private void pruneSimilar(TopDocs docs) throws IOException {
if (docs.scoreDocs.length == 0) {
return;
}
int cutoff = docs.scoreDocs.length;
double threshold = 0.005 * docs.scoreDocs[0].score;
for (int i = 0, j = 100; j < docs.scoreDocs.length; i++, j++) {
float delta = docs.scoreDocs[i].score - docs.scoreDocs[j].score;
if (delta < threshold) {
cutoff = j;
break;
}
}
if (cutoff < docs.scoreDocs.length) {
// LOG.info("pruned results from " + docs.scoreDocs.length + " to " + cutoff);
docs.scoreDocs = ArrayUtils.subarray(docs.scoreDocs, 0, cutoff);
}
}
public TopDocs topDocs(int start, int howMany) {
try {
TopDocs mainDocs = mainCollector.topDocs(0, Math.max(reRankDocs, length));
if (mainDocs.totalHits == 0 || mainDocs.scoreDocs.length == 0) {
return mainDocs;
}
if (boostedPriority != null) {
SolrRequestInfo info = SolrRequestInfo.getRequestInfo();
Map requestContext = null;
if (info != null) {
requestContext = info.getReq().getContext();
}
IntIntOpenHashMap boostedDocs =
QueryElevationComponent.getBoostDocs(
(SolrIndexSearcher) searcher, boostedPriority, requestContext);
ScoreDoc[] mainScoreDocs = mainDocs.scoreDocs;
ScoreDoc[] reRankScoreDocs = new ScoreDoc[Math.min(mainScoreDocs.length, reRankDocs)];
System.arraycopy(mainScoreDocs, 0, reRankScoreDocs, 0, reRankScoreDocs.length);
mainDocs.scoreDocs = reRankScoreDocs;
Map<Integer, Float> scoreMap = getScoreMap(mainDocs.scoreDocs, mainDocs.scoreDocs.length);
TopDocs rescoredDocs =
new QueryRescorer(reRankQuery) {
@Override
protected float combine(
float firstPassScore, boolean secondPassMatches, float secondPassScore) {
float score = firstPassScore;
if (secondPassMatches) {
score += reRankWeight * secondPassScore;
}
return score;
}
}.rescore(searcher, mainDocs, mainDocs.scoreDocs.length);
Arrays.sort(
rescoredDocs.scoreDocs,
new BoostedComp(boostedDocs, mainDocs.scoreDocs, rescoredDocs.getMaxScore()));
// Lower howMany if we've collected fewer documents.
howMany = Math.min(howMany, mainScoreDocs.length);
if (howMany == rescoredDocs.scoreDocs.length) {
if (scale) {
scaleScores(rescoredDocs, scoreMap);
}
return rescoredDocs; // Just return the rescoredDocs
} else if (howMany > rescoredDocs.scoreDocs.length) {
// We need to return more then we've reRanked, so create the combined page.
ScoreDoc[] scoreDocs = new ScoreDoc[howMany];
System.arraycopy(
mainScoreDocs, 0, scoreDocs, 0, scoreDocs.length); // lay down the initial docs
System.arraycopy(
rescoredDocs.scoreDocs,
0,
scoreDocs,
0,
rescoredDocs.scoreDocs.length); // overlay the re-ranked docs.
rescoredDocs.scoreDocs = scoreDocs;
if (scale) {
scaleScores(rescoredDocs, scoreMap);
}
return rescoredDocs;
} else {
// We've rescored more then we need to return.
ScoreDoc[] scoreDocs = new ScoreDoc[howMany];
System.arraycopy(rescoredDocs.scoreDocs, 0, scoreDocs, 0, howMany);
rescoredDocs.scoreDocs = scoreDocs;
if (scale) {
scaleScores(rescoredDocs, scoreMap);
}
return rescoredDocs;
}
} else {
ScoreDoc[] mainScoreDocs = mainDocs.scoreDocs;
/*
* Create the array for the reRankScoreDocs.
*/
ScoreDoc[] reRankScoreDocs = new ScoreDoc[Math.min(mainScoreDocs.length, reRankDocs)];
/*
* Copy the initial results into the reRankScoreDocs array.
*/
System.arraycopy(mainScoreDocs, 0, reRankScoreDocs, 0, reRankScoreDocs.length);
mainDocs.scoreDocs = reRankScoreDocs;
Map<Integer, Float> scoreMap = getScoreMap(mainDocs.scoreDocs, mainDocs.scoreDocs.length);
TopDocs rescoredDocs =
new QueryRescorer(reRankQuery) {
@Override
protected float combine(
float firstPassScore, boolean secondPassMatches, float secondPassScore) {
float score = firstPassScore;
if (secondPassMatches) {
score += reRankWeight * secondPassScore;
}
return score;
}
}.rescore(searcher, mainDocs, mainDocs.scoreDocs.length);
// Lower howMany to return if we've collected fewer documents.
howMany = Math.min(howMany, mainScoreDocs.length);
if (howMany == rescoredDocs.scoreDocs.length) {
if (scale) {
scaleScores(rescoredDocs, scoreMap);
}
return rescoredDocs; // Just return the rescoredDocs
} else if (howMany > rescoredDocs.scoreDocs.length) {
// We need to return more then we've reRanked, so create the combined page.
ScoreDoc[] scoreDocs = new ScoreDoc[howMany];
// lay down the initial docs
System.arraycopy(mainScoreDocs, 0, scoreDocs, 0, scoreDocs.length);
// overlay the rescoreds docs
System.arraycopy(
rescoredDocs.scoreDocs, 0, scoreDocs, 0, rescoredDocs.scoreDocs.length);
rescoredDocs.scoreDocs = scoreDocs;
if (scale) {
assert (scoreMap != null);
scaleScores(rescoredDocs, scoreMap);
}
return rescoredDocs;
} else {
// We've rescored more then we need to return.
ScoreDoc[] scoreDocs = new ScoreDoc[howMany];
System.arraycopy(rescoredDocs.scoreDocs, 0, scoreDocs, 0, howMany);
rescoredDocs.scoreDocs = scoreDocs;
if (scale) {
scaleScores(rescoredDocs, scoreMap);
}
return rescoredDocs;
}
}
} catch (Exception e) {
throw new SolrException(SolrException.ErrorCode.BAD_REQUEST, e);
}
}