Examples with MapReduceSpecification co.cask.cdap.api.mapreduce.MapReduceSpecification used on opensource projects

Example 6 with MapReduceSpecification

use of co.cask.cdap.api.mapreduce.MapReduceSpecification in project cdap by caskdata.

the class DefaultMapReduceConfigurer method createSpecification.

public MapReduceSpecification createSpecification() {
    Set<String> datasets = new HashSet<>();
    Reflections.visit(mapReduce, mapReduce.getClass(), new PropertyFieldExtractor(properties), new DataSetFieldExtractor(datasets));
    return new MapReduceSpecification(mapReduce.getClass().getName(), name, description, inputDataset, outputDataset, datasets, properties, driverResources, mapperResources, reducerResources);
}

Example 7 with MapReduceSpecification

use of co.cask.cdap.api.mapreduce.MapReduceSpecification in project cdap by caskdata.

the class ApplicationRegistrationStage method registerDatasets.

// Register dataset usage, based upon the program specifications.
// Note that worker specifications' datasets are not registered upon app deploy because the useDataset of the
// WorkerConfigurer is deprecated. Workers' access to datasets is aimed to be completely dynamic. Other programs are
// moving in this direction.
// Also, SparkSpecifications are the same in that a Spark program's dataset access is completely dynamic.
private void registerDatasets(ApplicationWithPrograms input) {
    ApplicationSpecification appSpec = input.getSpecification();
    ApplicationId appId = input.getApplicationId();
    NamespaceId namespaceId = appId.getParent();
    for (FlowSpecification flow : appSpec.getFlows().values()) {
        ProgramId programId = appId.flow(flow.getName());
        for (FlowletConnection connection : flow.getConnections()) {
            if (connection.getSourceType().equals(FlowletConnection.Type.STREAM)) {
                usageRegistry.register(programId, namespaceId.stream(connection.getSourceName()));
            }
        }
        for (FlowletDefinition flowlet : flow.getFlowlets().values()) {
            for (String dataset : flowlet.getDatasets()) {
                usageRegistry.register(programId, namespaceId.dataset(dataset));
            }
        }
    }
    for (MapReduceSpecification program : appSpec.getMapReduce().values()) {
        ProgramId programId = appId.mr(program.getName());
        for (String dataset : program.getDataSets()) {
            usageRegistry.register(programId, namespaceId.dataset(dataset));
        }
    }
    for (SparkSpecification sparkSpec : appSpec.getSpark().values()) {
        ProgramId programId = appId.spark(sparkSpec.getName());
        for (String dataset : sparkSpec.getDatasets()) {
            usageRegistry.register(programId, namespaceId.dataset(dataset));
        }
    }
    for (ServiceSpecification serviceSpecification : appSpec.getServices().values()) {
        ProgramId programId = appId.service(serviceSpecification.getName());
        for (HttpServiceHandlerSpecification handlerSpecification : serviceSpecification.getHandlers().values()) {
            for (String dataset : handlerSpecification.getDatasets()) {
                usageRegistry.register(programId, namespaceId.dataset(dataset));
            }
        }
    }
}

Example 8 with MapReduceSpecification

use of co.cask.cdap.api.mapreduce.MapReduceSpecification in project cdap by caskdata.

the class ApplicationSpecificationCodec method deserialize.

@Override
public ApplicationSpecification deserialize(JsonElement json, Type typeOfT, JsonDeserializationContext context) throws JsonParseException {
    JsonObject jsonObj = json.getAsJsonObject();
    String name = jsonObj.get("name").getAsString();
    String appVersion = ApplicationId.DEFAULT_VERSION;
    if (jsonObj.has("appVersion")) {
        appVersion = jsonObj.get("appVersion").getAsString();
    }
    String description = jsonObj.get("description").getAsString();
    String configuration = null;
    if (jsonObj.has("configuration")) {
        configuration = jsonObj.get("configuration").getAsString();
    }
    ArtifactId artifactId = context.deserialize(jsonObj.get("artifactId"), ArtifactId.class);
    Map<String, StreamSpecification> streams = deserializeMap(jsonObj.get("streams"), context, StreamSpecification.class);
    Map<String, String> datasetModules = deserializeMap(jsonObj.get("datasetModules"), context, String.class);
    Map<String, DatasetCreationSpec> datasetInstances = deserializeMap(jsonObj.get("datasetInstances"), context, DatasetCreationSpec.class);
    Map<String, FlowSpecification> flows = deserializeMap(jsonObj.get("flows"), context, FlowSpecification.class);
    Map<String, MapReduceSpecification> mapReduces = deserializeMap(jsonObj.get("mapReduces"), context, MapReduceSpecification.class);
    Map<String, SparkSpecification> sparks = deserializeMap(jsonObj.get("sparks"), context, SparkSpecification.class);
    Map<String, WorkflowSpecification> workflows = deserializeMap(jsonObj.get("workflows"), context, WorkflowSpecification.class);
    Map<String, ServiceSpecification> services = deserializeMap(jsonObj.get("services"), context, ServiceSpecification.class);
    Map<String, ScheduleSpecification> schedules = deserializeMap(jsonObj.get("schedules"), context, ScheduleSpecification.class);
    Map<String, ScheduleCreationSpec> programSchedules = deserializeMap(jsonObj.get("programSchedules"), context, ScheduleCreationSpec.class);
    Map<String, WorkerSpecification> workers = deserializeMap(jsonObj.get("workers"), context, WorkerSpecification.class);
    Map<String, Plugin> plugins = deserializeMap(jsonObj.get("plugins"), context, Plugin.class);
    return new DefaultApplicationSpecification(name, appVersion, description, configuration, artifactId, streams, datasetModules, datasetInstances, flows, mapReduces, sparks, workflows, services, schedules, programSchedules, workers, plugins);
}

Example 9 with MapReduceSpecification

use of co.cask.cdap.api.mapreduce.MapReduceSpecification in project cdap by caskdata.

the class DefaultAppConfigurer method addMapReduce.

@Override
public void addMapReduce(MapReduce mapReduce) {
    Preconditions.checkArgument(mapReduce != null, "MapReduce cannot be null.");
    DefaultMapReduceConfigurer configurer = new DefaultMapReduceConfigurer(mapReduce, deployNamespace, artifactId, artifactRepository, pluginInstantiator);
    mapReduce.configure(configurer);
    addDatasetsAndPlugins(configurer);
    MapReduceSpecification spec = configurer.createSpecification();
    mapReduces.put(spec.getName(), spec);
}

Example 10 with MapReduceSpecification

use of co.cask.cdap.api.mapreduce.MapReduceSpecification in project cdap by caskdata.

the class DistributedMapReduceProgramRunner method setupLaunchConfig.

@Override
protected void setupLaunchConfig(LaunchConfig launchConfig, Program program, ProgramOptions options, CConfiguration cConf, Configuration hConf, File tempDir) throws IOException {
    ApplicationSpecification appSpec = program.getApplicationSpecification();
    MapReduceSpecification spec = appSpec.getMapReduce().get(program.getName());
    // Get the resource for the container that runs the mapred client that will launch the actual mapred job.
    Map<String, String> clientArgs = RuntimeArguments.extractScope("task", "client", options.getUserArguments().asMap());
    Resources resources = SystemArguments.getResources(clientArgs, spec.getDriverResources());
    // Add runnable. Only one instance for the MR driver
    launchConfig.addRunnable(spec.getName(), new MapReduceTwillRunnable(spec.getName()), resources, 1).addExtraResources(MapReduceContainerHelper.localizeFramework(hConf, new HashMap<String, LocalizeResource>())).addExtraClasspath(MapReduceContainerHelper.addMapReduceClassPath(hConf, new ArrayList<String>())).addExtraDependencies(YarnClientProtocolProvider.class);
}