@Override
public Query prefixQuery(
Object value,
@Nullable MultiTermQuery.RewriteMethod method,
@Nullable QueryParseContext context) {
if (fieldType.indexOptions() != IndexOptions.NONE || context == null) {
return super.prefixQuery(value, method, context);
}
Collection<String> queryTypes = context.queryTypes();
if (queryTypes.size() == 1) {
PrefixQuery prefixQuery =
new PrefixQuery(
new Term(
UidFieldMapper.NAME,
Uid.createUidAsBytes(
Iterables.getFirst(queryTypes, null), BytesRefs.toBytesRef(value))));
if (method != null) {
prefixQuery.setRewriteMethod(method);
}
return prefixQuery;
}
BooleanQuery query = new BooleanQuery();
for (String queryType : queryTypes) {
PrefixQuery prefixQuery =
new PrefixQuery(
new Term(
UidFieldMapper.NAME,
Uid.createUidAsBytes(queryType, BytesRefs.toBytesRef(value))));
if (method != null) {
prefixQuery.setRewriteMethod(method);
}
query.add(prefixQuery, BooleanClause.Occur.SHOULD);
}
return query;
}
@Test(groups = "ch07")
public void testPrefixQuery() throws Exception {
FullTextSession session = Search.getFullTextSession(openSession());
Transaction tx = session.beginTransaction();
buildIndex(session, tx);
String userInput = "sea";
tx = session.beginTransaction();
PrefixQuery query = new PrefixQuery(new Term("title", userInput));
System.out.println(query.toString());
org.hibernate.search.FullTextQuery hibQuery = session.createFullTextQuery(query, Dvd.class);
List<Dvd> results = hibQuery.list();
assert results.size() == 4 : "incorrect hit count";
for (Dvd dvd : results) {
assert dvd.getTitle().indexOf("Sea") >= 0;
System.out.println(dvd.getTitle());
}
for (Object element : session.createQuery("from " + Dvd.class.getName()).list())
session.delete(element);
tx.commit();
session.close();
}
/**
* Construct a span prefix query
*
* @param scope
* @param q
* @return
*/
private SpanQuery buildPrefixQuery(String scope, Query q) {
PrefixQuery pq = (PrefixQuery) q;
ArrayList<SpanQuery> spans = new ArrayList<SpanQuery>();
spans.add(new SpanTermQuery(new Term(scope, getTag(scope, pq.getField(), TAG_TYPE.BEG_TAG))));
spans.add(new SpanTermQuery(new Term(scope, getTag(scope, pq.getField(), TAG_TYPE.END_TAG))));
PrefixQuery prefix = new PrefixQuery(new Term(scope, pq.getPrefix().text()));
spans.add(new SpanMultiTermQueryWrapper<PrefixQuery>(prefix));
return new SpanBetweenQuery(spans.toArray(new SpanQuery[spans.size()]));
}
@Override
public Query prefixQuery(
String value,
@Nullable MultiTermQuery.RewriteMethod method,
@Nullable QueryParseContext context) {
if (indexOptions() != IndexOptions.NONE || context == null) {
return super.prefixQuery(value, method, context);
}
Collection<String> queryTypes = context.queryTypes();
BooleanQuery.Builder query = new BooleanQuery.Builder();
for (String queryType : queryTypes) {
PrefixQuery prefixQuery =
new PrefixQuery(
new Term(
UidFieldMapper.NAME,
Uid.createUidAsBytes(queryType, BytesRefs.toBytesRef(value))));
if (method != null) {
prefixQuery.setRewriteMethod(method);
}
query.add(prefixQuery, BooleanClause.Occur.SHOULD);
}
return query.build();
}
@Test
public void testVectorHighlighterPrefixQuery() throws Exception {
Directory dir = new RAMDirectory();
IndexWriter indexWriter =
new IndexWriter(dir, Lucene.STANDARD_ANALYZER, true, IndexWriter.MaxFieldLength.UNLIMITED);
indexWriter.addDocument(
doc()
.add(field("_id", "1"))
.add(
field(
"content",
"the big bad dog",
Field.Store.YES,
Field.Index.ANALYZED,
Field.TermVector.WITH_POSITIONS_OFFSETS))
.build());
IndexReader reader = indexWriter.getReader();
IndexSearcher searcher = new IndexSearcher(reader);
TopDocs topDocs = searcher.search(new TermQuery(new Term("_id", "1")), 1);
assertThat(topDocs.totalHits, equalTo(1));
FastVectorHighlighter highlighter = new FastVectorHighlighter();
PrefixQuery prefixQuery = new PrefixQuery(new Term("content", "ba"));
assertThat(
prefixQuery.getRewriteMethod().getClass().getName(),
equalTo(PrefixQuery.CONSTANT_SCORE_AUTO_REWRITE_DEFAULT.getClass().getName()));
String fragment =
highlighter.getBestFragment(
highlighter.getFieldQuery(prefixQuery),
reader,
topDocs.scoreDocs[0].doc,
"content",
30);
assertThat(fragment, nullValue());
prefixQuery.setRewriteMethod(PrefixQuery.SCORING_BOOLEAN_QUERY_REWRITE);
Query rewriteQuery = prefixQuery.rewrite(reader);
fragment =
highlighter.getBestFragment(
highlighter.getFieldQuery(rewriteQuery),
reader,
topDocs.scoreDocs[0].doc,
"content",
30);
assertThat(fragment, notNullValue());
System.out.println(fragment);
// now check with the custom field query
prefixQuery = new PrefixQuery(new Term("content", "ba"));
assertThat(
prefixQuery.getRewriteMethod().getClass().getName(),
equalTo(PrefixQuery.CONSTANT_SCORE_AUTO_REWRITE_DEFAULT.getClass().getName()));
CustomFieldQuery.reader.set(reader);
fragment =
highlighter.getBestFragment(
new CustomFieldQuery(prefixQuery, highlighter),
reader,
topDocs.scoreDocs[0].doc,
"content",
30);
assertThat(fragment, notNullValue());
System.out.println(fragment);
}
/**
* Returns a Query instance for doing prefix searches on this field type. Also, other QueryParser
* implementations may have different semantics.
*
* <p>Sub-classes should override this method to provide their own range query implementation.
*
* @param parser the {@link org.apache.solr.search.QParser} calling the method
* @param sf the schema field
* @param termStr the term string for prefix query
* @return a Query instance to perform prefix search
*/
public Query getPrefixQuery(QParser parser, SchemaField sf, String termStr) {
PrefixQuery query = new PrefixQuery(new Term(sf.getName(), termStr));
query.setRewriteMethod(sf.getType().getRewriteMethod(parser, sf));
return query;
}
private static void planPrefixQuery(final StringBuilder builder, final PrefixQuery query) {
builder.append(query.getPrefix());
}
/** @see #toString(Query,IndexSchema) */
public static void toString(Query query, IndexSchema schema, Appendable out, int flags)
throws IOException {
boolean writeBoost = true;
if (query instanceof TermQuery) {
TermQuery q = (TermQuery) query;
Term t = q.getTerm();
FieldType ft = writeFieldName(t.field(), schema, out, flags);
writeFieldVal(t.bytes(), ft, out, flags);
} else if (query instanceof TermRangeQuery) {
TermRangeQuery q = (TermRangeQuery) query;
String fname = q.getField();
FieldType ft = writeFieldName(fname, schema, out, flags);
out.append(q.includesLower() ? '[' : '{');
BytesRef lt = q.getLowerTerm();
BytesRef ut = q.getUpperTerm();
if (lt == null) {
out.append('*');
} else {
writeFieldVal(lt, ft, out, flags);
}
out.append(" TO ");
if (ut == null) {
out.append('*');
} else {
writeFieldVal(ut, ft, out, flags);
}
out.append(q.includesUpper() ? ']' : '}');
} else if (query instanceof NumericRangeQuery) {
NumericRangeQuery q = (NumericRangeQuery) query;
String fname = q.getField();
FieldType ft = writeFieldName(fname, schema, out, flags);
out.append(q.includesMin() ? '[' : '{');
Number lt = q.getMin();
Number ut = q.getMax();
if (lt == null) {
out.append('*');
} else {
out.append(lt.toString());
}
out.append(" TO ");
if (ut == null) {
out.append('*');
} else {
out.append(ut.toString());
}
out.append(q.includesMax() ? ']' : '}');
} else if (query instanceof BooleanQuery) {
BooleanQuery q = (BooleanQuery) query;
boolean needParens = false;
if (q.getBoost() != 1.0 || q.getMinimumNumberShouldMatch() != 0 || q.isCoordDisabled()) {
needParens = true;
}
if (needParens) {
out.append('(');
}
boolean first = true;
for (BooleanClause c : q.clauses()) {
if (!first) {
out.append(' ');
} else {
first = false;
}
if (c.isProhibited()) {
out.append('-');
} else if (c.isRequired()) {
out.append('+');
}
Query subQuery = c.getQuery();
boolean wrapQuery = false;
// TODO: may need to put parens around other types
// of queries too, depending on future syntax.
if (subQuery instanceof BooleanQuery) {
wrapQuery = true;
}
if (wrapQuery) {
out.append('(');
}
toString(subQuery, schema, out, flags);
if (wrapQuery) {
out.append(')');
}
}
if (needParens) {
out.append(')');
}
if (q.getMinimumNumberShouldMatch() > 0) {
out.append('~');
out.append(Integer.toString(q.getMinimumNumberShouldMatch()));
}
if (q.isCoordDisabled()) {
out.append("/no_coord");
}
} else if (query instanceof PrefixQuery) {
PrefixQuery q = (PrefixQuery) query;
Term prefix = q.getPrefix();
FieldType ft = writeFieldName(prefix.field(), schema, out, flags);
out.append(prefix.text());
out.append('*');
} else if (query instanceof WildcardQuery) {
out.append(query.toString());
writeBoost = false;
} else if (query instanceof FuzzyQuery) {
out.append(query.toString());
writeBoost = false;
} else if (query instanceof ConstantScoreQuery) {
out.append(query.toString());
writeBoost = false;
} else if (query instanceof WrappedQuery) {
WrappedQuery q = (WrappedQuery) query;
out.append(q.getOptions());
toString(q.getWrappedQuery(), schema, out, flags);
writeBoost = false; // we don't use the boost on wrapped queries
} else {
out.append(query.getClass().getSimpleName() + '(' + query.toString() + ')');
writeBoost = false;
}
if (writeBoost && query.getBoost() != 1.0f) {
out.append("^");
out.append(Float.toString(query.getBoost()));
}
}
/**
* Extracts all MultiTermQueries for {@code field}, and returns equivalent automata that will
* match terms.
*/
static CharacterRunAutomaton[] extractAutomata(Query query, String field) {
List<CharacterRunAutomaton> list = new ArrayList<>();
if (query instanceof BooleanQuery) {
for (BooleanClause clause : (BooleanQuery) query) {
if (!clause.isProhibited()) {
list.addAll(Arrays.asList(extractAutomata(clause.getQuery(), field)));
}
}
} else if (query instanceof ConstantScoreQuery) {
list.addAll(Arrays.asList(extractAutomata(((ConstantScoreQuery) query).getQuery(), field)));
} else if (query instanceof DisjunctionMaxQuery) {
for (Query sub : ((DisjunctionMaxQuery) query).getDisjuncts()) {
list.addAll(Arrays.asList(extractAutomata(sub, field)));
}
} else if (query instanceof SpanOrQuery) {
for (Query sub : ((SpanOrQuery) query).getClauses()) {
list.addAll(Arrays.asList(extractAutomata(sub, field)));
}
} else if (query instanceof SpanNearQuery) {
for (Query sub : ((SpanNearQuery) query).getClauses()) {
list.addAll(Arrays.asList(extractAutomata(sub, field)));
}
} else if (query instanceof SpanNotQuery) {
list.addAll(Arrays.asList(extractAutomata(((SpanNotQuery) query).getInclude(), field)));
} else if (query instanceof SpanPositionCheckQuery) {
list.addAll(
Arrays.asList(extractAutomata(((SpanPositionCheckQuery) query).getMatch(), field)));
} else if (query instanceof SpanMultiTermQueryWrapper) {
list.addAll(
Arrays.asList(
extractAutomata(((SpanMultiTermQueryWrapper<?>) query).getWrappedQuery(), field)));
} else if (query instanceof AutomatonQuery) {
final AutomatonQuery aq = (AutomatonQuery) query;
if (aq.getField().equals(field)) {
list.add(
new CharacterRunAutomaton(aq.getAutomaton()) {
@Override
public String toString() {
return aq.toString();
}
});
}
} else if (query instanceof PrefixQuery) {
final PrefixQuery pq = (PrefixQuery) query;
Term prefix = pq.getPrefix();
if (prefix.field().equals(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 (fq.getField().equals(field)) {
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 (tq.getField().equals(field)) {
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();
final Comparator<CharsRef> comparator = CharsRef.getUTF16SortedAsUTF8Comparator();
// this is *not* an automaton, but it's 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();
}
});
}
}
return list.toArray(new CharacterRunAutomaton[list.size()]);
}