JAVA API :

PreviewContext

interface PreviewContext {
   
	 // all records from the stage identified by identifier    
	<T> Iterator<T> getOutputRecords();
   
	 // all metrics from the stage identified by identifier
	Collection<MetricTimeSeries> getMetrics();
  
 	 // all logs from the stage identified by identifier
	String getLogs();
}

PreviewEmitter

interface PreviewEmitter {	
	void emit(T record);
}

class DefaultPreview implements PreviewContext, PreviewEmitter {
   // implementation for methods, emit, getEmittedRecords, getLogs and getMetrics.
}

@Singleton
class PreviewHelper {
	Map<PreviewId, DefaultPreview> previewIdToDefaultPreview;
 
	PreviewEmitter getPreviewEmitter(PreviewId identifier) {
		// create DefaultPreview for identifier in previewIdToDefaultPreview and return if doesn't exist already, else get and return.	
	}
 
	PreviewContext getPreviewContext(PreviewId identifier) {
		// create DefaultPreview for identifier in previewIdToDefaultPreview and return if doesn't exist already, else get and return.
	}
}

Implementation:

SDK:

Preview Execution Isolation:

Requirement:

We want the program runs we execute, datasets created during preview for preview purpose, logs and metrics emitted during preview to be isolated from the regular Standalone execution which is used to publish and run the pipeline.
In Preview, pipeline could have lookup datasets in a transform which reads from the datasets in Standalone. so we want a way to share datasets in preview with datasets in standalone.
In Preview, we want to skip writing meta data and lineage information as they are unnecessary.

Preview Injector vs Standalone Injector:

Service	Standalone (Yes/No)	Preview (Yes/No)
userInterfaceService	Yes	No
trackerAppCreationService	Yes	No
router	Yes	No
streamService	Yes	Yes
exploreExecutorService	Yes	No
exploreClient	Yes	No
metadataService	Yes	No
serviceStore (set/get service instances)	Yes	No
appFabricServer	Yes	No
previewServer	No	Yes
datasetService	Yes	Yes
metricsQueryService	Yes	No (Can call MetricStore query)
txService	Yes	Yes
externalAuthenticationServer (if security enabled)	Yes	Yes
logAppenderInitializer	Yes	Yes
kafkaClient(if audit enabled)	Yes	No
zkClient (if audit enabled)	Yes	No
authorizerInstantiator (started by default)	Yes	Yes?

AppFabricServer vs PreviewServer :

This is a subset of services started in app-fabric server.

Services	AppFabricServer	PreviewServer
notificationService	Yes	No
schedulerService	Yes	No
applicationLifecycleService	Yes	Yes
systemArtifactLoader	Yes	Yes
programRuntimeService	Yes	Yes
streamCoordinatorClient	Yes	Yes
programLifecycleService	Yes	Yes
pluginService	Yes	Yes
httpService	Yes	Yes (but only with preview handler).

PreviewDatasetFramework

Requirements:

1) Pipeline want's to read from a dataset source (or) pipeline wants to write to a dataset sink (or) transform uses a lookup table. These datasets are in CDAP Standalone space.

2) Pipeline run's records, Pipeline run metrics, program status, etc are stored in System datasets in Preview space.

3) Error dataset : Its not clear if using error dataset should cause creating an error dataset in CDAP standalone space. I feel it might not be required to created in Standalone space. In which case if its a dataset then it's the only user level dataset that has to be created in Preview space, we can say we would have an in-memory implementation for maintaining error records.

Assumptions :

1) All Datasets in System Namespace will be using the "LocalDatasetFramework"

2) All Datasets in User's Namespaces will be using the "RemoteDatasetFramework"

PreviewDatasetFramework

... snippet
@Nullable
@Override
public <T extends Dataset> T getDataset(Id.DatasetInstance datasetInstanceId,
                                        @Nullable Map<String, String> arguments,
                                        @Nullable ClassLoader classLoader)
  throws DatasetManagementException, IOException {
  if (datasetInstanceId.getNamespace().equals(Id.Namespace.SYSTEM)) {
    return localDatasetFramework.getDataset(datasetInstanceId, arguments, classLoader);
  } else {
    return remoteDatasetFramework.getDataset(datasetInstanceId, arguments, classLoader);
  }
}

Adapting to Cluster

Having a LocalDatasetFramework for system namespace would make it useful for adapting to cluster, where the container's local directory will be used to store the system datasets and we can use the RemoteDatasetFramework of CDAP master for datasets in other namespace.

Java API Design and Implementation Details.

SDK:

Preview Execution Isolation:

Requirement:

PreviewDatasetFramework

Adapting to Cluster