use of org.apache.lucene.search.ConstantScoreQuery in project lucene-solr by apache.
the class TestTaxonomyFacetSumValueSource method testSumScoreAggregator.
public void testSumScoreAggregator() throws Exception {
Directory indexDir = newDirectory();
Directory taxoDir = newDirectory();
DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
IndexWriter iw = new IndexWriter(indexDir, newIndexWriterConfig(new MockAnalyzer(random())));
FacetsConfig config = new FacetsConfig();
for (int i = atLeast(30); i > 0; --i) {
Document doc = new Document();
if (random().nextBoolean()) {
// don't match all documents
doc.add(new StringField("f", "v", Field.Store.NO));
}
doc.add(new FacetField("dim", "a"));
iw.addDocument(config.build(taxoWriter, doc));
}
DirectoryReader r = DirectoryReader.open(iw);
DirectoryTaxonomyReader taxoReader = new DirectoryTaxonomyReader(taxoWriter);
FacetsCollector fc = new FacetsCollector(true);
BoostQuery csq = new BoostQuery(new ConstantScoreQuery(new MatchAllDocsQuery()), 2f);
TopDocs td = FacetsCollector.search(newSearcher(r), csq, 10, fc);
Facets facets = new TaxonomyFacetSumValueSource(taxoReader, config, fc, DoubleValuesSource.SCORES);
int expected = (int) (td.getMaxScore() * td.totalHits);
assertEquals(expected, facets.getSpecificValue("dim", "a").intValue());
iw.close();
IOUtils.close(taxoWriter, taxoReader, taxoDir, r, indexDir);
}
use of org.apache.lucene.search.ConstantScoreQuery in project lucene-solr by apache.
the class SynonymTokenizer method testGetBestFragmentsConstantScore.
public void testGetBestFragmentsConstantScore() throws Exception {
TestHighlightRunner helper = new TestHighlightRunner() {
@Override
public void run() throws Exception {
numHighlights = 0;
if (random().nextBoolean()) {
BooleanQuery.Builder bq = new BooleanQuery.Builder();
bq.add(new ConstantScoreQuery(new TermQuery(new Term(FIELD_NAME, "kennedy"))), Occur.MUST);
bq.add(new ConstantScoreQuery(new TermQuery(new Term(FIELD_NAME, "kennedy"))), Occur.MUST);
doSearching(bq.build());
} else {
doSearching(new ConstantScoreQuery(new TermQuery(new Term(FIELD_NAME, "kennedy"))));
}
doStandardHighlights(analyzer, searcher, hits, query, HighlighterTest.this);
assertTrue("Failed to find correct number of highlights " + numHighlights + " found", numHighlights == 4);
}
};
helper.start();
}
use of org.apache.lucene.search.ConstantScoreQuery in project lucene-solr by apache.
the class MultiTermHighlighting method extractAutomata.
/**
* Extracts MultiTermQueries that match the provided field predicate.
* Returns equivalent automata that will match terms.
*/
public static CharacterRunAutomaton[] extractAutomata(Query query, Predicate<String> fieldMatcher, boolean lookInSpan, Function<Query, Collection<Query>> preRewriteFunc) {
// TODO Lucene needs a Query visitor API! LUCENE-3041
List<CharacterRunAutomaton> list = new ArrayList<>();
Collection<Query> customSubQueries = preRewriteFunc.apply(query);
if (customSubQueries != null) {
for (Query sub : customSubQueries) {
list.addAll(Arrays.asList(extractAutomata(sub, fieldMatcher, lookInSpan, preRewriteFunc)));
}
} else if (query instanceof BooleanQuery) {
for (BooleanClause clause : (BooleanQuery) query) {
if (!clause.isProhibited()) {
list.addAll(Arrays.asList(extractAutomata(clause.getQuery(), fieldMatcher, lookInSpan, preRewriteFunc)));
}
}
} else if (query instanceof ConstantScoreQuery) {
list.addAll(Arrays.asList(extractAutomata(((ConstantScoreQuery) query).getQuery(), fieldMatcher, lookInSpan, preRewriteFunc)));
} else if (query instanceof BoostQuery) {
list.addAll(Arrays.asList(extractAutomata(((BoostQuery) query).getQuery(), fieldMatcher, lookInSpan, preRewriteFunc)));
} else if (query instanceof DisjunctionMaxQuery) {
for (Query sub : ((DisjunctionMaxQuery) query).getDisjuncts()) {
list.addAll(Arrays.asList(extractAutomata(sub, fieldMatcher, lookInSpan, preRewriteFunc)));
}
} else if (lookInSpan && query instanceof SpanOrQuery) {
for (Query sub : ((SpanOrQuery) query).getClauses()) {
list.addAll(Arrays.asList(extractAutomata(sub, fieldMatcher, lookInSpan, preRewriteFunc)));
}
} else if (lookInSpan && query instanceof SpanNearQuery) {
for (Query sub : ((SpanNearQuery) query).getClauses()) {
list.addAll(Arrays.asList(extractAutomata(sub, fieldMatcher, lookInSpan, preRewriteFunc)));
}
} else if (lookInSpan && query instanceof SpanNotQuery) {
list.addAll(Arrays.asList(extractAutomata(((SpanNotQuery) query).getInclude(), fieldMatcher, lookInSpan, preRewriteFunc)));
} else if (lookInSpan && query instanceof SpanPositionCheckQuery) {
list.addAll(Arrays.asList(extractAutomata(((SpanPositionCheckQuery) query).getMatch(), fieldMatcher, lookInSpan, preRewriteFunc)));
} else if (lookInSpan && query instanceof SpanBoostQuery) {
list.addAll(Arrays.asList(extractAutomata(((SpanBoostQuery) query).getQuery(), fieldMatcher, lookInSpan, preRewriteFunc)));
} else if (lookInSpan && query instanceof SpanMultiTermQueryWrapper) {
list.addAll(Arrays.asList(extractAutomata(((SpanMultiTermQueryWrapper<?>) query).getWrappedQuery(), fieldMatcher, lookInSpan, preRewriteFunc)));
} else if (query instanceof PrefixQuery) {
final PrefixQuery pq = (PrefixQuery) query;
Term prefix = pq.getPrefix();
if (fieldMatcher.test(prefix.field())) {
list.add(new CharacterRunAutomaton(Operations.concatenate(Automata.makeString(prefix.text()), Automata.makeAnyString())) {
@Override
public String toString() {
return pq.toString();
}
});
}
} else if (query instanceof FuzzyQuery) {
final FuzzyQuery fq = (FuzzyQuery) query;
if (fieldMatcher.test(fq.getField())) {
String utf16 = fq.getTerm().text();
int[] termText = new int[utf16.codePointCount(0, utf16.length())];
for (int cp, i = 0, j = 0; i < utf16.length(); i += Character.charCount(cp)) {
termText[j++] = cp = utf16.codePointAt(i);
}
int termLength = termText.length;
int prefixLength = Math.min(fq.getPrefixLength(), termLength);
String suffix = UnicodeUtil.newString(termText, prefixLength, termText.length - prefixLength);
LevenshteinAutomata builder = new LevenshteinAutomata(suffix, fq.getTranspositions());
String prefix = UnicodeUtil.newString(termText, 0, prefixLength);
Automaton automaton = builder.toAutomaton(fq.getMaxEdits(), prefix);
list.add(new CharacterRunAutomaton(automaton) {
@Override
public String toString() {
return fq.toString();
}
});
}
} else if (query instanceof TermRangeQuery) {
final TermRangeQuery tq = (TermRangeQuery) query;
if (fieldMatcher.test(tq.getField())) {
final CharsRef lowerBound;
if (tq.getLowerTerm() == null) {
lowerBound = null;
} else {
lowerBound = new CharsRef(tq.getLowerTerm().utf8ToString());
}
final CharsRef upperBound;
if (tq.getUpperTerm() == null) {
upperBound = null;
} else {
upperBound = new CharsRef(tq.getUpperTerm().utf8ToString());
}
final boolean includeLower = tq.includesLower();
final boolean includeUpper = tq.includesUpper();
final CharsRef scratch = new CharsRef();
@SuppressWarnings("deprecation") final Comparator<CharsRef> comparator = CharsRef.getUTF16SortedAsUTF8Comparator();
// this is *not* an automaton, but its very simple
list.add(new CharacterRunAutomaton(Automata.makeEmpty()) {
@Override
public boolean run(char[] s, int offset, int length) {
scratch.chars = s;
scratch.offset = offset;
scratch.length = length;
if (lowerBound != null) {
int cmp = comparator.compare(scratch, lowerBound);
if (cmp < 0 || (!includeLower && cmp == 0)) {
return false;
}
}
if (upperBound != null) {
int cmp = comparator.compare(scratch, upperBound);
if (cmp > 0 || (!includeUpper && cmp == 0)) {
return false;
}
}
return true;
}
@Override
public String toString() {
return tq.toString();
}
});
}
} else if (query instanceof AutomatonQuery) {
final AutomatonQuery aq = (AutomatonQuery) query;
if (fieldMatcher.test(aq.getField())) {
list.add(new CharacterRunAutomaton(aq.getAutomaton()) {
@Override
public String toString() {
return aq.toString();
}
});
}
}
return list.toArray(new CharacterRunAutomaton[list.size()]);
}
use of org.apache.lucene.search.ConstantScoreQuery in project lucene-solr by apache.
the class ConstantScoreQueryBuilder method getQuery.
@Override
public Query getQuery(Element e) throws ParserException {
Element queryElem = DOMUtils.getFirstChildOrFail(e);
Query q = new ConstantScoreQuery(queryFactory.getQuery(queryElem));
float boost = DOMUtils.getAttribute(e, "boost", 1.0f);
if (boost != 1f) {
q = new BoostQuery(q, boost);
}
return q;
}
use of org.apache.lucene.search.ConstantScoreQuery in project lucene-solr by apache.
the class TestFieldCacheSortRandom method testRandomStringSort.
private void testRandomStringSort(SortField.Type type) throws Exception {
Random random = new Random(random().nextLong());
final int NUM_DOCS = atLeast(100);
final Directory dir = newDirectory();
final RandomIndexWriter writer = new RandomIndexWriter(random, dir);
final boolean allowDups = random.nextBoolean();
final Set<String> seen = new HashSet<>();
final int maxLength = TestUtil.nextInt(random, 5, 100);
if (VERBOSE) {
System.out.println("TEST: NUM_DOCS=" + NUM_DOCS + " maxLength=" + maxLength + " allowDups=" + allowDups);
}
int numDocs = 0;
final List<BytesRef> docValues = new ArrayList<>();
// TODO: deletions
while (numDocs < NUM_DOCS) {
final Document doc = new Document();
// 10% of the time, the document is missing the value:
final BytesRef br;
if (random().nextInt(10) != 7) {
final String s;
if (random.nextBoolean()) {
s = TestUtil.randomSimpleString(random, maxLength);
} else {
s = TestUtil.randomUnicodeString(random, maxLength);
}
if (!allowDups) {
if (seen.contains(s)) {
continue;
}
seen.add(s);
}
if (VERBOSE) {
System.out.println(" " + numDocs + ": s=" + s);
}
doc.add(new StringField("stringdv", s, Field.Store.NO));
docValues.add(new BytesRef(s));
} else {
br = null;
if (VERBOSE) {
System.out.println(" " + numDocs + ": <missing>");
}
docValues.add(null);
}
doc.add(new IntPoint("id", numDocs));
doc.add(new StoredField("id", numDocs));
writer.addDocument(doc);
numDocs++;
if (random.nextInt(40) == 17) {
// force flush
writer.getReader().close();
}
}
Map<String, UninvertingReader.Type> mapping = new HashMap<>();
mapping.put("stringdv", Type.SORTED);
mapping.put("id", Type.INTEGER_POINT);
final IndexReader r = UninvertingReader.wrap(writer.getReader(), mapping);
writer.close();
if (VERBOSE) {
System.out.println(" reader=" + r);
}
final IndexSearcher s = newSearcher(r, false);
final int ITERS = atLeast(100);
for (int iter = 0; iter < ITERS; iter++) {
final boolean reverse = random.nextBoolean();
final TopFieldDocs hits;
final SortField sf;
final boolean sortMissingLast;
final boolean missingIsNull;
sf = new SortField("stringdv", type, reverse);
sortMissingLast = random().nextBoolean();
missingIsNull = true;
if (sortMissingLast) {
sf.setMissingValue(SortField.STRING_LAST);
}
final Sort sort;
if (random.nextBoolean()) {
sort = new Sort(sf);
} else {
sort = new Sort(sf, SortField.FIELD_DOC);
}
final int hitCount = TestUtil.nextInt(random, 1, r.maxDoc() + 20);
final RandomQuery f = new RandomQuery(random.nextLong(), random.nextFloat(), docValues);
int queryType = random.nextInt(2);
if (queryType == 0) {
hits = s.search(new ConstantScoreQuery(f), hitCount, sort, random.nextBoolean(), random.nextBoolean());
} else {
hits = s.search(f, hitCount, sort, random.nextBoolean(), random.nextBoolean());
}
if (VERBOSE) {
System.out.println("\nTEST: iter=" + iter + " " + hits.totalHits + " hits; topN=" + hitCount + "; reverse=" + reverse + "; sortMissingLast=" + sortMissingLast + " sort=" + sort);
}
// Compute expected results:
Collections.sort(f.matchValues, new Comparator<BytesRef>() {
@Override
public int compare(BytesRef a, BytesRef b) {
if (a == null) {
if (b == null) {
return 0;
}
if (sortMissingLast) {
return 1;
} else {
return -1;
}
} else if (b == null) {
if (sortMissingLast) {
return -1;
} else {
return 1;
}
} else {
return a.compareTo(b);
}
}
});
if (reverse) {
Collections.reverse(f.matchValues);
}
final List<BytesRef> expected = f.matchValues;
if (VERBOSE) {
System.out.println(" expected:");
for (int idx = 0; idx < expected.size(); idx++) {
BytesRef br = expected.get(idx);
if (br == null && missingIsNull == false) {
br = new BytesRef();
}
System.out.println(" " + idx + ": " + (br == null ? "<missing>" : br.utf8ToString()));
if (idx == hitCount - 1) {
break;
}
}
}
if (VERBOSE) {
System.out.println(" actual:");
for (int hitIDX = 0; hitIDX < hits.scoreDocs.length; hitIDX++) {
final FieldDoc fd = (FieldDoc) hits.scoreDocs[hitIDX];
BytesRef br = (BytesRef) fd.fields[0];
System.out.println(" " + hitIDX + ": " + (br == null ? "<missing>" : br.utf8ToString()) + " id=" + s.doc(fd.doc).get("id"));
}
}
for (int hitIDX = 0; hitIDX < hits.scoreDocs.length; hitIDX++) {
final FieldDoc fd = (FieldDoc) hits.scoreDocs[hitIDX];
BytesRef br = expected.get(hitIDX);
if (br == null && missingIsNull == false) {
br = new BytesRef();
}
// Normally, the old codecs (that don't support
// docsWithField via doc values) will always return
// an empty BytesRef for the missing case; however,
// if all docs in a given segment were missing, in
// that case it will return null! So we must map
// null here, too:
BytesRef br2 = (BytesRef) fd.fields[0];
if (br2 == null && missingIsNull == false) {
br2 = new BytesRef();
}
assertEquals(br, br2);
}
}
r.close();
dir.close();
}
Aggregations