自定义评分的第一种实现方式，详细内容可以查看代码注释
package com.johnny.lucene04.advance_search.selfScore;
import java.io.IOException;
import java.text.SimpleDateFormat;
import java.util.Date;
import org.apache.lucene.document.Document;
import org.apache.lucene.index.AtomicReaderContext;
import org.apache.lucene.index.DirectoryReader;
import org.apache.lucene.index.Term;
import org.apache.lucene.queries.CustomScoreProvider;
import org.apache.lucene.queries.CustomScoreQuery;
import org.apache.lucene.search.FieldCache;
import org.apache.lucene.search.FieldCache.Longs;
import org.apache.lucene.search.IndexSearcher;
import org.apache.lucene.search.Query;
import org.apache.lucene.search.ScoreDoc;
import org.apache.lucene.search.TermQuery;
import org.apache.lucene.search.TopDocs;
import com.johnny.lucene04.advance_search.FileIndexUtils;
/**
* 自定义评分第一种写法
* （1）创建类并继承CustomScoreQuery
（2）覆盖重写类中的getCusomScoreProvider方法
（3）创建类并继承CustomScoreProvider
（4）覆盖重写类中的customScore确定新的评分规则
* @author Johnny
*
*/
public class MySelfScore {
public void searchBySelfScore(){
try{
IndexSearcher search = new IndexSearcher(DirectoryReader.open(FileIndexUtils.getDirectory()));
Query q = new TermQuery(new Term("content","java"));
MyCustomScoreQuery myQuery = new MyCustomScoreQuery(q);
TopDocs tds = search.search(myQuery, 200);
SimpleDateFormat sdf = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss");
for(ScoreDoc sd:tds.scoreDocs){
Document d = search.doc(sd.doc);
System.out.println(sd.doc+":("+sd.score+")" +
"["+d.get("filename")+"【"+d.get("path")+"】--->"+
d.get("size")+"-----"+sdf.format(new Date(Long.valueOf(d.get("date"))))+"]");
}
System.out.println("-----------Total result:"+tds.scoreDocs.length);
}catch(Exception e){
e.printStackTrace();
}
}
/**
*重写评分的实现方式
* **/
private class MyScoreProvider extends CustomScoreProvider{
private AtomicReaderContext context;
public MyScoreProvider(AtomicReaderContext context) {
super(context);
this.context = context;
}
/**重写评分方法，假定需求为文档size大于1000的评分/1000**/
@Override
public float customScore(int doc, float subQueryScore, float valSrcScore)
throws IOException {
// 从域缓存中加载索引字段信息
Longs longs= FieldCache.DEFAULT.getLongs(context.reader(), "size", false);
//doc实际上就是Lucene中得docId
long size = longs.get(doc);
float ff = 1f;//判断加权
if(size>1000){
ff = 1f/1000;
}
/*
* 通过得分相乘放大分数
* 此处可以控制与原有得分结合的方式，加减乘除都可以
* **/
return subQueryScore*valSrcScore*ff;
}
}
/**
* 重写CustomScoreQuery 的getCustomScoreProvider方法
* 引用自定义的Provider
*/
private class MyCustomScoreQuery extends CustomScoreQuery{
public MyCustomScoreQuery(Query subQuery) {
super(subQuery);
}
@Override
protected CustomScoreProvider getCustomScoreProvider(
AtomicReaderContext context) throws IOException {
/**注册使用自定义的评分实现方式**/
return new MyScoreProvider(context);
}
}
}
按照评分进行排序：
public class ScoreSearch {
/**
* 按照评分进行排序
*/
public void searchByScore(String queryStr,Sort sort){
try{
IndexSearcher search = new IndexSearcher(DirectoryReader.open(FileIndexUtils.getDirectory()));
QueryParser qp = new QueryParser("content", new StandardAnalyzer());
Query q = qp.parse(queryStr);
TopDocs tds = null;
if(sort!=null){
tds = search.search(q, 200,sort);
}else{
tds = search.search(q, 200);
}
for(ScoreDoc sd :tds.scoreDocs){
Document d = search.doc(sd.doc);
System.out.println(sd.doc+":("+sd.score+")" +
"["+d.get("filename")+"【"+d.get("path")+"】---"+d.get("score")+"--->"+
d.get("size")+"]");
}
}catch(Exception e){
e.printStackTrace();
}
}
测试方法如下：
@Test
public void testSearchByScore(){
ScoreSearch ss = new ScoreSearch();
/**sort没有set sort时，默认按照关联性进行排序**/
Sort sort = new Sort();
//sort.setSort(new SortField("score", Type.INT,true));//true表示倒叙，默认和false表示正序
sort.setSort(new SortField("size", Type.LONG),new SortField("score",Type.INT));
ss.searchByScore("java",sort);
}
@Test
public void testSelfScore(){
MySelfScore mss = new MySelfScore();
/**
* 如果使用标准分词器，那么java类中得java.io.IOException会被认为是一个词，不会进行拆分
*/
mss.searchBySelfScore();
}