Search in sources :

Example 1 with BasicSparkExecutionPluginContext

use of io.cdap.cdap.etl.spark.batch.BasicSparkExecutionPluginContext in project cdap by caskdata.

the class DStreamCollection method compute.

@Override
public <U> SparkCollection<U> compute(StageSpec stageSpec, SparkCompute<T, U> compute) throws Exception {
    SparkCompute<T, U> wrappedCompute = new DynamicSparkCompute<>(new DynamicDriverContext(stageSpec, sec, new NoopStageStatisticsCollector()), compute);
    Transactionals.execute(sec, new TxRunnable() {

        @Override
        public void run(DatasetContext datasetContext) throws Exception {
            PipelineRuntime pipelineRuntime = new SparkPipelineRuntime(sec);
            SparkExecutionPluginContext sparkPluginContext = new BasicSparkExecutionPluginContext(sec, JavaSparkContext.fromSparkContext(stream.context().sparkContext()), datasetContext, pipelineRuntime, stageSpec);
            wrappedCompute.initialize(sparkPluginContext);
        }
    }, Exception.class);
    return wrap(stream.transform(new ComputeTransformFunction<>(sec, stageSpec, wrappedCompute)));
}
Also used : DynamicSparkCompute(io.cdap.cdap.etl.spark.streaming.function.DynamicSparkCompute) NoopStageStatisticsCollector(io.cdap.cdap.etl.common.NoopStageStatisticsCollector) ComputeTransformFunction(io.cdap.cdap.etl.spark.streaming.function.ComputeTransformFunction) PipelineRuntime(io.cdap.cdap.etl.common.PipelineRuntime) SparkPipelineRuntime(io.cdap.cdap.etl.spark.SparkPipelineRuntime) SparkPipelineRuntime(io.cdap.cdap.etl.spark.SparkPipelineRuntime) BasicSparkExecutionPluginContext(io.cdap.cdap.etl.spark.batch.BasicSparkExecutionPluginContext) BasicSparkExecutionPluginContext(io.cdap.cdap.etl.spark.batch.BasicSparkExecutionPluginContext) SparkExecutionPluginContext(io.cdap.cdap.etl.api.batch.SparkExecutionPluginContext) TxRunnable(io.cdap.cdap.api.TxRunnable) DatasetContext(io.cdap.cdap.api.data.DatasetContext)

Example 2 with BasicSparkExecutionPluginContext

use of io.cdap.cdap.etl.spark.batch.BasicSparkExecutionPluginContext in project cdap by caskdata.

the class DynamicSparkCompute method lazyInit.

// when checkpointing is enabled, and Spark is loading DStream operations from an existing checkpoint,
// delegate will be null and the initialize() method won't have been called. So we need to instantiate
// the delegate and initialize it.
private void lazyInit(final JavaSparkContext jsc) throws Exception {
    if (delegate == null) {
        PluginFunctionContext pluginFunctionContext = dynamicDriverContext.getPluginFunctionContext();
        delegate = pluginFunctionContext.createPlugin();
        final StageSpec stageSpec = pluginFunctionContext.getStageSpec();
        final JavaSparkExecutionContext sec = dynamicDriverContext.getSparkExecutionContext();
        Transactionals.execute(sec, new TxRunnable() {

            @Override
            public void run(DatasetContext datasetContext) throws Exception {
                PipelineRuntime pipelineRuntime = new SparkPipelineRuntime(sec);
                SparkExecutionPluginContext sparkPluginContext = new BasicSparkExecutionPluginContext(sec, jsc, datasetContext, pipelineRuntime, stageSpec);
                delegate.initialize(sparkPluginContext);
            }
        }, Exception.class);
    }
}
Also used : BasicSparkExecutionPluginContext(io.cdap.cdap.etl.spark.batch.BasicSparkExecutionPluginContext) PluginFunctionContext(io.cdap.cdap.etl.spark.function.PluginFunctionContext) BasicSparkExecutionPluginContext(io.cdap.cdap.etl.spark.batch.BasicSparkExecutionPluginContext) SparkExecutionPluginContext(io.cdap.cdap.etl.api.batch.SparkExecutionPluginContext) SparkPipelineRuntime(io.cdap.cdap.etl.spark.SparkPipelineRuntime) PipelineRuntime(io.cdap.cdap.etl.common.PipelineRuntime) SparkPipelineRuntime(io.cdap.cdap.etl.spark.SparkPipelineRuntime) TxRunnable(io.cdap.cdap.api.TxRunnable) StageSpec(io.cdap.cdap.etl.proto.v2.spec.StageSpec) JavaSparkExecutionContext(io.cdap.cdap.api.spark.JavaSparkExecutionContext) DatasetContext(io.cdap.cdap.api.data.DatasetContext)

Aggregations

TxRunnable (io.cdap.cdap.api.TxRunnable)2 DatasetContext (io.cdap.cdap.api.data.DatasetContext)2 SparkExecutionPluginContext (io.cdap.cdap.etl.api.batch.SparkExecutionPluginContext)2 PipelineRuntime (io.cdap.cdap.etl.common.PipelineRuntime)2 SparkPipelineRuntime (io.cdap.cdap.etl.spark.SparkPipelineRuntime)2 BasicSparkExecutionPluginContext (io.cdap.cdap.etl.spark.batch.BasicSparkExecutionPluginContext)2 JavaSparkExecutionContext (io.cdap.cdap.api.spark.JavaSparkExecutionContext)1 NoopStageStatisticsCollector (io.cdap.cdap.etl.common.NoopStageStatisticsCollector)1 StageSpec (io.cdap.cdap.etl.proto.v2.spec.StageSpec)1 PluginFunctionContext (io.cdap.cdap.etl.spark.function.PluginFunctionContext)1 ComputeTransformFunction (io.cdap.cdap.etl.spark.streaming.function.ComputeTransformFunction)1 DynamicSparkCompute (io.cdap.cdap.etl.spark.streaming.function.DynamicSparkCompute)1