001    package org.apache.lucene.demo.facet;
002    
003    /*
004     * Licensed to the Apache Software Foundation (ASF) under one or more
005     * contributor license agreements.  See the NOTICE file distributed with
006     * this work for additional information regarding copyright ownership.
007     * The ASF licenses this file to You under the Apache License, Version 2.0
008     * (the "License"); you may not use this file except in compliance with
009     * the License.  You may obtain a copy of the License at
010     *
011     *     http://www.apache.org/licenses/LICENSE-2.0
012     *
013     * Unless required by applicable law or agreed to in writing, software
014     * distributed under the License is distributed on an "AS IS" BASIS,
015     * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
016     * See the License for the specific language governing permissions and
017     * limitations under the License.
018     */
019    
020    import java.io.IOException;
021    import java.util.ArrayList;
022    import java.util.List;
023    
024    import org.apache.lucene.analysis.core.WhitespaceAnalyzer;
025    import org.apache.lucene.document.Document;
026    import org.apache.lucene.facet.DrillDownQuery;
027    import org.apache.lucene.facet.FacetResult;
028    import org.apache.lucene.facet.Facets;
029    import org.apache.lucene.facet.FacetsCollector;
030    import org.apache.lucene.facet.FacetsConfig;
031    import org.apache.lucene.facet.sortedset.DefaultSortedSetDocValuesReaderState;
032    import org.apache.lucene.facet.sortedset.SortedSetDocValuesFacetCounts;
033    import org.apache.lucene.facet.sortedset.SortedSetDocValuesFacetField;
034    import org.apache.lucene.facet.sortedset.SortedSetDocValuesReaderState;
035    import org.apache.lucene.index.DirectoryReader;
036    import org.apache.lucene.index.IndexWriter;
037    import org.apache.lucene.index.IndexWriterConfig;
038    import org.apache.lucene.search.IndexSearcher;
039    import org.apache.lucene.search.MatchAllDocsQuery;
040    import org.apache.lucene.store.Directory;
041    import org.apache.lucene.store.RAMDirectory;
042    
043    /** Shows simple usage of faceted indexing and search,
044     *  using {@link SortedSetDocValuesFacetField} and {@link
045     *  SortedSetDocValuesFacetCounts}.  */
046    
047    public class SimpleSortedSetFacetsExample {
048    
049      private final Directory indexDir = new RAMDirectory();
050      private final FacetsConfig config = new FacetsConfig();
051    
052      /** Empty constructor */
053      public SimpleSortedSetFacetsExample() {
054      }
055    
056      /** Build the example index. */
057      private void index() throws IOException {
058        IndexWriter indexWriter = new IndexWriter(indexDir, new IndexWriterConfig(FacetExamples.EXAMPLES_VER, 
059            new WhitespaceAnalyzer(FacetExamples.EXAMPLES_VER)));
060        Document doc = new Document();
061        doc.add(new SortedSetDocValuesFacetField("Author", "Bob"));
062        doc.add(new SortedSetDocValuesFacetField("Publish Year", "2010"));
063        indexWriter.addDocument(config.build(doc));
064    
065        doc = new Document();
066        doc.add(new SortedSetDocValuesFacetField("Author", "Lisa"));
067        doc.add(new SortedSetDocValuesFacetField("Publish Year", "2010"));
068        indexWriter.addDocument(config.build(doc));
069    
070        doc = new Document();
071        doc.add(new SortedSetDocValuesFacetField("Author", "Lisa"));
072        doc.add(new SortedSetDocValuesFacetField("Publish Year", "2012"));
073        indexWriter.addDocument(config.build(doc));
074    
075        doc = new Document();
076        doc.add(new SortedSetDocValuesFacetField("Author", "Susan"));
077        doc.add(new SortedSetDocValuesFacetField("Publish Year", "2012"));
078        indexWriter.addDocument(config.build(doc));
079    
080        doc = new Document();
081        doc.add(new SortedSetDocValuesFacetField("Author", "Frank"));
082        doc.add(new SortedSetDocValuesFacetField("Publish Year", "1999"));
083        indexWriter.addDocument(config.build(doc));
084        
085        indexWriter.close();
086      }
087    
088      /** User runs a query and counts facets. */
089      private List<FacetResult> search() throws IOException {
090        DirectoryReader indexReader = DirectoryReader.open(indexDir);
091        IndexSearcher searcher = new IndexSearcher(indexReader);
092        SortedSetDocValuesReaderState state = new DefaultSortedSetDocValuesReaderState(indexReader);
093    
094        // Aggregatses the facet counts
095        FacetsCollector fc = new FacetsCollector();
096    
097        // MatchAllDocsQuery is for "browsing" (counts facets
098        // for all non-deleted docs in the index); normally
099        // you'd use a "normal" query:
100        FacetsCollector.search(searcher, new MatchAllDocsQuery(), 10, fc);
101    
102        // Retrieve results
103        Facets facets = new SortedSetDocValuesFacetCounts(state, fc);
104    
105        List<FacetResult> results = new ArrayList<FacetResult>();
106        results.add(facets.getTopChildren(10, "Author"));
107        results.add(facets.getTopChildren(10, "Publish Year"));
108        indexReader.close();
109        
110        return results;
111      }
112      
113      /** User drills down on 'Publish Year/2010'. */
114      private FacetResult drillDown() throws IOException {
115        DirectoryReader indexReader = DirectoryReader.open(indexDir);
116        IndexSearcher searcher = new IndexSearcher(indexReader);
117        SortedSetDocValuesReaderState state = new DefaultSortedSetDocValuesReaderState(indexReader);
118    
119        // Now user drills down on Publish Year/2010:
120        DrillDownQuery q = new DrillDownQuery(config);
121        q.add("Publish Year", "2010");
122        FacetsCollector fc = new FacetsCollector();
123        FacetsCollector.search(searcher, q, 10, fc);
124    
125        // Retrieve results
126        Facets facets = new SortedSetDocValuesFacetCounts(state, fc);
127        FacetResult result = facets.getTopChildren(10, "Author");
128        indexReader.close();
129        
130        return result;
131      }
132    
133      /** Runs the search example. */
134      public List<FacetResult> runSearch() throws IOException {
135        index();
136        return search();
137      }
138      
139      /** Runs the drill-down example. */
140      public FacetResult runDrillDown() throws IOException {
141        index();
142        return drillDown();
143      }
144    
145      /** Runs the search and drill-down examples and prints the results. */
146      public static void main(String[] args) throws Exception {
147        System.out.println("Facet counting example:");
148        System.out.println("-----------------------");
149        SimpleSortedSetFacetsExample example = new SimpleSortedSetFacetsExample();
150        List<FacetResult> results = example.runSearch();
151        System.out.println("Author: " + results.get(0));
152        System.out.println("Publish Year: " + results.get(0));
153    
154        System.out.println("\n");
155        System.out.println("Facet drill-down example (Publish Year/2010):");
156        System.out.println("---------------------------------------------");
157        System.out.println("Author: " + example.runDrillDown());
158      }
159    }