1 package org.apache.lucene.queryParser;
4 * Licensed to the Apache Software Foundation (ASF) under one or more
5 * contributor license agreements. See the NOTICE file distributed with
6 * this work for additional information regarding copyright ownership.
7 * The ASF licenses this file to You under the Apache License, Version 2.0
8 * (the "License"); you may not use this file except in compliance with
9 * the License. You may obtain a copy of the License at
11 * http://www.apache.org/licenses/LICENSE-2.0
13 * Unless required by applicable law or agreed to in writing, software
14 * distributed under the License is distributed on an "AS IS" BASIS,
15 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
16 * See the License for the specific language governing permissions and
17 * limitations under the License.
20 import java.io.IOException;
21 import java.io.Reader;
23 import org.apache.lucene.analysis.Analyzer;
24 import org.apache.lucene.analysis.BaseTokenStreamTestCase;
25 import org.apache.lucene.analysis.LowerCaseFilter;
26 import org.apache.lucene.analysis.TokenFilter;
27 import org.apache.lucene.analysis.TokenStream;
28 import org.apache.lucene.analysis.standard.StandardTokenizer;
29 import org.apache.lucene.analysis.tokenattributes.OffsetAttribute;
30 import org.apache.lucene.analysis.tokenattributes.PositionIncrementAttribute;
31 import org.apache.lucene.analysis.tokenattributes.CharTermAttribute;
32 import org.apache.lucene.analysis.tokenattributes.TypeAttribute;
33 import org.apache.lucene.search.Query;
34 import org.apache.lucene.util.Version;
37 * Test QueryParser's ability to deal with Analyzers that return more
38 * than one token per position or that return tokens with a position
42 public class TestMultiAnalyzer extends BaseTokenStreamTestCase {
44 private static int multiToken = 0;
46 public void testMultiAnalyzer() throws ParseException {
48 QueryParser qp = new QueryParser(TEST_VERSION_CURRENT, "", new MultiAnalyzer());
50 // trivial, no multiple tokens:
51 assertEquals("foo", qp.parse("foo").toString());
52 assertEquals("foo", qp.parse("\"foo\"").toString());
53 assertEquals("foo foobar", qp.parse("foo foobar").toString());
54 assertEquals("\"foo foobar\"", qp.parse("\"foo foobar\"").toString());
55 assertEquals("\"foo foobar blah\"", qp.parse("\"foo foobar blah\"").toString());
57 // two tokens at the same position:
58 assertEquals("(multi multi2) foo", qp.parse("multi foo").toString());
59 assertEquals("foo (multi multi2)", qp.parse("foo multi").toString());
60 assertEquals("(multi multi2) (multi multi2)", qp.parse("multi multi").toString());
61 assertEquals("+(foo (multi multi2)) +(bar (multi multi2))",
62 qp.parse("+(foo multi) +(bar multi)").toString());
63 assertEquals("+(foo (multi multi2)) field:\"bar (multi multi2)\"",
64 qp.parse("+(foo multi) field:\"bar multi\"").toString());
67 assertEquals("\"(multi multi2) foo\"", qp.parse("\"multi foo\"").toString());
68 assertEquals("\"foo (multi multi2)\"", qp.parse("\"foo multi\"").toString());
69 assertEquals("\"foo (multi multi2) foobar (multi multi2)\"",
70 qp.parse("\"foo multi foobar multi\"").toString());
73 assertEquals("(field:multi field:multi2) field:foo", qp.parse("field:multi field:foo").toString());
74 assertEquals("field:\"(multi multi2) foo\"", qp.parse("field:\"multi foo\"").toString());
76 // three tokens at one position:
77 assertEquals("triplemulti multi3 multi2", qp.parse("triplemulti").toString());
78 assertEquals("foo (triplemulti multi3 multi2) foobar",
79 qp.parse("foo triplemulti foobar").toString());
81 // phrase with non-default slop:
82 assertEquals("\"(multi multi2) foo\"~10", qp.parse("\"multi foo\"~10").toString());
84 // phrase with non-default boost:
85 assertEquals("\"(multi multi2) foo\"^2.0", qp.parse("\"multi foo\"^2").toString());
87 // phrase after changing default slop
89 assertEquals("\"(multi multi2) foo\"~99 bar",
90 qp.parse("\"multi foo\" bar").toString());
91 assertEquals("\"(multi multi2) foo\"~99 \"foo bar\"~2",
92 qp.parse("\"multi foo\" \"foo bar\"~2").toString());
95 // non-default operator:
96 qp.setDefaultOperator(QueryParser.AND_OPERATOR);
97 assertEquals("+(multi multi2) +foo", qp.parse("multi foo").toString());
101 public void testMultiAnalyzerWithSubclassOfQueryParser() throws ParseException {
103 DumbQueryParser qp = new DumbQueryParser("", new MultiAnalyzer());
104 qp.setPhraseSlop(99); // modified default slop
106 // direct call to (super's) getFieldQuery to demonstrate differnce
107 // between phrase and multiphrase with modified default slop
108 assertEquals("\"foo bar\"~99",
109 qp.getSuperFieldQuery("","foo bar", true).toString());
110 assertEquals("\"(multi multi2) bar\"~99",
111 qp.getSuperFieldQuery("","multi bar", true).toString());
114 // ask sublcass to parse phrase with modified default slop
115 assertEquals("\"(multi multi2) foo\"~99 bar",
116 qp.parse("\"multi foo\" bar").toString());
120 public void testPosIncrementAnalyzer() throws ParseException {
121 QueryParser qp = new QueryParser(Version.LUCENE_24, "", new PosIncrementAnalyzer());
122 assertEquals("quick brown", qp.parse("the quick brown").toString());
123 assertEquals("\"quick brown\"", qp.parse("\"the quick brown\"").toString());
124 assertEquals("quick brown fox", qp.parse("the quick brown fox").toString());
125 assertEquals("\"quick brown fox\"", qp.parse("\"the quick brown fox\"").toString());
129 * Expands "multi" to "multi" and "multi2", both at the same position,
130 * and expands "triplemulti" to "triplemulti", "multi3", and "multi2".
132 private class MultiAnalyzer extends Analyzer {
134 public MultiAnalyzer() {
138 public TokenStream tokenStream(String fieldName, Reader reader) {
139 TokenStream result = new StandardTokenizer(TEST_VERSION_CURRENT, reader);
140 result = new TestFilter(result);
141 result = new LowerCaseFilter(TEST_VERSION_CURRENT, result);
146 private final class TestFilter extends TokenFilter {
148 private String prevType;
149 private int prevStartOffset;
150 private int prevEndOffset;
152 private final CharTermAttribute termAtt;
153 private final PositionIncrementAttribute posIncrAtt;
154 private final OffsetAttribute offsetAtt;
155 private final TypeAttribute typeAtt;
157 public TestFilter(TokenStream in) {
159 termAtt = addAttribute(CharTermAttribute.class);
160 posIncrAtt = addAttribute(PositionIncrementAttribute.class);
161 offsetAtt = addAttribute(OffsetAttribute.class);
162 typeAtt = addAttribute(TypeAttribute.class);
166 public final boolean incrementToken() throws java.io.IOException {
167 if (multiToken > 0) {
168 termAtt.setEmpty().append("multi"+(multiToken+1));
169 offsetAtt.setOffset(prevStartOffset, prevEndOffset);
170 typeAtt.setType(prevType);
171 posIncrAtt.setPositionIncrement(0);
175 boolean next = input.incrementToken();
179 prevType = typeAtt.type();
180 prevStartOffset = offsetAtt.startOffset();
181 prevEndOffset = offsetAtt.endOffset();
182 String text = termAtt.toString();
183 if (text.equals("triplemulti")) {
186 } else if (text.equals("multi")) {
195 public void reset() throws IOException {
197 this.prevType = null;
198 this.prevStartOffset = 0;
199 this.prevEndOffset = 0;
204 * Analyzes "the quick brown" as: quick(incr=2) brown(incr=1).
205 * Does not work correctly for input other than "the quick brown ...".
207 private class PosIncrementAnalyzer extends Analyzer {
209 public PosIncrementAnalyzer() {
213 public TokenStream tokenStream(String fieldName, Reader reader) {
214 TokenStream result = new StandardTokenizer(TEST_VERSION_CURRENT, reader);
215 result = new TestPosIncrementFilter(result);
216 result = new LowerCaseFilter(TEST_VERSION_CURRENT, result);
221 private final class TestPosIncrementFilter extends TokenFilter {
223 CharTermAttribute termAtt;
224 PositionIncrementAttribute posIncrAtt;
226 public TestPosIncrementFilter(TokenStream in) {
228 termAtt = addAttribute(CharTermAttribute.class);
229 posIncrAtt = addAttribute(PositionIncrementAttribute.class);
233 public final boolean incrementToken () throws java.io.IOException {
234 while(input.incrementToken()) {
235 if (termAtt.toString().equals("the")) {
236 // stopword, do nothing
237 } else if (termAtt.toString().equals("quick")) {
238 posIncrAtt.setPositionIncrement(2);
241 posIncrAtt.setPositionIncrement(1);
249 /** a very simple subclass of QueryParser */
250 private final static class DumbQueryParser extends QueryParser {
252 public DumbQueryParser(String f, Analyzer a) {
253 super(TEST_VERSION_CURRENT, f, a);
256 /** expose super's version */
257 public Query getSuperFieldQuery(String f, String t, boolean quoted)
258 throws ParseException {
259 return super.getFieldQuery(f,t,quoted);
261 /** wrap super's version */
263 protected Query getFieldQuery(String f, String t, boolean quoted)
264 throws ParseException {
265 return new DumbQueryWrapper(getSuperFieldQuery(f,t,quoted));
270 * A very simple wrapper to prevent instanceof checks but uses
271 * the toString of the query it wraps.
273 private final static class DumbQueryWrapper extends Query {
276 public DumbQueryWrapper(Query q) {
281 public String toString(String f) {
282 return q.toString(f);