(new ETLStage("condition", MockCondition.getPlugin("condition"))).addStage(new ETLStage("action1", MockAction.getPlugin(actionTable, "row1", "key1", "val1"))).addStage(new ETLStage("action2", MockAction.getPlugin(actionTable, "row2", "key2", "val2", "row1key1", "val1"))).addStage(new ETLStage("action3", MockAction.getPlugin(actionTable, "row3", "key3", "val3"))).addStage(new ETLStage("action4", MockAction.getPlugin(actionTable, "row4", "key4", "val4", "row3key3", "val3"))).addConnection("action1", "action2").addConnection("action3", "action4").addConnection("action2", "condition").addConnection("action4", "condition").addConnection("condition", "trueSource", true).addConnection("condition", "falseSource", false).addConnection("trueSource", "trueSink").addConnection("falseSource", "falseSink").build();
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT_RANGE, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app(appName);
    ApplicationManager appManager = deployApplication(appId, appRequest);
    StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
    StructuredRecord recordBob = StructuredRecord.builder(schema).set("name", "bob").build();
    for (String branch : Arrays.asList("true", "false")) {
        // write records to source
        String source = branch.equals("true") ? trueSource : falseSource;
        String sink = branch.equals("true") ? trueSink : falseSink;
        DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(source));
        MockSource.writeInput(inputManager, ImmutableList.of(recordSamuel, recordBob));
        WorkflowManager workflowManager = appManager.getWorkflowManager(SmartWorkflow.NAME);
        workflowManager.start(ImmutableMap.of("condition.branch.to.execute", branch));
        if (branch.equals("true")) {
            workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
        } else {
            workflowManager.waitForRuns(ProgramRunStatus.COMPLETED, 2, 5, TimeUnit.MINUTES);
        }
        // check sink
        DataSetManager<Table> sinkManager = getDataset(sink);
        Set<StructuredRecord> expected = ImmutableSet.of(recordSamuel, recordBob);
        Set<StructuredRecord> actual = Sets.newHashSet(MockSink.readOutput(sinkManager));
        Assert.assertEquals(expected, actual);
        validateMetric(2, appId, branch + "Source.records.out");
        validateMetric(2, appId, branch + "Sink.records.in");
        // check Action1 and Action2 is executed correctly
        DataSetManager<Table> actionTableDS = getDataset(actionTable);
        Assert.assertEquals("val1", MockAction.readOutput(actionTableDS, "row1", "key1"));
        Assert.assertEquals("val2", MockAction.readOutput(actionTableDS, "row2", "key2"));
        Assert.assertEquals("val3", MockAction.readOutput(actionTableDS, "row3", "key3"));
        Assert.assertEquals("val4", MockAction.readOutput(actionTableDS, "row4", "key4"));
    }
}

Example 12 with ETLStage

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

the class DataPipelineTest method testPipelineWithActions.

private void testPipelineWithActions(Engine engine) throws Exception {
    String actionTable = "actionTable-" + engine;
    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";
    String sourceName = "actionSource-" + engine;
    String sinkName = "actionSink-" + engine;
    String sourceTableName = "actionSourceTable-" + engine;
    String sinkTableName = "actionSinkTable-" + engine;
    Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
    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))).addStage(new ETLStage(sourceName, MockSource.getPlugin(sourceTableName, schema))).addStage(new ETLStage(sinkName, MockSink.getPlugin(sinkTableName))).addConnection(sourceName, sinkName).addConnection("action1", "action2").addConnection("action2", sourceName).addConnection(sinkName, "action3").setEngine(engine).build();
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app("MyApp-" + engine);
    ApplicationManager appManager = deployApplication(appId, appRequest);
    StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
    StructuredRecord recordBob = StructuredRecord.builder(schema).set("name", "bob").build();
    // write records to source
    DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(sourceTableName));
    MockSource.writeInput(inputManager, ImmutableList.of(recordSamuel, recordBob));
    WorkflowManager workflowManager = appManager.getWorkflowManager(SmartWorkflow.NAME);
    workflowManager.start();
    workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
    // check sink
    DataSetManager<Table> sinkManager = getDataset(sinkTableName);
    Set<StructuredRecord> expected = ImmutableSet.of(recordSamuel, recordBob);
    Set<StructuredRecord> actual = Sets.newHashSet(MockSink.readOutput(sinkManager));
    Assert.assertEquals(expected, actual);
    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));
    validateMetric(2, appId, sourceName + ".records.out");
    validateMetric(2, appId, sinkName + ".records.in");
    List<RunRecord> history = workflowManager.getHistory(ProgramRunStatus.COMPLETED);
    Assert.assertEquals(1, history.size());
    String runId = history.get(0).getPid();
    for (WorkflowToken.Scope scope : Arrays.asList(WorkflowToken.Scope.SYSTEM, WorkflowToken.Scope.USER)) {
        WorkflowTokenDetail token = workflowManager.getToken(runId, scope, null);
        for (Map.Entry<String, List<WorkflowTokenDetail.NodeValueDetail>> tokenData : token.getTokenData().entrySet()) {
            Assert.assertTrue(!tokenData.getKey().startsWith(co.cask.cdap.etl.common.Constants.StageStatistics.PREFIX));
        }
    }
}

Example 13 with ETLStage

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

the class DataPipelineTest method testRuntimeArgs.

private void testRuntimeArgs(Engine engine) throws Exception {
    String sourceName = "runtimeArgInput-" + engine;
    String sinkName = "runtimeArgOutput-" + engine;
    ETLBatchConfig etlConfig = ETLBatchConfig.builder("* * * * *").addStage(new ETLStage("action", MockAction.getPlugin("dumy", "val", "ue", "dwayne"))).addStage(new ETLStage("source", MockSource.getPlugin(sourceName))).addStage(new ETLStage("filter", StringValueFilterTransform.getPlugin("name", "samuel"))).addStage(new ETLStage("sink", MockSink.getPlugin(sinkName))).addConnection("action", "source").addConnection("source", "filter").addConnection("filter", "sink").setEngine(engine).build();
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app("RuntimeArgApp-" + engine);
    ApplicationManager appManager = deployApplication(appId, appRequest);
    // there should be only two programs - one workflow and one mapreduce/spark
    Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
    StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
    StructuredRecord recordDwayne = StructuredRecord.builder(schema).set("name", "dwayne").build();
    // write one record to each source
    DataSetManager<Table> inputManager = getDataset(sourceName);
    MockSource.writeInput(inputManager, ImmutableList.of(recordSamuel, recordDwayne));
    WorkflowManager workflowManager = appManager.getWorkflowManager(SmartWorkflow.NAME);
    workflowManager.start();
    workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
    // check sink
    DataSetManager<Table> sinkManager = getDataset(sinkName);
    Set<StructuredRecord> expected = ImmutableSet.of(recordSamuel);
    Set<StructuredRecord> actual = Sets.newHashSet(MockSink.readOutput(sinkManager));
    Assert.assertEquals(expected, actual);
}

Example 14 with ETLStage

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

the class DataPipelineTest method testSimpleCondition.

private void testSimpleCondition(Engine engine) throws Exception {
    Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
    /*
     * source --> condition --> trueSink
     *              |
     *              |-------> falseSink
     *
     */
    String appName = "SimpleCondition-" + engine;
    String source = appName + "Source-" + engine;
    String trueSink = "true" + appName + "Sink-" + engine;
    String falseSink = "false" + appName + "Sink-" + engine;
    String conditionTableName = "condition-" + engine;
    ETLBatchConfig etlConfig = ETLBatchConfig.builder("* * * * *").addStage(new ETLStage("source", MockSource.getPlugin(source, schema))).addStage(new ETLStage("trueSink", MockSink.getPlugin(trueSink))).addStage(new ETLStage("falseSink", MockSink.getPlugin(falseSink))).addStage(new ETLStage("condition", MockCondition.getPlugin("condition", conditionTableName))).addConnection("source", "condition").addConnection("condition", "trueSink", true).addConnection("condition", "falseSink", false).setEngine(engine).build();
    AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT_RANGE, etlConfig);
    ApplicationId appId = NamespaceId.DEFAULT.app(appName);
    ApplicationManager appManager = deployApplication(appId, appRequest);
    StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
    StructuredRecord recordBob = StructuredRecord.builder(schema).set("name", "bob").build();
    // write records to source
    DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(source));
    MockSource.writeInput(inputManager, ImmutableList.of(recordSamuel, recordBob));
    WorkflowManager workflowManager = null;
    for (String branch : Arrays.asList("true", "false")) {
        String sink = branch.equals("true") ? trueSink : falseSink;
        workflowManager = appManager.getWorkflowManager(SmartWorkflow.NAME);
        workflowManager.start(ImmutableMap.of("condition.branch.to.execute", branch));
        if (branch.equals("true")) {
            workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
        } else {
            workflowManager.waitForRuns(ProgramRunStatus.COMPLETED, 2, 5, TimeUnit.MINUTES);
        }
        // check sink
        DataSetManager<Table> sinkManager = getDataset(sink);
        Set<StructuredRecord> expected = ImmutableSet.of(recordSamuel, recordBob);
        Set<StructuredRecord> actual = Sets.newHashSet(MockSink.readOutput(sinkManager));
        Assert.assertEquals(expected, actual);
        // check condition table
        DataSetManager<Table> conditionTableDS = getDataset(conditionTableName);
        Assert.assertEquals("2", MockCondition.readOutput(conditionTableDS, "stats", "source.input.records"));
        Assert.assertEquals("2", MockCondition.readOutput(conditionTableDS, "stats", "source.output.records"));
        Assert.assertEquals("0", MockCondition.readOutput(conditionTableDS, "stats", "source.error.records"));
        validateMetric(branch.equals("true") ? 2 : 4, appId, "source.records.out");
        validateMetric(2, appId, branch + "Sink.records.in");
    }
    boolean foundStatisticsInToken = false;
    if (workflowManager != null) {
        List<RunRecord> history = workflowManager.getHistory(ProgramRunStatus.COMPLETED);
        // Checking for single run should be fine.
        String runId = history.get(0).getPid();
        for (WorkflowToken.Scope scope : Arrays.asList(WorkflowToken.Scope.SYSTEM, WorkflowToken.Scope.USER)) {
            WorkflowTokenDetail token = workflowManager.getToken(runId, scope, null);
            for (Map.Entry<String, List<WorkflowTokenDetail.NodeValueDetail>> tokenData : token.getTokenData().entrySet()) {
                if (tokenData.getKey().startsWith(co.cask.cdap.etl.common.Constants.StageStatistics.PREFIX)) {
                    foundStatisticsInToken = true;
                    break;
                }
            }
        }
    }
    Assert.assertTrue(foundStatisticsInToken);
}

Example 15 with ETLStage

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

the class DataPipelineTest method testServiceUrl.

public 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.waitForStatus(true);
    URL url = new URL(serviceManager.getServiceURL(), "name");
    HttpRequest httpRequest = HttpRequest.post(url).withBody("bob").build();
    HttpResponse httpResponse = HttpRequests.execute(httpRequest);
    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);
}