1 package org.apache.lucene.search.highlight;
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.util.HashMap;
22 import java.util.HashSet;
26 import org.apache.lucene.analysis.CachingTokenFilter;
27 import org.apache.lucene.analysis.TokenStream;
28 import org.apache.lucene.analysis.tokenattributes.CharTermAttribute;
29 import org.apache.lucene.analysis.tokenattributes.PositionIncrementAttribute;
30 import org.apache.lucene.index.IndexReader;
31 import org.apache.lucene.index.memory.MemoryIndex;
32 import org.apache.lucene.search.Query;
33 import org.apache.lucene.search.spans.SpanQuery;
34 import org.apache.lucene.util.StringHelper;
37 * {@link Scorer} implementation which scores text fragments by the number of
38 * unique query terms found. This class converts appropriate {@link Query}s to
39 * {@link SpanQuery}s and attempts to score only those terms that participated in
40 * generating the 'hit' on the document.
42 public class QueryScorer implements Scorer {
43 private float totalScore;
44 private Set<String> foundTerms;
45 private Map<String,WeightedSpanTerm> fieldWeightedSpanTerms;
46 private float maxTermWeight;
47 private int position = -1;
48 private String defaultField;
49 private CharTermAttribute termAtt;
50 private PositionIncrementAttribute posIncAtt;
51 private boolean expandMultiTermQuery = true;
54 private IndexReader reader;
55 private boolean skipInitExtractor;
56 private boolean wrapToCaching = true;
57 private int maxCharsToAnalyze;
60 * @param query Query to use for highlighting
62 public QueryScorer(Query query) {
63 init(query, null, null, true);
67 * @param query Query to use for highlighting
68 * @param field Field to highlight - pass null to ignore fields
70 public QueryScorer(Query query, String field) {
71 init(query, field, null, true);
75 * @param query Query to use for highlighting
76 * @param field Field to highlight - pass null to ignore fields
77 * @param reader {@link IndexReader} to use for quasi tf/idf scoring
79 public QueryScorer(Query query, IndexReader reader, String field) {
80 init(query, field, reader, true);
85 * @param query to use for highlighting
86 * @param reader {@link IndexReader} to use for quasi tf/idf scoring
87 * @param field to highlight - pass null to ignore fields
90 public QueryScorer(Query query, IndexReader reader, String field, String defaultField) {
91 this.defaultField = StringHelper.intern(defaultField);
92 init(query, field, reader, true);
96 * @param defaultField - The default field for queries with the field name unspecified
98 public QueryScorer(Query query, String field, String defaultField) {
99 this.defaultField = StringHelper.intern(defaultField);
100 init(query, field, null, true);
104 * @param weightedTerms an array of pre-created {@link WeightedSpanTerm}s
106 public QueryScorer(WeightedSpanTerm[] weightedTerms) {
107 this.fieldWeightedSpanTerms = new HashMap<String,WeightedSpanTerm>(weightedTerms.length);
109 for (int i = 0; i < weightedTerms.length; i++) {
110 WeightedSpanTerm existingTerm = fieldWeightedSpanTerms.get(weightedTerms[i].term);
112 if ((existingTerm == null) ||
113 (existingTerm.weight < weightedTerms[i].weight)) {
114 // if a term is defined more than once, always use the highest
116 fieldWeightedSpanTerms.put(weightedTerms[i].term, weightedTerms[i]);
117 maxTermWeight = Math.max(maxTermWeight, weightedTerms[i].getWeight());
120 skipInitExtractor = true;
126 * @see org.apache.lucene.search.highlight.Scorer#getFragmentScore()
128 public float getFragmentScore() {
134 * @return The highest weighted term (useful for passing to
135 * GradientFormatter to set top end of coloring scale).
137 public float getMaxTermWeight() {
138 return maxTermWeight;
144 * @see org.apache.lucene.search.highlight.Scorer#getTokenScore(org.apache.lucene.analysis.Token,
147 public float getTokenScore() {
148 position += posIncAtt.getPositionIncrement();
149 String termText = termAtt.toString();
151 WeightedSpanTerm weightedSpanTerm;
153 if ((weightedSpanTerm = fieldWeightedSpanTerms.get(
154 termText)) == null) {
158 if (weightedSpanTerm.positionSensitive &&
159 !weightedSpanTerm.checkPosition(position)) {
163 float score = weightedSpanTerm.getWeight();
165 // found a query term - is it unique in this doc?
166 if (!foundTerms.contains(termText)) {
168 foundTerms.add(termText);
175 * @see org.apache.lucene.search.highlight.Scorer#init(org.apache.lucene.analysis.TokenStream)
177 public TokenStream init(TokenStream tokenStream) throws IOException {
179 termAtt = tokenStream.addAttribute(CharTermAttribute.class);
180 posIncAtt = tokenStream.addAttribute(PositionIncrementAttribute.class);
181 if(!skipInitExtractor) {
182 if(fieldWeightedSpanTerms != null) {
183 fieldWeightedSpanTerms.clear();
185 return initExtractor(tokenStream);
191 * Retrieve the {@link WeightedSpanTerm} for the specified token. Useful for passing
192 * Span information to a {@link Fragmenter}.
194 * @param token to get {@link WeightedSpanTerm} for
195 * @return WeightedSpanTerm for token
197 public WeightedSpanTerm getWeightedSpanTerm(String token) {
198 return fieldWeightedSpanTerms.get(token);
203 private void init(Query query, String field, IndexReader reader, boolean expandMultiTermQuery) {
204 this.reader = reader;
205 this.expandMultiTermQuery = expandMultiTermQuery;
210 private TokenStream initExtractor(TokenStream tokenStream) throws IOException {
211 WeightedSpanTermExtractor qse = defaultField == null ? new WeightedSpanTermExtractor()
212 : new WeightedSpanTermExtractor(defaultField);
213 qse.setMaxDocCharsToAnalyze(maxCharsToAnalyze);
214 qse.setExpandMultiTermQuery(expandMultiTermQuery);
215 qse.setWrapIfNotCachingTokenFilter(wrapToCaching);
216 if (reader == null) {
217 this.fieldWeightedSpanTerms = qse.getWeightedSpanTerms(query,
220 this.fieldWeightedSpanTerms = qse.getWeightedSpanTermsWithScores(query,
221 tokenStream, field, reader);
223 if(qse.isCachedTokenStream()) {
224 return qse.getTokenStream();
233 * @see org.apache.lucene.search.highlight.Scorer#startFragment(org.apache.lucene.search.highlight.TextFragment)
235 public void startFragment(TextFragment newFragment) {
236 foundTerms = new HashSet<String>();
241 * @return true if multi-term queries should be expanded
243 public boolean isExpandMultiTermQuery() {
244 return expandMultiTermQuery;
248 * Controls whether or not multi-term queries are expanded
249 * against a {@link MemoryIndex} {@link IndexReader}.
251 * @param expandMultiTermQuery true if multi-term queries should be expanded
253 public void setExpandMultiTermQuery(boolean expandMultiTermQuery) {
254 this.expandMultiTermQuery = expandMultiTermQuery;
258 * By default, {@link TokenStream}s that are not of the type
259 * {@link CachingTokenFilter} are wrapped in a {@link CachingTokenFilter} to
260 * ensure an efficient reset - if you are already using a different caching
261 * {@link TokenStream} impl and you don't want it to be wrapped, set this to
266 public void setWrapIfNotCachingTokenFilter(boolean wrap) {
267 this.wrapToCaching = wrap;
270 public void setMaxDocCharsToAnalyze(int maxDocCharsToAnalyze) {
271 this.maxCharsToAnalyze = maxDocCharsToAnalyze;