groovyx.gpars.dataflow.operator
[Java] Class DataflowProcessor

java.lang.Object
  groovyx.gpars.dataflow.operator.DataflowProcessor

@SuppressWarnings({"RawUseOfParameterizedType", "AccessingNonPublicFieldOfAnotherObject", "unchecked", "AbstractClassWithoutAbstractMethods", "ConstantDeclaredInAbstractClass"})
public abstract class DataflowProcessor

Dataflow selectors and operators (processors) form the basic units in dataflow networks. They are typically combined into oriented graphs that transform data. They accept a set of input and output dataflow channels and following specific strategies they transform input values from the input channels into new values written to the output channels. The output channels at the same time are suitable to be used as input channels by some other dataflow processors. The channels allow processors to communicate.

Dataflow selectors and operators enable creation of highly concurrent applications yet the abstraction hides the low-level concurrency primitives and exposes much friendlier API. Since selectors and operators internally leverage the actor implementation, they reuse a pool of threads and so the actual number of threads used by the calculation can be kept much lower than the actual number of processors used in the network.

Authors:

Vaclav Pech Date: Sep 9, 2009

INPUTS

public static final java.lang.String INPUTS

LISTENERS

private static final java.lang.String LISTENERS

MAX_FORKS

public static final java.lang.String MAX_FORKS

OUTPUTS

public static final java.lang.String OUTPUTS

STATE_OBJECT

public static final java.lang.String STATE_OBJECT

actor

protected DataflowProcessorActor actor

The internal actor performing on behalf of the processor

listeners

final java.util.Collection listeners

stateObject

protected final java.lang.Object stateObject

May hold custom state provided at construction time and read within the body

DataflowProcessor

protected DataflowProcessor(java.util.Map channels, groovy.lang.Closure code)

Creates a processor After creation the processor needs to be started using the start() method.

Parameters:

channels - A map specifying "inputs" and "outputs" - dataflow channels (instances of the DataflowQueue or DataflowVariable classes) to use for inputs and outputs

code - The processor's body to run each time all inputs have a value to read

addDataflowEventListener

public final void addDataflowEventListener(DataflowEventListener listener)

bindAllOutputValues

public final void bindAllOutputValues(java.lang.Object... values)

Used by the processor's body to send a value to all output channels. The values passed as arguments will each be sent to an output channel with identical position index.

If the maxForks value is set to a value greater than 1, calls to bindAllOutputs may result in values written to different channels to be in different order. If this is a problem for the application logic, the bindAllOutputsAtomically method should be considered instead.

Parameters:

values - Values to send to output channels of the same position index

bindAllOutputValuesAtomically

public final void bindAllOutputValuesAtomically(java.lang.Object... values)

Used by the processor's body to send a value to all output channels, while guaranteeing atomicity of the operation and preventing other calls to bindAllOutputsAtomically() from interfering with one another. The values passed as arguments will each be sent to an output channel with identical position index.

Parameters:

values - Values to send to output channels of the same position index

bindAllOutputs

public final void bindAllOutputs(java.lang.Object value)

Used by the processor's body to send a value to all output channels. If the maxForks value is set to a value greater than 1, calls to bindAllOutputs may result in values written to different channels to be in different order. If this is a problem for the application logic, the bindAllOutputsAtomically method should be considered instead.

Parameters:

value - The value to bind

bindAllOutputsAtomically

public final void bindAllOutputsAtomically(java.lang.Object value)

Used by the processor's body to send a value to all output channels, while guaranteeing atomicity of the operation and preventing other calls to bindAllOutputsAtomically() from interfering with one another.

Parameters:

value - The value to bind

bindOutput

public final void bindOutput(int idx, java.lang.Object value)

Used by the processor's body to send a value to the given output channel

Parameters:

idx - The index of the channel to bind

value - The value to bind

bindOutput

public final void bindOutput(java.lang.Object value)

Used by the processor's body to send a value to the first / only output channel

Parameters:

value - The value to bind

checkMaxForks

protected static void checkMaxForks(java.util.Map channels)

extractInputs

static java.util.List extractInputs(java.util.Map channels)

extractListeners

private static java.util.Collection extractListeners(java.util.Map channels)

extractOutputs

static java.util.List extractOutputs(java.util.Map channels)

extractState

private static java.lang.Object extractState(java.util.Map channels)

fireAfterRun

protected final void fireAfterRun(java.util.List messages)

fireAfterStart

protected final void fireAfterStart()

fireAfterStop

protected final void fireAfterStop()

fireBeforeRun

protected final java.util.List fireBeforeRun(java.util.List messages)

fireControlMessageArrived

protected final java.lang.Object fireControlMessageArrived(DataflowReadChannel channel, int index, java.lang.Object message)

fireCustomEvent

public final java.lang.Object fireCustomEvent(java.lang.Object data)

fireMessageArrived

protected final java.lang.Object fireMessageArrived(DataflowReadChannel channel, int index, java.lang.Object message)

fireMessageSentOut

protected final java.lang.Object fireMessageSentOut(DataflowWriteChannel channel, int index, java.lang.Object message)

fireOnException

protected final boolean fireOnException(java.lang.Throwable e)

getOutput

public final DataflowWriteChannel getOutput()

The processor's first / only output channel

Returns:

The particular DataflowWriteChannel instance

getOutputs

public final DataflowWriteChannel getOutputs(int idx)

The processor's output channel of the given index

Parameters:

idx - The index of the channel to retrieve

Returns:

The particular DataflowWriteChannel instance

getOutputs

public final java.util.List getOutputs()

The processor's all output channels

Returns:

A List holding all output channels

getStateObject

public final java.lang.Object getStateObject()

Retrieves the custom state object

Returns:

The state object associated with the operator

join

public final void join()

Joins the processor waiting for it to finish

throws:

InterruptedException If the thread gets interrupted

registerChannelListenersToAllInputs

public final void registerChannelListenersToAllInputs(DataflowChannelListener handler)

Registers the provided handler to all input channels

Parameters:

handler - The closure to invoke whenever a value gets bound to any of the input channels

removeDataflowEventListener

public final void removeDataflowEventListener(DataflowEventListener listener)

reportError

@SuppressWarnings({"UseOfSystemOutOrSystemErr"})
final void reportError(java.lang.Throwable e)

Is invoked in case the actor throws an exception.

Parameters:

e - The reported exception

shouldBeMultiThreaded

static boolean shouldBeMultiThreaded(java.util.Map channels)

start

public final DataflowProcessor start(PGroup group)

Starts a processor using the specified parallel group

Parameters:

group - The parallel group to use with the processor

Returns:

This operator instance

start

public final DataflowProcessor start()

Starts a processor using the specified parallel group

Returns:

This operator instance

terminate

public final void terminate()

Stops the processor immediately, potentially loosing unhandled messages

terminateAfterNextRun

public final void terminateAfterNextRun()

Gently stops the processor after the next set of messages is handled. Unlike with terminate(), no messages will get lost. If the operator never gets triggered after calling the terminateAfterNextRun() method, the operator never really stops.

Copyright © 2008–2012 Václav Pech. All Rights Reserved.