001package 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
020import java.io.Closeable;
021import java.io.IOException;
022import java.text.ParseException;
023
024import org.apache.lucene.analysis.core.WhitespaceAnalyzer;
025import org.apache.lucene.document.Document;
026import org.apache.lucene.document.DoubleField;
027import org.apache.lucene.document.Field;
028import org.apache.lucene.document.NumericDocValuesField;
029import org.apache.lucene.expressions.Expression;
030import org.apache.lucene.expressions.SimpleBindings;
031import org.apache.lucene.expressions.js.JavascriptCompiler;
032import org.apache.lucene.facet.DrillDownQuery;
033import org.apache.lucene.facet.DrillSideways;
034import org.apache.lucene.facet.FacetResult;
035import org.apache.lucene.facet.Facets;
036import org.apache.lucene.facet.FacetsCollector;
037import org.apache.lucene.facet.FacetsConfig;
038import org.apache.lucene.facet.range.DoubleRange;
039import org.apache.lucene.facet.range.DoubleRangeFacetCounts;
040import org.apache.lucene.facet.taxonomy.TaxonomyReader;
041import org.apache.lucene.index.DirectoryReader;
042import org.apache.lucene.index.IndexWriter;
043import org.apache.lucene.index.IndexWriterConfig;
044import org.apache.lucene.index.IndexWriterConfig.OpenMode;
045import org.apache.lucene.queries.function.ValueSource;
046import org.apache.lucene.search.BooleanClause;
047import org.apache.lucene.search.BooleanQuery;
048import org.apache.lucene.search.Filter;
049import org.apache.lucene.search.IndexSearcher;
050import org.apache.lucene.search.MatchAllDocsQuery;
051import org.apache.lucene.search.NumericRangeQuery;
052import org.apache.lucene.search.QueryWrapperFilter;
053import org.apache.lucene.search.SortField;
054import org.apache.lucene.search.TopDocs;
055import org.apache.lucene.store.Directory;
056import org.apache.lucene.store.RAMDirectory;
057import org.apache.lucene.util.SloppyMath;
058
059/** Shows simple usage of dynamic range faceting, using the
060 *  expressions module to calculate distance. */
061public class DistanceFacetsExample implements Closeable {
062
063  final DoubleRange ONE_KM = new DoubleRange("< 1 km", 0.0, true, 1.0, false);
064  final DoubleRange TWO_KM = new DoubleRange("< 2 km", 0.0, true, 2.0, false);
065  final DoubleRange FIVE_KM = new DoubleRange("< 5 km", 0.0, true, 5.0, false);
066  final DoubleRange TEN_KM = new DoubleRange("< 10 km", 0.0, true, 10.0, false);
067
068  private final Directory indexDir = new RAMDirectory();
069  private IndexSearcher searcher;
070  private final FacetsConfig config = new FacetsConfig();
071
072  /** The "home" latitude. */
073  public final static double ORIGIN_LATITUDE = 40.7143528;
074
075  /** The "home" longitude. */
076  public final static double ORIGIN_LONGITUDE = -74.0059731;
077
078  /** Radius of the Earth in KM
079   *
080   * NOTE: this is approximate, because the earth is a bit
081   * wider at the equator than the poles.  See
082   * http://en.wikipedia.org/wiki/Earth_radius */
083  public final static double EARTH_RADIUS_KM = 6371.01;
084
085  /** Empty constructor */
086  public DistanceFacetsExample() {}
087  
088  /** Build the example index. */
089  public void index() throws IOException {
090    IndexWriter writer = new IndexWriter(indexDir, new IndexWriterConfig(
091        new WhitespaceAnalyzer()).setOpenMode(OpenMode.CREATE));
092
093    // TODO: we could index in radians instead ... saves all the conversions in getBoundingBoxFilter
094
095    // Add documents with latitude/longitude location:
096    // we index these both as DoubleFields (for bounding box/ranges) and as NumericDocValuesFields (for scoring)
097    Document doc = new Document();
098    doc.add(new DoubleField("latitude", 40.759011, Field.Store.NO));
099    doc.add(new NumericDocValuesField("latitude", Double.doubleToRawLongBits(40.759011)));
100    doc.add(new DoubleField("longitude", -73.9844722, Field.Store.NO));
101    doc.add(new NumericDocValuesField("longitude", Double.doubleToRawLongBits(-73.9844722)));
102    writer.addDocument(doc);
103    
104    doc = new Document();
105    doc.add(new DoubleField("latitude", 40.718266, Field.Store.NO));
106    doc.add(new NumericDocValuesField("latitude", Double.doubleToRawLongBits(40.718266)));
107    doc.add(new DoubleField("longitude", -74.007819, Field.Store.NO));
108    doc.add(new NumericDocValuesField("longitude", Double.doubleToRawLongBits(-74.007819)));
109    writer.addDocument(doc);
110    
111    doc = new Document();
112    doc.add(new DoubleField("latitude", 40.7051157, Field.Store.NO));
113    doc.add(new NumericDocValuesField("latitude", Double.doubleToRawLongBits(40.7051157)));
114    doc.add(new DoubleField("longitude", -74.0088305, Field.Store.NO));
115    doc.add(new NumericDocValuesField("longitude", Double.doubleToRawLongBits(-74.0088305)));
116    writer.addDocument(doc);
117
118    // Open near-real-time searcher
119    searcher = new IndexSearcher(DirectoryReader.open(writer, true));
120    writer.close();
121  }
122
123  private ValueSource getDistanceValueSource() {
124    Expression distance;
125    try {
126      distance = JavascriptCompiler.compile(
127                  "haversin(" + ORIGIN_LATITUDE + "," + ORIGIN_LONGITUDE + ",latitude,longitude)");
128    } catch (ParseException pe) {
129      // Should not happen
130      throw new RuntimeException(pe);
131    }
132    SimpleBindings bindings = new SimpleBindings();
133    bindings.add(new SortField("latitude", SortField.Type.DOUBLE));
134    bindings.add(new SortField("longitude", SortField.Type.DOUBLE));
135
136    return distance.getValueSource(bindings);
137  }
138
139  /** Given a latitude and longitude (in degrees) and the
140   *  maximum great circle (surface of the earth) distance,
141   *  returns a simple Filter bounding box to "fast match"
142   *  candidates. */
143  public static Filter getBoundingBoxFilter(double originLat, double originLng, double maxDistanceKM) {
144
145    // Basic bounding box geo math from
146    // http://JanMatuschek.de/LatitudeLongitudeBoundingCoordinates,
147    // licensed under creative commons 3.0:
148    // http://creativecommons.org/licenses/by/3.0
149
150    // TODO: maybe switch to recursive prefix tree instead
151    // (in lucene/spatial)?  It should be more efficient
152    // since it's a 2D trie...
153
154    // Degrees -> Radians:
155    double originLatRadians = Math.toRadians(originLat);
156    double originLngRadians = Math.toRadians(originLng);
157
158    double angle = maxDistanceKM / (SloppyMath.earthDiameter(originLat) / 2.0);
159
160    double minLat = originLatRadians - angle;
161    double maxLat = originLatRadians + angle;
162
163    double minLng;
164    double maxLng;
165    if (minLat > Math.toRadians(-90) && maxLat < Math.toRadians(90)) {
166      double delta = Math.asin(Math.sin(angle)/Math.cos(originLatRadians));
167      minLng = originLngRadians - delta;
168      if (minLng < Math.toRadians(-180)) {
169        minLng += 2 * Math.PI;
170      }
171      maxLng = originLngRadians + delta;
172      if (maxLng > Math.toRadians(180)) {
173        maxLng -= 2 * Math.PI;
174      }
175    } else {
176      // The query includes a pole!
177      minLat = Math.max(minLat, Math.toRadians(-90));
178      maxLat = Math.min(maxLat, Math.toRadians(90));
179      minLng = Math.toRadians(-180);
180      maxLng = Math.toRadians(180);
181    }
182
183    BooleanQuery f = new BooleanQuery();
184
185    // Add latitude range filter:
186    f.add(NumericRangeQuery.newDoubleRange("latitude", Math.toDegrees(minLat), Math.toDegrees(maxLat), true, true),
187          BooleanClause.Occur.FILTER);
188
189    // Add longitude range filter:
190    if (minLng > maxLng) {
191      // The bounding box crosses the international date
192      // line:
193      BooleanQuery lonF = new BooleanQuery();
194      lonF.add(NumericRangeQuery.newDoubleRange("longitude", Math.toDegrees(minLng), null, true, true),
195               BooleanClause.Occur.SHOULD);
196      lonF.add(NumericRangeQuery.newDoubleRange("longitude", null, Math.toDegrees(maxLng), true, true),
197               BooleanClause.Occur.SHOULD);
198      f.add(lonF, BooleanClause.Occur.MUST);
199    } else {
200      f.add(NumericRangeQuery.newDoubleRange("longitude", Math.toDegrees(minLng), Math.toDegrees(maxLng), true, true),
201            BooleanClause.Occur.FILTER);
202    }
203
204    return new QueryWrapperFilter(f);
205  }
206
207  /** User runs a query and counts facets. */
208  public FacetResult search() throws IOException {
209
210    FacetsCollector fc = new FacetsCollector();
211
212    searcher.search(new MatchAllDocsQuery(), fc);
213
214    Facets facets = new DoubleRangeFacetCounts("field", getDistanceValueSource(), fc,
215                                               getBoundingBoxFilter(ORIGIN_LATITUDE, ORIGIN_LONGITUDE, 10.0),
216                                               ONE_KM,
217                                               TWO_KM,
218                                               FIVE_KM,
219                                               TEN_KM);
220
221    return facets.getTopChildren(10, "field");
222  }
223
224  /** User drills down on the specified range. */
225  public TopDocs drillDown(DoubleRange range) throws IOException {
226
227    // Passing no baseQuery means we drill down on all
228    // documents ("browse only"):
229    DrillDownQuery q = new DrillDownQuery(null);
230    final ValueSource vs = getDistanceValueSource();
231    q.add("field", range.getFilter(getBoundingBoxFilter(ORIGIN_LATITUDE, ORIGIN_LONGITUDE, range.max), vs));
232    DrillSideways ds = new DrillSideways(searcher, config, (TaxonomyReader) null) {
233        @Override
234        protected Facets buildFacetsResult(FacetsCollector drillDowns, FacetsCollector[] drillSideways, String[] drillSidewaysDims) throws IOException {        
235          assert drillSideways.length == 1;
236          return new DoubleRangeFacetCounts("field", vs, drillSideways[0], ONE_KM, TWO_KM, FIVE_KM, TEN_KM);
237        }
238      };
239    return ds.search(q, 10).hits;
240  }
241
242  @Override
243  public void close() throws IOException {
244    searcher.getIndexReader().close();
245    indexDir.close();
246  }
247
248  /** Runs the search and drill-down examples and prints the results. */
249  public static void main(String[] args) throws Exception {
250    DistanceFacetsExample example = new DistanceFacetsExample();
251    example.index();
252
253    System.out.println("Distance facet counting example:");
254    System.out.println("-----------------------");
255    System.out.println(example.search());
256
257    System.out.println("Distance facet drill-down example (field/< 2 km):");
258    System.out.println("---------------------------------------------");
259    TopDocs hits = example.drillDown(example.TWO_KM);
260    System.out.println(hits.totalHits + " totalHits");
261
262    example.close();
263  }
264}