/** * Set all of the {@code inputs} as predecessors of the {@code operatorContext} each of the {@code outputs}. * * @param inputs input {@link ChannelInstance}s * @param executionLineageNode in-between {@link ExecutionLineageNode} * @param outputs output {@link ChannelInstance}s * @see #addPredecessor(LazyExecutionLineageNode) */ public static void connectAll(ChannelInstance[] inputs, ExecutionLineageNode executionLineageNode, ChannelInstance[] outputs) { for (ChannelInstance input : inputs) { if (input != null) executionLineageNode.addPredecessor(input.getLineage()); } for (ChannelInstance output : outputs) { if (output != null) output.getLineage().addPredecessor(executionLineageNode); } }
/** * Utility method to forward a {@link JavaChannelInstance} to another. * * @param input that should be forwarded * @param output to that should be forwarded */ static void forward(ChannelInstance input, ChannelInstance output) { // Do the forward. if (output instanceof CollectionChannel.Instance) { ((CollectionChannel.Instance) output).accept(((CollectionChannel.Instance) input).provideCollection()); } else if (output instanceof StreamChannel.Instance) { ((StreamChannel.Instance) output).accept(((JavaChannelInstance) input).provideStream()); } else { throw new RheemException(String.format("Cannot forward %s to %s.", input, output)); } // Manipulate the lineage. output.getLineage().addPredecessor(input.getLineage()); }
@Override public Tuple<Collection<ExecutionLineageNode>, Collection<ChannelInstance>> evaluate( ChannelInstance[] inputs, ChannelInstance[] outputs, JavaExecutor javaExecutor, OptimizationContext.OperatorContext operatorContext) { assert inputs.length == this.getNumInputs(); assert outputs.length == this.getNumOutputs(); String url = this.getInputUrl().trim(); FileSystem fs = FileSystems.getFileSystem(url).orElseThrow( () -> new RheemException(String.format("Cannot access file system of %s.", url)) ); try { final InputStream inputStream = fs.open(url); Stream<String> lines = new BufferedReader(new InputStreamReader(inputStream)).lines(); ((StreamChannel.Instance) outputs[0]).accept(lines); } catch (IOException e) { throw new RheemException(String.format("Reading %s failed.", url), e); } ExecutionLineageNode prepareLineageNode = new ExecutionLineageNode(operatorContext); prepareLineageNode.add(LoadProfileEstimators.createFromSpecification( "rheem.java.textfilesource.load.prepare", javaExecutor.getConfiguration() )); ExecutionLineageNode mainLineageNode = new ExecutionLineageNode(operatorContext); mainLineageNode.add(LoadProfileEstimators.createFromSpecification( "rheem.java.textfilesource.load.main", javaExecutor.getConfiguration() )); outputs[0].getLineage().addPredecessor(mainLineageNode); return prepareLineageNode.collectAndMark(); }
outputs[0].getLineage().addPredecessor(probingExecutionLineageNode);
/** * Utility method to forward a {@link RddChannel.Instance} to another. * * @param input that should be forwarded * @param output to that should be forwarded */ public void forward(ChannelInstance input, ChannelInstance output) { final RddChannel.Instance rddInput = (RddChannel.Instance) input; final RddChannel.Instance rddOutput = (RddChannel.Instance) output; // Do the forward. assert rddInput.getChannel().getDescriptor() == RddChannel.CACHED_DESCRIPTOR || rddOutput.getChannel().getDescriptor() != RddChannel.CACHED_DESCRIPTOR; rddOutput.accept(rddInput.provideRdd(), this); // Manipulate the lineage. output.getLineage().addPredecessor(input.getLineage()); }
output.getLineage().addPredecessor(probingExecutionLineageNode); return new Tuple<>(executionLineageNodes, producedChannelInstances);
"rheem.graphchi.pagerank.load.output", configuration )); outputChannelInstance.getLineage().addPredecessor(outputExecutionLineage);
@Override public Tuple<Collection<ExecutionLineageNode>, Collection<ChannelInstance>> evaluate( ChannelInstance[] inputs, ChannelInstance[] outputs, SparkExecutor sparkExecutor, OptimizationContext.OperatorContext operatorContext) { assert inputs.length == this.getNumInputs(); assert outputs.length == this.getNumOutputs(); RddChannel.Instance output = (RddChannel.Instance) outputs[0]; final JavaRDD<String> rdd = sparkExecutor.sc.textFile(this.getInputUrl()); this.name(rdd); output.accept(rdd, sparkExecutor); ExecutionLineageNode prepareLineageNode = new ExecutionLineageNode(operatorContext); prepareLineageNode.add(LoadProfileEstimators.createFromSpecification( "rheem.spark.textfilesource.load.prepare", sparkExecutor.getConfiguration() )); ExecutionLineageNode mainLineageNode = new ExecutionLineageNode(operatorContext); mainLineageNode.add(LoadProfileEstimators.createFromSpecification( "rheem.spark.textfilesource.load.main", sparkExecutor.getConfiguration() )); output.getLineage().addPredecessor(mainLineageNode); return prepareLineageNode.collectAndMark(); }
outputs[0].getLineage().addPredecessor(probingExecutionLineageNode);