com.sparsity.sparksee.algorithms

Class CommunitiesSCD

java.lang.Object

com.sparsity.sparksee.algorithms.CommunityDetection

com.sparsity.sparksee.algorithms.DisjointCommunityDetection

com.sparsity.sparksee.algorithms.CommunitiesSCD

All Implemented Interfaces:

java.io.Closeable, java.lang.AutoCloseable

public class CommunitiesSCD
extends DisjointCommunityDetection

CommunitiesSCD class.

Implementation of the community detection algorithm "Scalable Community Detection" based on the paper "High quality, scalable and parallel community detection for large real graphs" by Arnau Prat-Perez, David Dominguez-Sal, Josep-Lluis Larriba-Pey - WWW 2014.

The purpose of this algorithm is to find disjoint communities in an undirected graph or in a directed graph which will be considered as an undirected one.

It is possible to set some restrictions after constructing a new instance of this class and before running it in order to limit the results.

After the execution, we can retrieve the results stored in an instance of the DisjointCommunities class using the getCommunities method.

Check out the 'Algorithms' section in the SPARKSEE User Manual for more details on this.

Author:

Sparsity Technologies http://www.sparsity-technologies.com

Constructor Summary

Constructors

Constructor and Description
CommunitiesSCD(Session session) Creates a new instance of CommunitiesSCD.

Method Summary

Modifier and Type	Method and Description
void	addAllEdgeTypes() Allows connectivity through all edge types of the graph.
void	addAllNodeTypes() Allows connectivity through all node types of the graph.
void	addEdgeType(int type) Allows connectivity through edges of the given type.
void	addNodeType(int type) Allows connectivity through nodes of the given type.
void	excludeEdges(Objects edges) Set which edges can't be used.
void	excludeNodes(Objects nodes) Set which nodes can't be used.
DisjointCommunities	getCommunities() Returns the results generated by the execution of the algorithm.
void	includeEdges(Objects edges) Set additional edges that can be used.
void	includeNodes(Objects nodes) Set additional nodes that can be used.
void	run() Executes the algorithm.
void	setLookAhead(int lookahead) Sets the size of the lookahead iterations to look (5 by default).
void	setMaterializedAttribute(java.lang.String attributeName) Creates a new common attribute type for all node types in the graph in order to store, persistently, the results related to the disjoint communities found while executing this algorithm.

Methods inherited from class com.sparsity.sparksee.algorithms.CommunityDetection

close, isClosed

Methods inherited from class java.lang.Object

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

Constructor Detail

CommunitiesSCD

public CommunitiesSCD(Session session)

Creates a new instance of CommunitiesSCD.

After creating this instance is required to indicate the set of edge types and the set of node types which will be navigated through while traversing the graph in order to find the communities.

Parameters:

session - [in] Session to get the graph from and calculate the communities

Method Detail

excludeNodes

public void excludeNodes(Objects nodes)

Set which nodes can't be used.

This will replace any previously specified set of excluded nodes. Should only be used to exclude the usage of specific nodes from allowed node types because it's less efficient than not allowing a node type.

Overrides:

excludeNodes in class DisjointCommunityDetection

Parameters:

nodes - [in] A set of node identifiers that must be kept intact until the destruction of the class.

addAllEdgeTypes

public void addAllEdgeTypes()

Allows connectivity through all edge types of the graph.

The edges can be used in Any direction.

Overrides:

addAllEdgeTypes in class DisjointCommunityDetection

getCommunities

public DisjointCommunities getCommunities()

Returns the results generated by the execution of the algorithm.

These results contain information related to the disjoint communities found as the number of different components, the set of nodes contained in each component or many other data.

Overrides:

getCommunities in class DisjointCommunityDetection

Returns:

Returns an instance of the class DisjointCommunities which contain information related to the disjoint communities found.

run

public void run()

Executes the algorithm.

Overrides:

run in class DisjointCommunityDetection

addEdgeType

public void addEdgeType(int type)

Allows connectivity through edges of the given type.

The edges can be used in Any direction.

Overrides:

addEdgeType in class DisjointCommunityDetection

Parameters:

type - [in] Edge type.

excludeEdges

public void excludeEdges(Objects edges)

Set which edges can't be used.

This will replace any previously specified set of excluded edges. Should only be used to exclude the usage of specific edges from allowed edge types because it's less efficient than not allowing an edge type.

Overrides:

excludeEdges in class DisjointCommunityDetection

Parameters:

edges - [in] A set of edge identifiers that must be kept intact until the destruction of the class.

setMaterializedAttribute

public void setMaterializedAttribute(java.lang.String attributeName)

Creates a new common attribute type for all node types in the graph in order to store, persistently, the results related to the disjoint communities found while executing this algorithm.

Whenever the user wants to retrieve the results, even when the graph has been closed and opened again, it is only necessary to create a new instance of the class DisjointCommunities indicating the graph and the name of the common attribute type which stores the results. This instance will have all the information related to the disjoint communities found in the moment of the execution of the algorithm that stored this data.

It is possible to run the algorithm without specifying this parameter in order to avoid materializing the results of the execution.

Overrides:

setMaterializedAttribute in class DisjointCommunityDetection

Parameters:

attributeName - [in] The name of the common attribute type for all node types in the graph which will store persistently the results generated by the execution of the algorithm.

addNodeType

public void addNodeType(int type)

Allows connectivity through nodes of the given type.

Overrides:

addNodeType in class DisjointCommunityDetection

Parameters:

type - null

includeNodes

public void includeNodes(Objects nodes)

Set additional nodes that can be used.

This will replace any previously specified set of include nodes. Using this optional method adds valid nodes to the nodes of any node type explicitly set as a valid type. Should only be used to include specific small sets of nodes because it's less efficient than just using a node type.

Overrides:

includeNodes in class DisjointCommunityDetection

Parameters:

nodes - [in] A set of node identifiers that must be kept intact until the destruction of the class.

setLookAhead

public void setLookAhead(int lookahead)

Sets the size of the lookahead iterations to look (5 by default).

Parameters:

lookahead - [in] Number of iterations. It must be positive or zero.

addAllNodeTypes

public void addAllNodeTypes()

Allows connectivity through all node types of the graph.

Overrides:

addAllNodeTypes in class DisjointCommunityDetection

includeEdges

public void includeEdges(Objects edges)

Set additional edges that can be used.

This will replace any previously specified set of include edges. Using this optional method adds valid edges to the edges of any edge type explicitly set as a valid type. Should only be used to include specific small sets of edges because it's less efficient than just using an edge type. For any edge to be used, both nodes must be also valid.

Overrides:

includeEdges in class DisjointCommunityDetection

Parameters:

edges - [in] A set of edge identifiers that must be kept intact until the destruction of the class.