@Override
public DocScoreList mostSimilar(int wpId, int maxResults, TIntSet validIds) throws IOException {
if (hasCachedMostSimilar(wpId)) {
return getCachedMostSimilar(wpId, maxResults, validIds);
}
MoreLikeThis mlt = getMoreLikeThis();
int luceneId = esaHelper.wpIdToLuceneId(wpId);
Query query;
if (luceneId >= 0) {
query = mlt.like(luceneId);
} else if (textHelper != null && textHelper.wpIdToLuceneId(wpId) >= 0) {
Document d = textHelper.wpIdToLuceneDoc(wpId);
String text = d.get(Page.FIELD_TEXT);
query = mlt.like(new StringReader(text), Page.FIELD_TEXT);
} else {
return null;
}
TopDocs similarDocs = searcher.search(query, esaHelper.getWpIdFilter(validIds), maxResults);
pruneSimilar(similarDocs);
DocScoreList scores = new DocScoreList(similarDocs.scoreDocs.length);
for (int i = 0; i < similarDocs.scoreDocs.length; i++) {
ScoreDoc sd = similarDocs.scoreDocs[i];
scores.set(i, esaHelper.luceneIdToWpId(sd.doc), similarDocs.scoreDocs[i].score);
}
return normalize(scores);
}
@Override
public DocScoreList mostSimilar(String phrase, int maxResults, TIntSet validIds)
throws IOException {
final TIntDoubleHashMap scores = getConceptVector(phrase, validIds);
Integer luceneIds[] = ArrayUtils.toObject(scores.keys());
Arrays.sort(
luceneIds,
new Comparator<Integer>() {
@Override
public int compare(Integer id1, Integer id2) {
return -1 * new Double(scores.get(id1)).compareTo(scores.get(id2));
}
});
DocScoreList result = new DocScoreList(Math.min(luceneIds.length, maxResults));
for (int i = 0; i < result.numDocs(); i++) {
result.set(i, esaHelper.luceneIdToWpId(luceneIds[i]), scores.get(luceneIds[i]));
}
return normalize(result);
}