org.apache.nutch.clustering
Interface HitsCluster

All Known Implementing Classes:

HitsClusterAdapter

public interface HitsCluster

An interface representing a group (cluster) of related hits.

If isJunkCluster() method returns true then this cluster contains documents that are grouped together, but no clear semantic relation has been detected. Such clusters may be hidden in the user interface layer, unless someone wishes to see an explicit group of documents that didn't belong anywhere else.

Version:

$Id: HitsCluster.java 823614 2009-10-09 17:02:32Z ab $

Author:

Dawid Weiss

Method Summary

String[]	getDescriptionLabels()
HitDetails[]	getHits()
HitsCluster[]	getSubclusters()
boolean	isJunkCluster() Returns true if this cluster constains documents that did not fit anywhere else (presentation layer may discard such clusters).

Method Detail

getSubclusters

HitsCluster[] getSubclusters()

Returns:

Returns an array of HitsCluster objects that are sub-groups of the current group, or null if this cluster has no sub-groups.

getHits

HitDetails[] getHits()

Returns:

Returns a relevance-ordered array of the hits belonging to this cluster or null if this cluster has no associated documents (it may have subclusters only).

getDescriptionLabels

String[] getDescriptionLabels()

Returns:

Returns an array of labels for this cluster. The labels should be sorted according to their relevance to the cluster's content. Not all of the labels must be displayed - the application is free to set a cutoff threshold and display only the topmost labels.

isJunkCluster

boolean isJunkCluster()

Returns true if this cluster constains documents that did not fit anywhere else (presentation layer may discard such clusters).

Subclusters of this cluster are also junk clusters, even if they don't have this property set to true