Lucene3.6.2入门系列

文章15 |   阅读 6823 |   点赞0

来源：https://blog.csdn.net/jadyer/category_1571227.html

【Lucene3.6.2入门系列】第08节_高级搜索之自定义评分

转x33g5p2x  于2021-12-24 转载在 其他  
字(7.1k)|赞(0)|评价(0)|浏览(436)

完整版见 https://jadyer.github.io/2013/08/19/lucene-advanced-search-score/

  1. package com.jadyer.lucene;
  3. import java.io.File;
  4. import java.io.IOException;
  5. import java.util.Random;
  7. import org.apache.lucene.analysis.standard.StandardAnalyzer;
  8. import org.apache.lucene.document.Document;
  9. import org.apache.lucene.document.Field;
  10. import org.apache.lucene.document.NumericField;
  11. import org.apache.lucene.index.CorruptIndexException;
  12. import org.apache.lucene.index.IndexReader;
  13. import org.apache.lucene.index.IndexWriter;
  14. import org.apache.lucene.index.IndexWriterConfig;
  15. import org.apache.lucene.index.Term;
  16. import org.apache.lucene.search.IndexSearcher;
  17. import org.apache.lucene.search.Query;
  18. import org.apache.lucene.search.ScoreDoc;
  19. import org.apache.lucene.search.TermQuery;
  20. import org.apache.lucene.search.TopDocs;
  21. import org.apache.lucene.store.Directory;
  22. import org.apache.lucene.store.FSDirectory;
  23. import org.apache.lucene.util.Version;
  25. import com.jadyer.custom.MyNameScoreQuery;
  27. /**
  28.  * 【Lucene3.6.2入门系列】第08节_高级搜索之自定义评分
  29.  * @create Aug 19, 2013 12:13:14 PM
  30.  * @author 玄玉<http://blog.csdn.net/jadyer>
  31.  */
  32. public class AdvancedSearchByScore {
  33. 	private Directory directory;
  34. 	private IndexReader reader;
  35. 	
  36. 	public AdvancedSearchByScore(){
  37. 		/**文件大小*/
  38. 		int[] sizes = {90, 10, 20, 10, 60, 50};
  39. 		/**文件名*/
  40. 		String[] names = {"Michael.java", "Scofield.ini", "Tbag.txt", "Jack", "Jade", "Jadyer"};
  41. 		/**文件内容*/
  42. 		String[] contents = {"my java blog is http://blog.csdn.net/jadyer",
  43. 							 "my Java Website is http://www.jadyer.cn",
  44. 							 "my name is jadyer",
  45. 							 "I am a Java Developer",
  46. 							 "I am from Haerbin",
  47. 							 "I like java of Lucene"};
  48. 		IndexWriter writer = null;
  49. 		Document doc = null;
  50. 		try {
  51. 			directory = FSDirectory.open(new File("myExample/01_index/"));
  52. 			writer = new IndexWriter(directory, new IndexWriterConfig(Version.LUCENE_36, new StandardAnalyzer(Version.LUCENE_36)));
  53. 			writer.deleteAll();
  54. 			for(int i=0; i<sizes.length; i++){
  55. 				doc = new Document();
  56. 				doc.add(new NumericField("size", Field.Store.YES, true).setIntValue(sizes[i]));
  57. 				doc.add(new Field("name", names[i], Field.Store.YES, Field.Index.ANALYZED_NO_NORMS));
  58. 				doc.add(new Field("content", contents[i], Field.Store.NO, Field.Index.ANALYZED));
  59. 				//添加一个评分域,专门在自定义评分时使用
  60. 				//此时默认为Field.Store.NO和Field.Index.ANALYZED_NO_NORMS
  61. 				doc.add(new NumericField("fileScore").setIntValue(new Random().nextInt(600)));
  62. 				writer.addDocument(doc);
  63. 			}
  64. 		} catch (Exception e) {
  65. 			e.printStackTrace();
  66. 		} finally {
  67. 			if(null != writer){
  68. 				try {
  69. 					writer.close();
  70. 				} catch (IOException ce) {
  71. 					ce.printStackTrace();
  72. 				}
  73. 			}
  74. 		}
  75. 	}
  76. 	
  77. 	/**
  78. 	 * 获取IndexReader实例
  79. 	 */
  80. 	private IndexReader getIndexReader(){
  81. 		try {
  82. 			if(reader == null){
  83. 				reader = IndexReader.open(directory);
  84. 			}else{
  85. 				//if the index was changed since the provided reader was opened, open and return a new reader; else,return null
  86. 				//如果当前reader在打开期间index发生改变,则打开并返回一个新的IndexReader,否则返回null
  87. 				IndexReader ir = IndexReader.openIfChanged(reader);
  88. 				if(ir != null){
  89. 					reader.close(); //关闭原reader
  90. 					reader = ir;    //赋予新reader
  91. 				}
  92. 			}
  93. 			return reader;
  94. 		}catch(Exception e) {
  95. 			e.printStackTrace();
  96. 		}
  97. 		return null; //发生异常则返回null
  98. 	}
  99. 	
  100. 	
  101. 	/**
  102. 	 * 自定义评分搜索
  103. 	 */
  104. 	public void searchByCustomScoreQuery(){
  105. 		IndexSearcher searcher = new IndexSearcher(this.getIndexReader());
  106. //		//创建一个评分域
  107. //		FieldScoreQuery fsq = new FieldScoreQuery("fileScore", FieldScoreQuery.Type.INT);
  108. //		//创建自定义的CustomScoreQuery对象
  109. //		Query query = new MyCustomScoreQuery(new TermQuery(new Term("content", "java")), fsq);
  110. 		Query query = new MyNameScoreQuery(new TermQuery(new Term("content", "java")));
  111. 		try {
  112. 			TopDocs tds = searcher.search(query, 10);
  113. 			for(ScoreDoc sd : tds.scoreDocs){
  114. 				Document doc = searcher.doc(sd.doc);
  115. 				System.out.print("文档编号=" + sd.doc + "  文档权值=" + doc.getBoost() + "  文档评分=" + sd.score + "    ");
  116. 				System.out.println("size=" + doc.get("size") + "  name=" + doc.get("name"));
  117. 			}
  118. 		} catch (CorruptIndexException e) {
  119. 			e.printStackTrace();
  120. 		} catch (IOException e) {
  121. 			e.printStackTrace();
  122. 		} finally {
  123. 			if(searcher != null){
  124. 				try {
  125. 					searcher.close();
  126. 				} catch (IOException e) {
  127. 					e.printStackTrace();
  128. 				}
  129. 			}
  130. 		}
  131. 	}
  132. 	
  133. 	
  134. 	/**
  135. 	 * 测试一下评分效果
  136. 	 */
  137. 	public static void main(String[] args) {
  138. 		new AdvancedSearchByScore().searchByCustomScoreQuery();
  139. 	}
  140. }

下面是我们自定义的评分类MyCustomScoreQuery.java

  1. package com.jadyer.custom;
  3. import java.io.IOException;
  5. import org.apache.lucene.index.IndexReader;
  6. import org.apache.lucene.search.Query;
  7. import org.apache.lucene.search.function.CustomScoreProvider;
  8. import org.apache.lucene.search.function.CustomScoreQuery;
  9. import org.apache.lucene.search.function.ValueSourceQuery;
  11. /**
  12.  * 自定义评分的步骤
  13.  * @see 1)创建一个类继承于CustomScoreQuery
  14.  * @see 2)覆盖CustomScoreQuery.getCustomScoreProvider()方法
  15.  * @see 3)创建一个类继承于CustomScoreProvider
  16.  * @see 4)覆盖CustomScoreProvider.customScore()方法(我们的自定义评分主要就是在此方法中完成的)
  17.  * @create Aug 6, 2013 10:30:46 AM
  18.  * @author 玄玉<http://blog.csdn.net/jadyer>
  19.  */
  20. public class MyCustomScoreQuery extends CustomScoreQuery {
  21. 	private static final long serialVersionUID = -2373017691291184609L;
  22. 	
  23. 	public MyCustomScoreQuery(Query subQuery, ValueSourceQuery valSrcQuery) {
  24. 		//ValueSourceQuery参数就是指专门用来做评分的Query,即评分域的FieldScoreQuery
  25. 		super(subQuery, valSrcQuery);
  26. 	}
  27. 	
  28. 	@Override
  29. 	protected CustomScoreProvider getCustomScoreProvider(IndexReader reader) throws IOException {
  30. 		//如果直接返回super的,就表示使用原有的评分规则,即通过[原有的评分*传入的评分域所获取的评分]来确定最终评分
  31. 		//return super.getCustomScoreProvider(reader);
  32. 		return new MyCustomScoreProvider(reader);
  33. 	}
  34. 	
  35. 	
  36. 	private class MyCustomScoreProvider extends CustomScoreProvider {
  37. 		public MyCustomScoreProvider(IndexReader reader) {
  38. 			super(reader);
  39. 		}
  40. 		@Override
  41. 		public float customScore(int doc, float subQueryScore, float valSrcScore) throws IOException {
  42. 			//subQueryScore--表示默认文档的打分,valSrcScore--表示评分域的打分
  43. 			//该方法的返回值就是文档评分,即ScoreDoc.score获取的结果
  44. 			System.out.println("subQueryScore=" + subQueryScore + "    valSrcScore=" + valSrcScore);
  45. 			return subQueryScore/valSrcScore;
  46. 		}
  47. 	}
  48. }

下面是自定义的采用特殊文件名作为评分标准的评分类MyNameScoreQuery.java

  1. package com.jadyer.custom;
  3. import java.io.IOException;
  5. import org.apache.lucene.index.IndexReader;
  6. import org.apache.lucene.search.FieldCache;
  7. import org.apache.lucene.search.Query;
  8. import org.apache.lucene.search.function.CustomScoreProvider;
  9. import org.apache.lucene.search.function.CustomScoreQuery;
  11. /**
  12.  * 采用特殊文件名作为评分标准
  13.  * @create Aug 6, 2013 2:17:13 PM
  14.  * @author 玄玉<http://blog.csdn.net/jadyer>
  15.  */
  16. public class MyNameScoreQuery extends CustomScoreQuery {
  17. 	private static final long serialVersionUID = -2813985445544972520L;
  18. 	
  19. 	public MyNameScoreQuery(Query subQuery) {
  20. 		//由于这里是打算根据文件名来自定义评分,所以重写构造方法时不必传入评分域的ValueSourceQuery
  21. 		super(subQuery);
  22. 	}
  23. 	
  24. 	@Override
  25. 	protected CustomScoreProvider getCustomScoreProvider(IndexReader reader) throws IOException {
  26. 		return new FilenameScoreProvider(reader);
  27. 	}
  28. 	
  29. 	
  30. 	private class FilenameScoreProvider extends CustomScoreProvider {
  31. 		String[] filenames;
  32. 		public FilenameScoreProvider(IndexReader reader) {
  33. 			super(reader);
  34. 			try {
  35. 				//在IndexReader没有关闭之前,所有的数据都会存储到一个预缓存中(缺点是占用大量内存)
  36. 				//所以我们可以通过预缓存获取name域的值(获取到的是name域所有值,故使用数组)
  37. 				this.filenames = FieldCache.DEFAULT.getStrings(reader, "name");
  38. 			} catch (IOException e) {
  39. 				e.printStackTrace();
  40. 			}
  41. 		}
  42. 		@Override
  43. 		public float customScore(int doc, float subQueryScore, float valSrcScore) throws IOException {
  44. 			//由于FilenameScoreQuery构造方法没有传入ValueSourceQuery,故此处ValueSourceQuery默认为1.0
  45. 			System.out.println("subQueryScore=" + subQueryScore + "    valSrcScore=" + valSrcScore);
  46. 			if(filenames[doc].endsWith(".java") || filenames[doc].endsWith(".ini")){
  47. 				//只加大java文件和ini文件的评分
  48. 				return subQueryScore*1.5f;
  49. 			}else{
  50. 				return subQueryScore/1.5f;
  51. 			}
  52. 		}
  53. 	}
  54. }

版权说明 : 本文为转载文章, 版权归原作者所有 版权申明

原文链接 : https://blog.csdn.net/jadyer/article/details/10068351

内容来源于网络，如有侵权，请联系作者删除！

solrlucene搜索score评分

相关文章

    查看更多
    • 蜀ICP备13028337号-1 大数据知识库 https://www.saoniuhuo.com © All rights reserved
    • 本站内容来源互联网,如果侵犯您的权益请联系我们删除, 联系方式：448109455@qq.com