private static void planFuzzyQuery(final StringBuilder builder, final FuzzyQuery query) {
builder.append(query.getTerm());
builder.append(",prefixLength=");
builder.append(query.getPrefixLength());
builder.append(",minSimilarity=");
builder.append(query.getMinSimilarity());
}
private final void populateTerm(
final HQuery hQuery, final boolean isShould, final boolean isMust, final Query subQueryL)
throws FederatedSearchException {
HTerm hTerm = new HTerm();
hTerm.isShould = isShould;
hTerm.isMust = isMust;
hTerm.boost = subQueryL.getBoost();
hQuery.terms.add(hTerm);
if (subQueryL instanceof TermQuery) {
TermQuery lTerm = (TermQuery) subQueryL;
hTerm.type = lTerm.getTerm().field();
hTerm.text = lTerm.getTerm().text();
} else if (subQueryL instanceof FuzzyQuery) {
FuzzyQuery lTerm = (FuzzyQuery) subQueryL;
hTerm.isFuzzy = true;
hTerm.type = lTerm.getTerm().field();
hTerm.text = lTerm.getTerm().text();
} else if (subQueryL instanceof TermRangeQuery) {
TermRangeQuery lTerm = (TermRangeQuery) subQueryL;
hTerm.isFuzzy = false;
hTerm.type = lTerm.getField();
hTerm.minRange = lTerm.getLowerTerm();
hTerm.maxRange = lTerm.getUpperTerm();
} else {
throw new FederatedSearchException(
"HQueryParser: Not Implemented Query :" + subQueryL.getClass().toString());
}
}
public void assertFromTestData(int codePointTable[]) throws Exception {
InputStream stream = getClass().getResourceAsStream("fuzzyTestData.txt");
BufferedReader reader = new BufferedReader(new InputStreamReader(stream, "UTF-8"));
int bits = Integer.parseInt(reader.readLine());
int terms = (int) Math.pow(2, bits);
Directory dir = newDirectory();
RandomIndexWriter writer =
new RandomIndexWriter(
random,
dir,
newIndexWriterConfig(
TEST_VERSION_CURRENT, new MockAnalyzer(MockTokenizer.KEYWORD, false))
.setMergePolicy(newLogMergePolicy()));
Document doc = new Document();
Field field = newField("field", "", Field.Store.NO, Field.Index.ANALYZED);
doc.add(field);
for (int i = 0; i < terms; i++) {
field.setValue(mapInt(codePointTable, i));
writer.addDocument(doc);
}
IndexReader r = writer.getReader();
IndexSearcher searcher = newSearcher(r);
writer.close();
String line;
while ((line = reader.readLine()) != null) {
String params[] = line.split(",");
String query = mapInt(codePointTable, Integer.parseInt(params[0]));
int prefix = Integer.parseInt(params[1]);
int pqSize = Integer.parseInt(params[2]);
float minScore = Float.parseFloat(params[3]);
FuzzyQuery q = new FuzzyQuery(new Term("field", query), minScore, prefix);
q.setRewriteMethod(new MultiTermQuery.TopTermsBoostOnlyBooleanQueryRewrite(pqSize));
int expectedResults = Integer.parseInt(reader.readLine());
TopDocs docs = searcher.search(q, expectedResults);
assertEquals(expectedResults, docs.totalHits);
for (int i = 0; i < expectedResults; i++) {
String scoreDoc[] = reader.readLine().split(",");
assertEquals(Integer.parseInt(scoreDoc[0]), docs.scoreDocs[i].doc);
assertEquals(Float.parseFloat(scoreDoc[1]), docs.scoreDocs[i].score, epsilon);
}
}
searcher.close();
r.close();
dir.close();
}
private static JsonObject convertFuzzyQuery(FuzzyQuery query) {
String field = query.getTerm().field();
JsonObject fuzzyQuery =
JsonBuilder.object()
.add(
"fuzzy",
JsonBuilder.object()
.add(
field,
JsonBuilder.object()
.addProperty("value", query.getTerm().text())
.addProperty("fuzziness", query.getMaxEdits())
.addProperty("prefix_length", query.getPrefixLength())
.addProperty("boost", query.getBoost())))
.build();
return wrapQueryForNestedIfRequired(field, fuzzyQuery);
}
public void testWildcard() throws Exception {
assertQueryEquals("term*", null, "term*");
assertQueryEquals("term*^2", null, "term*^2.0");
assertQueryEquals("term~", null, "term~0.5");
assertQueryEquals("term~0.7", null, "term~0.7");
assertQueryEquals("term~^2", null, "term~0.5^2.0");
assertQueryEquals("term^2~", null, "term~0.5^2.0");
assertQueryEquals("term*germ", null, "term*germ");
assertQueryEquals("term*germ^3", null, "term*germ^3.0");
assertTrue(getQuery("term*", null) instanceof PrefixQuery);
assertTrue(getQuery("term*^2", null) instanceof PrefixQuery);
assertTrue(getQuery("term~", null) instanceof FuzzyQuery);
assertTrue(getQuery("term~0.7", null) instanceof FuzzyQuery);
FuzzyQuery fq = (FuzzyQuery) getQuery("term~0.7", null);
assertEquals(0.7f, fq.getMinSimilarity(), 0.1f);
assertEquals(FuzzyQuery.defaultPrefixLength, fq.getPrefixLength());
fq = (FuzzyQuery) getQuery("term~", null);
assertEquals(0.5f, fq.getMinSimilarity(), 0.1f);
assertEquals(FuzzyQuery.defaultPrefixLength, fq.getPrefixLength());
assertQueryNodeException("term~1.1"); // value > 1, throws exception
assertTrue(getQuery("term*germ", null) instanceof WildcardQuery);
/*
* Tests to see that wild card terms are (or are not) properly lower-cased
* with propery parser configuration
*/
// First prefix queries:
// by default, convert to lowercase:
assertWildcardQueryEquals("Term*", true, "term*");
// explicitly set lowercase:
assertWildcardQueryEquals("term*", true, "term*");
assertWildcardQueryEquals("Term*", true, "term*");
assertWildcardQueryEquals("TERM*", true, "term*");
// explicitly disable lowercase conversion:
assertWildcardQueryEquals("term*", false, "term*");
assertWildcardQueryEquals("Term*", false, "Term*");
assertWildcardQueryEquals("TERM*", false, "TERM*");
// Then 'full' wildcard queries:
// by default, convert to lowercase:
assertWildcardQueryEquals("Te?m", "te?m");
// explicitly set lowercase:
assertWildcardQueryEquals("te?m", true, "te?m");
assertWildcardQueryEquals("Te?m", true, "te?m");
assertWildcardQueryEquals("TE?M", true, "te?m");
assertWildcardQueryEquals("Te?m*gerM", true, "te?m*germ");
// explicitly disable lowercase conversion:
assertWildcardQueryEquals("te?m", false, "te?m");
assertWildcardQueryEquals("Te?m", false, "Te?m");
assertWildcardQueryEquals("TE?M", false, "TE?M");
assertWildcardQueryEquals("Te?m*gerM", false, "Te?m*gerM");
// Fuzzy queries:
assertWildcardQueryEquals("Term~", "term~0.5");
assertWildcardQueryEquals("Term~", true, "term~0.5");
assertWildcardQueryEquals("Term~", false, "Term~0.5");
// Range queries:
// TODO: implement this on QueryParser
// Q0002E_INVALID_SYNTAX_CANNOT_PARSE: Syntax Error, cannot parse '[A TO
// C]': Lexical error at line 1, column 1. Encountered: "[" (91), after
// : ""
assertWildcardQueryEquals("[A TO C]", "[a TO c]");
assertWildcardQueryEquals("[A TO C]", true, "[a TO c]");
assertWildcardQueryEquals("[A TO C]", false, "[A TO C]");
// Test suffix queries: first disallow
try {
assertWildcardQueryEquals("*Term", true, "*term");
fail();
} catch (QueryNodeException pe) {
// expected exception
}
try {
assertWildcardQueryEquals("?Term", true, "?term");
fail();
} catch (QueryNodeException pe) {
// expected exception
}
// Test suffix queries: then allow
assertWildcardQueryEquals("*Term", true, "*term", true);
assertWildcardQueryEquals("?Term", true, "?term", true);
}
/**
* 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()]);
}
public void testWildcard() throws Exception {
assertQueryEquals("term*", null, "term*");
assertQueryEquals("term*^2", null, "term*^2.0");
assertQueryEquals("term~", null, "term~0.5");
assertQueryEquals("term~0.7", null, "term~0.7");
assertQueryEquals("term~^2", null, "term~0.5^2.0");
assertQueryEquals("term^2~", null, "term~0.5^2.0");
assertQueryEquals("term*germ", null, "term*germ");
assertQueryEquals("term*germ^3", null, "term*germ^3.0");
assertTrue(getQuery("term*", null) instanceof PrefixQuery);
assertTrue(getQuery("term*^2", null) instanceof PrefixQuery);
assertTrue(getQuery("term~", null) instanceof FuzzyQuery);
assertTrue(getQuery("term~0.7", null) instanceof FuzzyQuery);
FuzzyQuery fq = (FuzzyQuery) getQuery("term~0.7", null);
assertEquals(0.7f, fq.getMinSimilarity(), 0.1f);
assertEquals(FuzzyQuery.defaultPrefixLength, fq.getPrefixLength());
fq = (FuzzyQuery) getQuery("term~", null);
assertEquals(0.5f, fq.getMinSimilarity(), 0.1f);
assertEquals(FuzzyQuery.defaultPrefixLength, fq.getPrefixLength());
assertParseException("term~1.1"); // value > 1, throws exception
assertTrue(getQuery("term*germ", null) instanceof WildcardQuery);
/* Tests to see that wild card terms are (or are not) properly
* lower-cased with propery parser configuration
*/
// First prefix queries:
// by default, convert to lowercase:
assertWildcardQueryEquals("Term*", true, "term*");
// explicitly set lowercase:
assertWildcardQueryEquals("term*", true, "term*");
assertWildcardQueryEquals("Term*", true, "term*");
assertWildcardQueryEquals("TERM*", true, "term*");
// explicitly disable lowercase conversion:
assertWildcardQueryEquals("term*", false, "term*");
assertWildcardQueryEquals("Term*", false, "Term*");
assertWildcardQueryEquals("TERM*", false, "TERM*");
// Then 'full' wildcard queries:
// by default, convert to lowercase:
assertWildcardQueryEquals("Te?m", "te?m");
// explicitly set lowercase:
assertWildcardQueryEquals("te?m", true, "te?m");
assertWildcardQueryEquals("Te?m", true, "te?m");
assertWildcardQueryEquals("TE?M", true, "te?m");
assertWildcardQueryEquals("Te?m*gerM", true, "te?m*germ");
// explicitly disable lowercase conversion:
assertWildcardQueryEquals("te?m", false, "te?m");
assertWildcardQueryEquals("Te?m", false, "Te?m");
assertWildcardQueryEquals("TE?M", false, "TE?M");
assertWildcardQueryEquals("Te?m*gerM", false, "Te?m*gerM");
// Fuzzy queries:
assertWildcardQueryEquals("Term~", "term~0.5");
assertWildcardQueryEquals("Term~", true, "term~0.5");
assertWildcardQueryEquals("Term~", false, "Term~0.5");
// Range queries:
assertWildcardQueryEquals("[A TO C]", "[a TO c]");
assertWildcardQueryEquals("[A TO C]", true, "[a TO c]");
assertWildcardQueryEquals("[A TO C]", false, "[A TO C]");
// Test suffix queries: first disallow
try {
assertWildcardQueryEquals("*Term", true, "*term");
fail();
} catch (ParseException pe) {
// expected exception
}
try {
assertWildcardQueryEquals("?Term", true, "?term");
fail();
} catch (ParseException pe) {
// expected exception
}
// Test suffix queries: then allow
assertWildcardQueryEquals("*Term", true, "*term", true);
assertWildcardQueryEquals("?Term", true, "?term", true);
}