org.apache.giraph.graph

Class AbstractComputation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

java.lang.Object

org.apache.giraph.conf.DefaultImmutableClassesGiraphConfigurable<I,V,E>

org.apache.giraph.worker.WorkerAggregatorDelegator<I,V,E>

org.apache.giraph.graph.AbstractComputation<I,V,E,M1,M2>

Type Parameters:

I - Vertex id

V - Vertex data

E - Edge data

M1 - Incoming message type

M2 - Outgoing message type

All Implemented Interfaces:

AggregatorUsage, GiraphConfigurationSettable<I,V,E>, ImmutableClassesGiraphConfigurable<I,V,E>, TypesHolder<I,V,E,M1,M2>, Computation<I,V,E,M1,M2>, WorkerAggregatorUsage, WorkerBroadcastUsage, WorkerGlobalCommUsage, WorkerIndexUsage<I>, WorkerReduceUsage

Direct Known Subclasses:

BasicComputation, JythonGiraphComputation

public abstract class AbstractComputation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>
extends WorkerAggregatorDelegator<I,V,E>
implements Computation<I,V,E,M1,M2>

See Computation for explanation of the interface. This is a abstract class helper for users to implement their computations. It implements all of the methods required by the Computation interface except for the compute(Vertex, Iterable) which we leave to the user to define. In most cases users should inherit from this class when implementing their algorithms with Giraph.

Constructor Summary

Constructors

Constructor and Description
AbstractComputation()

Method Summary

Modifier and Type	Method and Description
void	addEdgeRequest(I sourceVertexId, Edge<I,E> edge) Request to add an edge of a vertex in the graph (processed just prior to the next superstep)
void	addVertexRequest(I id, V value) Sends a request to create a vertex that will be available during the next superstep.
void	addVertexRequest(I id, V value, OutEdges<I,E> edges) Sends a request to create a vertex that will be available during the next superstep.
abstract void	compute(Vertex<I,V,E> vertex, Iterable<M1> messages) Must be defined by user to do computation on a single Vertex.
org.apache.hadoop.mapreduce.Mapper.Context	getContext() Get the mapper context
int	getMyWorkerIndex() Get index for this worker
long	getSuperstep() Retrieves the current superstep.
long	getTotalNumEdges() Get the total (all workers) number of edges that existed in the previous superstep.
long	getTotalNumVertices() Get the total (all workers) number of vertices that existed in the previous superstep.
<W extends WorkerContext> W	getWorkerContext() Get the worker context
int	getWorkerCount() Get number of workers
int	getWorkerForVertex(I vertexId) Get worker index which will contain vertex with given id, if such vertex exists.
void	initialize(GraphState graphState, WorkerClientRequestProcessor<I,V,E> workerClientRequestProcessor, CentralizedServiceWorker<I,V,E> serviceWorker, WorkerGlobalCommUsage workerGlobalCommUsage) Initialize, called by infrastructure before the superstep starts.
void	postSuperstep() Finish computation.
void	preSuperstep() Prepare for computation.
void	removeEdgesRequest(I sourceVertexId, I targetVertexId) Request to remove all edges from a given source vertex to a given target vertex (processed just prior to the next superstep).
void	removeVertexRequest(I vertexId) Request to remove a vertex from the graph (applied just prior to the next superstep).
void	sendMessage(I id, M2 message) Send a message to a vertex id.
void	sendMessageToAllEdges(Vertex<I,V,E> vertex, M2 message) Send a message to all edges.
void	sendMessageToMultipleEdges(Iterator<I> vertexIdIterator, M2 message) Send a message to multiple target vertex ids in the iterator.

Methods inherited from class org.apache.giraph.worker.WorkerAggregatorDelegator

aggregate, getAggregatedValue, getBroadcast, reduce, reduceMerge, setWorkerGlobalCommUsage

Methods inherited from class org.apache.giraph.conf.DefaultImmutableClassesGiraphConfigurable

getConf, setConf

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface org.apache.giraph.conf.ImmutableClassesGiraphConfigurable

getConf

Methods inherited from interface org.apache.giraph.conf.GiraphConfigurationSettable

setConf

Methods inherited from interface org.apache.giraph.worker.WorkerBroadcastUsage

getBroadcast

Methods inherited from interface org.apache.giraph.worker.WorkerReduceUsage

reduce, reduceMerge

Methods inherited from interface org.apache.giraph.worker.WorkerAggregatorUsage

aggregate

Methods inherited from interface org.apache.giraph.aggregators.AggregatorUsage

getAggregatedValue

Constructor Detail

AbstractComputation

public AbstractComputation()

Method Detail

compute

public abstract void compute(Vertex<I,V,E> vertex,
                             Iterable<M1> messages)
                      throws IOException

Must be defined by user to do computation on a single Vertex.

Specified by:

compute in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Parameters:

vertex - Vertex

messages - Messages that were sent to this vertex in the previous superstep. Each message is only guaranteed to have a life expectancy as long as next() is not called.

Throws:

IOException

preSuperstep

public void preSuperstep()

Prepare for computation. This method is executed exactly once prior to compute(Vertex, Iterable) being called for any of the vertices in the partition.

Specified by:

preSuperstep in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

postSuperstep

public void postSuperstep()

Finish computation. This method is executed exactly once after computation for all vertices in the partition is complete.

Specified by:

postSuperstep in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

initialize

public void initialize(GraphState graphState,
                       WorkerClientRequestProcessor<I,V,E> workerClientRequestProcessor,
                       CentralizedServiceWorker<I,V,E> serviceWorker,
                       WorkerGlobalCommUsage workerGlobalCommUsage)

Initialize, called by infrastructure before the superstep starts. Shouldn't be called by user code.

Specified by:

initialize in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Parameters:

graphState - Graph state

workerClientRequestProcessor - Processor for handling requests

serviceWorker - Graph-wide BSP Mapper for this Vertex

workerGlobalCommUsage - Worker global communication usage

getSuperstep

public long getSuperstep()

Retrieves the current superstep.

Specified by:

getSuperstep in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Returns:

Current superstep

getTotalNumVertices

public long getTotalNumVertices()

Get the total (all workers) number of vertices that existed in the previous superstep.

Specified by:

getTotalNumVertices in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Returns:

Total number of vertices (-1 if first superstep)

getTotalNumEdges

public long getTotalNumEdges()

Get the total (all workers) number of edges that existed in the previous superstep.

Specified by:

getTotalNumEdges in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Returns:

Total number of edges (-1 if first superstep)

sendMessage

public void sendMessage(I id,
                        M2 message)

Send a message to a vertex id.

Specified by:

sendMessage in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Parameters:

id - Vertex id to send the message to

message - Message data to send

sendMessageToAllEdges

public void sendMessageToAllEdges(Vertex<I,V,E> vertex,
                                  M2 message)

Send a message to all edges.

Specified by:

sendMessageToAllEdges in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Parameters:

vertex - Vertex whose edges to send the message to.

message - Message sent to all edges.

sendMessageToMultipleEdges

public void sendMessageToMultipleEdges(Iterator<I> vertexIdIterator,
                                       M2 message)

Send a message to multiple target vertex ids in the iterator.

Specified by:

sendMessageToMultipleEdges in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Parameters:

vertexIdIterator - An iterator to multiple target vertex ids.

message - Message sent to all targets in the iterator.

addVertexRequest

public void addVertexRequest(I id,
                             V value,
                             OutEdges<I,E> edges)
                      throws IOException

Sends a request to create a vertex that will be available during the next superstep.

Specified by:

addVertexRequest in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Parameters:

id - Vertex id

value - Vertex value

edges - Initial edges

Throws:

IOException

addVertexRequest

public void addVertexRequest(I id,
                             V value)
                      throws IOException

Sends a request to create a vertex that will be available during the next superstep.

Specified by:

addVertexRequest in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Parameters:

id - Vertex id

value - Vertex value

Throws:

IOException

removeVertexRequest

public void removeVertexRequest(I vertexId)
                         throws IOException

Request to remove a vertex from the graph (applied just prior to the next superstep).

Specified by:

removeVertexRequest in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Parameters:

vertexId - Id of the vertex to be removed.

Throws:

IOException

addEdgeRequest

public void addEdgeRequest(I sourceVertexId,
                           Edge<I,E> edge)
                    throws IOException

Request to add an edge of a vertex in the graph (processed just prior to the next superstep)

Specified by:

addEdgeRequest in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Parameters:

sourceVertexId - Source vertex id of edge

edge - Edge to add

Throws:

IOException

removeEdgesRequest

public void removeEdgesRequest(I sourceVertexId,
                               I targetVertexId)
                        throws IOException

Request to remove all edges from a given source vertex to a given target vertex (processed just prior to the next superstep).

Specified by:

removeEdgesRequest in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Parameters:

sourceVertexId - Source vertex id

targetVertexId - Target vertex id

Throws:

IOException

getContext

public org.apache.hadoop.mapreduce.Mapper.Context getContext()

Get the mapper context

Specified by:

getContext in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Returns:

Mapper context

getWorkerContext

public <W extends WorkerContext> W getWorkerContext()

Get the worker context

Specified by:

getWorkerContext in interface Computation<I extends org.apache.hadoop.io.WritableComparable,V extends org.apache.hadoop.io.Writable,E extends org.apache.hadoop.io.Writable,M1 extends org.apache.hadoop.io.Writable,M2 extends org.apache.hadoop.io.Writable>

Type Parameters:

W - WorkerContext class

Returns:

WorkerContext context

getWorkerCount

public final int getWorkerCount()

Description copied from interface: WorkerIndexUsage

Get number of workers

Specified by:

getWorkerCount in interface WorkerIndexUsage<I extends org.apache.hadoop.io.WritableComparable>

Returns:

Number of workers

getMyWorkerIndex

public final int getMyWorkerIndex()

Description copied from interface: WorkerIndexUsage

Get index for this worker

Specified by:

getMyWorkerIndex in interface WorkerIndexUsage<I extends org.apache.hadoop.io.WritableComparable>

Returns:

Index of this worker

getWorkerForVertex

public final int getWorkerForVertex(I vertexId)

Description copied from interface: WorkerIndexUsage

Get worker index which will contain vertex with given id, if such vertex exists.

Specified by:

getWorkerForVertex in interface WorkerIndexUsage<I extends org.apache.hadoop.io.WritableComparable>

Parameters:

vertexId - vertex id

Returns:

worker index