org.nuxeo.lib.stream.computation

Interface Computation

All Known Implementing Classes:

AbstractBatchComputation, AbstractBulkComputation, AbstractComputation, AbstractTransientBlobComputation, AutomationBulkAction.AutomationComputation, BulkIndexComputation, BulkScrollerComputation, BulkStatusComputation, CSVProjectionComputation, ExposeBlob, IndexCompletionComputation, IndexRequestComputation, LatencyMonitorComputation, LatencyTrackerComputation, MakeBlob, RemoveProxyAction.RemoveProxyComputation, SetPropertiesAction.SetPropertyComputation, SetSystemPropertiesAction.SetSystemPropertyComputation, SortBlob, StreamAuditStorageWriter.AuditStorageLogWriterComputation, StreamAuditWriter.AuditLogWriterComputation, TrashAction.TrashComputation, WorkComputation, ZipBlob

public interface Computation

Computation receives records from input streams one at a time, it can produce record on its output streams. A timer processing can be used for windowing computation.

Since:

9.3

Method Summary

Modifier and Type	Method and Description
default void	destroy() Called when the framework is ready to shutdown the computation.
void	init(ComputationContext context) Called when the framework has registered the computation successfully.
ComputationMetadata	metadata() Identify the computation.
void	processFailure(ComputationContext context, Throwable failure) Called when processRecord(org.nuxeo.lib.stream.computation.ComputationContext, java.lang.String, org.nuxeo.lib.stream.computation.Record) or processTimer(org.nuxeo.lib.stream.computation.ComputationContext, java.lang.String, long) fails and cannot be retried.
void	processRecord(ComputationContext context, String inputStreamName, Record record) Process an incoming record on one of the computation's input streams.
void	processRetry(ComputationContext context, Throwable failure) Called after a failure in processRecord(org.nuxeo.lib.stream.computation.ComputationContext, java.lang.String, org.nuxeo.lib.stream.computation.Record) or processTimer(org.nuxeo.lib.stream.computation.ComputationContext, java.lang.String, long) before retrying.
void	processTimer(ComputationContext context, String key, long timestamp) Process a timer callback previously set via ComputationContext.setTimer(String, long).
default void	signalStop() A hook to inform that computation will be soon destroyed.

Method Detail

init

void init(ComputationContext context)

Called when the framework has registered the computation successfully. Gives users a first opportunity to schedule timer callbacks and produce records.

Parameters:

context - The computation context object provided by the system.

destroy

default void destroy()

Called when the framework is ready to shutdown the computation. Gives users a chance to perform some cleanup before the process is killed.

processRecord

void processRecord(ComputationContext context,
                   String inputStreamName,
                   Record record)

Process an incoming record on one of the computation's input streams.

Parameters:

context - The computation context object provided by the system.

inputStreamName - Name of the input stream that provides the record.

record - The record.

processTimer

void processTimer(ComputationContext context,
                  String key,
                  long timestamp)

Process a timer callback previously set via ComputationContext.setTimer(String, long).

Parameters:

context - The computation context object provided by the system.

key - The name of the timer.

timestamp - The timestamp (in ms) for which the callback was scheduled.

metadata

ComputationMetadata metadata()

Identify the computation.

Returns:

computation's metadata.

signalStop

default void signalStop()

A hook to inform that computation will be soon destroyed. It gives a way for long processing to cooperate to a quick shutdown.
This method is not invoked from the computation thread, it should only set some volatile flag and returns immediately.

Since:

10.2

processRetry

void processRetry(ComputationContext context,
                  Throwable failure)

Called after a failure in processRecord(org.nuxeo.lib.stream.computation.ComputationContext, java.lang.String, org.nuxeo.lib.stream.computation.Record) or processTimer(org.nuxeo.lib.stream.computation.ComputationContext, java.lang.String, long) before retrying.

Since:

10.3

processFailure

void processFailure(ComputationContext context,
                    Throwable failure)

Called when processRecord(org.nuxeo.lib.stream.computation.ComputationContext, java.lang.String, org.nuxeo.lib.stream.computation.Record) or processTimer(org.nuxeo.lib.stream.computation.ComputationContext, java.lang.String, long) fails and cannot be retried.

Since:

10.3