org.neo4j.graphalgo.impl.path
Class ShortestPath

java.lang.Object
  org.neo4j.graphalgo.impl.path.ShortestPath

All Implemented Interfaces:

PathFinder<Path>

public class ShortestPath
extends Object
implements PathFinder<Path>

Find (all or one) simple shortest path(s) between two nodes. It starts from both ends and goes one relationship at the time, alternating side between each traversal. It does so to minimize the traversal overhead if one side has a very large amount of relationships, but the other one very few. It performs well however the graph is proportioned. Relationships are traversed in the specified directions from the start node, but in the reverse direction ( Direction.reverse() ) from the end node. This doesn't affect Direction.BOTH.

Nested Class Summary

protected class

ShortestPath.DirectionData

Constructor Summary

ShortestPath(int maxDepth, RelationshipExpander relExpander)
Constructs a new shortest path algorithm.

ShortestPath(int maxDepth, RelationshipExpander relExpander, int maxResultCount)
Constructs a new shortest path algorithm.

ShortestPath(int maxDepth, RelationshipExpander relExpander, int maxResultCount, boolean findPathsOnMaxDepthOnly)
Constructs a new shortest path algorithm.

Method Summary

protected Collection<Node>	filterNextLevelNodes(Collection<Node> nextNodes)
Iterable<Path>	findAllPaths(Node start, Node end) Tries to find all paths between start and end nodes.
Path	findSinglePath(Node start, Node end) Tries to find a single path between start and end nodes.

Methods inherited from class java.lang.Object

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

Constructor Detail

ShortestPath

public ShortestPath(int maxDepth,
                    RelationshipExpander relExpander)

Constructs a new shortest path algorithm.

Parameters:

maxDepth - the maximum depth for the traversal. Returned paths will never have a greater Path.length() than maxDepth.

relExpander - the RelationshipExpander to use for deciding which relationships to expand for each Node.

ShortestPath

public ShortestPath(int maxDepth,
                    RelationshipExpander relExpander,
                    int maxResultCount)

Constructs a new shortest path algorithm.

Parameters:

maxDepth - the maximum depth for the traversal. Returned paths will never have a greater Path.length() than maxDepth.

relExpander - the RelationshipExpander to use for deciding which relationships to expand for each Node.

maxResultCount - the maximum number of hits to return. If this number of hits are encountered the traversal will stop.

ShortestPath

public ShortestPath(int maxDepth,
                    RelationshipExpander relExpander,
                    int maxResultCount,
                    boolean findPathsOnMaxDepthOnly)

Constructs a new shortest path algorithm.

Parameters:

maxDepth - the maximum depth for the traversal. Returned paths will never have a greater Path.length() than maxDepth.

relExpander - the RelationshipExpander to use for deciding which relationships to expand for each Node.

maxResultCount - the maximum number of hits to return. If this number of hits are encountered the traversal will stop.

findPathsOnMaxDepthOnly - if true then it will only try to find paths on that particular depth (maxDepth).

Method Detail

findAllPaths

public Iterable<Path> findAllPaths(Node start,
                                   Node end)

Description copied from interface: PathFinder

Tries to find all paths between start and end nodes. A collection of Paths is returned with all the found paths. If no paths are found an empty collection is returned.

Specified by:

findAllPaths in interface PathFinder<Path>

Parameters:

start - the start Node which defines the start of the path.

end - the end Node which defines the end of the path.

Returns:

all Paths between start and end.

findSinglePath

public Path findSinglePath(Node start,
                           Node end)

Description copied from interface: PathFinder

Tries to find a single path between start and end nodes. If a path is found a Path is returned with that path information, else null is returned. If more than one path is found, the implementation can decide itself upon which of those to return.

Specified by:

findSinglePath in interface PathFinder<Path>

Parameters:

start - the start Node which defines the start of the path.

end - the end Node which defines the end of the path.

Returns:

a single Path between start and end, or null if no path was found.

filterNextLevelNodes

protected Collection<Node> filterNextLevelNodes(Collection<Node> nextNodes)