/**
* Builds an "Other Topics" cluster that groups those documents from <code>allDocument</code> that
* were not referenced in any cluster in <code>clusters</code>.
*
* @param allDocuments all documents to check against
* @param clusters list of clusters with assigned documents
* @param label label for the "Other Topics" group
* @return the "Other Topics" cluster
*/
public static Cluster buildOtherTopics(
List<Document> allDocuments, List<Cluster> clusters, String label) {
final Set<Document> unclusteredDocuments = Sets.newLinkedHashSet(allDocuments);
final Set<Document> assignedDocuments = Sets.newHashSet();
for (Cluster cluster : clusters) {
collectAllDocuments(cluster, assignedDocuments);
}
unclusteredDocuments.removeAll(assignedDocuments);
final Cluster otherTopics = new Cluster(label);
otherTopics.addDocuments(unclusteredDocuments);
otherTopics.setOtherTopics(true);
return otherTopics;
}
/** A recursive routine for collecting unique documents from this cluster and subclusters. */
private static Set<Document> collectAllDocuments(Cluster cluster, Set<Document> docs) {
if (cluster == null) {
return docs;
}
docs.addAll(cluster.getDocuments());
final List<Cluster> subclusters = cluster.getSubclusters();
for (final Cluster subcluster : subclusters) {
collectAllDocuments(subcluster, docs);
}
return docs;
}