use of org.apache.solr.search.EarlyTerminatingCollectorException in project lucene-solr by apache.
the class SpellCheckCollator method collate.
public List<SpellCheckCollation> collate(SpellingResult result, String originalQuery, ResponseBuilder ultimateResponse) {
List<SpellCheckCollation> collations = new ArrayList<>();
QueryComponent queryComponent = null;
if (ultimateResponse.components != null) {
for (SearchComponent sc : ultimateResponse.components) {
if (sc instanceof QueryComponent) {
queryComponent = (QueryComponent) sc;
break;
}
}
}
boolean verifyCandidateWithQuery = true;
int maxTries = maxCollationTries;
int maxNumberToIterate = maxTries;
if (maxTries < 1) {
maxTries = 1;
maxNumberToIterate = maxCollations;
verifyCandidateWithQuery = false;
}
if (queryComponent == null && verifyCandidateWithQuery) {
LOG.info("Could not find an instance of QueryComponent. Disabling collation verification against the index.");
maxTries = 1;
verifyCandidateWithQuery = false;
}
docCollectionLimit = docCollectionLimit > 0 ? docCollectionLimit : 0;
int maxDocId = -1;
if (verifyCandidateWithQuery && docCollectionLimit > 0) {
IndexReader reader = ultimateResponse.req.getSearcher().getIndexReader();
maxDocId = reader.maxDoc();
}
int tryNo = 0;
int collNo = 0;
PossibilityIterator possibilityIter = new PossibilityIterator(result.getSuggestions(), maxNumberToIterate, maxCollationEvaluations, suggestionsMayOverlap);
while (tryNo < maxTries && collNo < maxCollations && possibilityIter.hasNext()) {
PossibilityIterator.RankedSpellPossibility possibility = possibilityIter.next();
String collationQueryStr = getCollation(originalQuery, possibility.corrections);
int hits = 0;
if (verifyCandidateWithQuery) {
tryNo++;
SolrParams origParams = ultimateResponse.req.getParams();
ModifiableSolrParams params = new ModifiableSolrParams(origParams);
Iterator<String> origParamIterator = origParams.getParameterNamesIterator();
int pl = SpellingParams.SPELLCHECK_COLLATE_PARAM_OVERRIDE.length();
while (origParamIterator.hasNext()) {
String origParamName = origParamIterator.next();
if (origParamName.startsWith(SpellingParams.SPELLCHECK_COLLATE_PARAM_OVERRIDE) && origParamName.length() > pl) {
String[] val = origParams.getParams(origParamName);
if (val.length == 1 && val[0].length() == 0) {
params.set(origParamName.substring(pl), (String[]) null);
} else {
params.set(origParamName.substring(pl), val);
}
}
}
params.set(CommonParams.Q, collationQueryStr);
params.remove(CommonParams.START);
params.set(CommonParams.ROWS, "" + docCollectionLimit);
// we don't want any stored fields
params.set(CommonParams.FL, ID);
// we'll sort by doc id to ensure no scoring is done.
params.set(CommonParams.SORT, "_docid_ asc");
// CursorMark does not like _docid_ sorting, and we don't need it.
params.remove(CursorMarkParams.CURSOR_MARK_PARAM);
// If a dismax query, don't add unnecessary clauses for scoring
params.remove(DisMaxParams.TIE);
params.remove(DisMaxParams.PF);
params.remove(DisMaxParams.PF2);
params.remove(DisMaxParams.PF3);
params.remove(DisMaxParams.BQ);
params.remove(DisMaxParams.BF);
// Collate testing does not support Grouping (see SOLR-2577)
params.remove(GroupParams.GROUP);
// Collate testing does not support the Collapse QParser (See SOLR-8807)
params.remove("expand");
String[] filters = params.getParams(CommonParams.FQ);
if (filters != null) {
List<String> filtersToApply = new ArrayList<>(filters.length);
for (String fq : filters) {
if (!fq.startsWith("{!collapse")) {
filtersToApply.add(fq);
}
}
params.set("fq", filtersToApply.toArray(new String[filtersToApply.size()]));
}
// creating a request here... make sure to close it!
ResponseBuilder checkResponse = new ResponseBuilder(new LocalSolrQueryRequest(ultimateResponse.req.getCore(), params), new SolrQueryResponse(), Arrays.<SearchComponent>asList(queryComponent));
checkResponse.setQparser(ultimateResponse.getQparser());
checkResponse.setFilters(ultimateResponse.getFilters());
checkResponse.setQueryString(collationQueryStr);
checkResponse.components = Arrays.<SearchComponent>asList(queryComponent);
try {
queryComponent.prepare(checkResponse);
if (docCollectionLimit > 0) {
int f = checkResponse.getFieldFlags();
checkResponse.setFieldFlags(f |= SolrIndexSearcher.TERMINATE_EARLY);
}
queryComponent.process(checkResponse);
hits = (Integer) checkResponse.rsp.getToLog().get("hits");
} catch (EarlyTerminatingCollectorException etce) {
assert (docCollectionLimit > 0);
assert 0 < etce.getNumberScanned();
assert 0 < etce.getNumberCollected();
if (etce.getNumberScanned() == maxDocId) {
hits = etce.getNumberCollected();
} else {
hits = (int) (((float) (maxDocId * etce.getNumberCollected())) / (float) etce.getNumberScanned());
}
} catch (Exception e) {
LOG.warn("Exception trying to re-query to check if a spell check possibility would return any hits.", e);
} finally {
checkResponse.req.close();
}
}
if (hits > 0 || !verifyCandidateWithQuery) {
collNo++;
SpellCheckCollation collation = new SpellCheckCollation();
collation.setCollationQuery(collationQueryStr);
collation.setHits(hits);
collation.setInternalRank(suggestionsMayOverlap ? ((possibility.rank * 1000) + possibility.index) : possibility.rank);
NamedList<String> misspellingsAndCorrections = new NamedList<>();
for (SpellCheckCorrection corr : possibility.corrections) {
misspellingsAndCorrections.add(corr.getOriginal().toString(), corr.getCorrection());
}
collation.setMisspellingsAndCorrections(misspellingsAndCorrections);
collations.add(collation);
}
if (LOG.isDebugEnabled()) {
LOG.debug("Collation: " + collationQueryStr + (verifyCandidateWithQuery ? (" will return " + hits + " hits.") : ""));
}
}
return collations;
}
Aggregations