【一】自定义评分
CustomScoreQuery类
构造函数
Constructor Summary | |
---|---|
CustomScoreQuery(Query subQuery) Create a CustomScoreQuery over input subQuery. |
|
CustomScoreQuery(Query subQuery, Query... scoringQueries) Create a CustomScoreQuery over input subQuery and a FunctionQuery . |
|
CustomScoreQuery(Query subQuery, Query scoringQuery) Create a CustomScoreQuery over input subQuery and a FunctionQuery . |
FunctionQuery,FunctionQuery(ValueSource valueSource),自定义评分query;
ValueSource,读取某个域的评分,ValueSource valueSource = new IntFieldSource("score");读取Score域的评分;
通过重写getCustomScoreProvider(AtomicReaderContext)修改评分计算方法
getCustomScoreProvider默认计算方法
super.getCustomScoreProvider(reader):原有评分*传入评分域评分
CustomScoreProvider类
重写 customScore(int doc, float subQueryScore, float valSrcScore)方法,
- doc是文件编号
- subQueryScore是原有评分值
- valSrcScore是传入域的评分值
在这个方法里,可以编写自己的算法,操作两个评分值
1、原有评分*传入的评分域打分
1 Query subquery = new TermQuery(new Term("content", "java"));
2 //评分域
3 ValueSource valueSource = new IntFieldSource("score");
4 FunctionQuery fq = new FunctionQuery(valueSource );//自定义评分query
5 //自定义评分query
6 MyCustomScoreQuery myCustomScoreQuery = new MyCustomScoreQuery(subquery ,fq );
7
8 private class MyCustomScoreQuery extends CustomScoreQuery {
9
10 public MyCustomScoreQuery(Query subQuery , FunctionQuery fq) {
11 super(subQuery ,fq );
12 // TODO Auto-generated constructor stub
13 }
14
15 protected CustomScoreProvider getCustomScoreProvider(AtomicReaderContext reader ) throws IOException {
16 return super .getCustomScoreProvider(reader);
17 }
18 }
2、原有评分/传入评分域打分
1 private class MyCustomScoreQuery extends CustomScoreQuery {
2
3 public MyCustomScoreQuery(Query subQuery , FunctionQuery fq) {
4 super(subQuery ,fq );
5 // TODO Auto-generated constructor stub
6 }
7
8 protected CustomScoreProvider getCustomScoreProvider(AtomicReaderContext reader ) throws IOException {
9 //return super.getCustomScoreProvider(reader);
10 return new MyCustomScoreProvider(reader);
11 }
12 }
13
14 private class MyCustomScoreProvider extends CustomScoreProvider {
15 public MyCustomScoreProvider(AtomicReaderContext reader){
16 super(reader );
17 }
18
19 public float customScore(int doc, float subQueryScore, float valSrcScore) throws IOException{
20 //return super.customScore(doc, subQueryScore, valSrcScore);
21 return subQueryScore /valSrcScore;
22 }
23 }
3、根据文件名评分,指定的特殊文件名获得更高的评分比重
[1]从reader的缓存中获得文件名
Lucene4.0前,Sring[] filenames = FieldCache. DEFAULT.getString( "filename" );
Lucene4.0后,没有了getString,getByte也变成了 Deprecate。
Lucene4.0后,没有了getString,getByte也变成了 Deprecate。
所以采用了下面的方法获取StringField
Binaryvalues filenames = FieldCache.DEFAULT.getTerms( atomicReader, "filename" , false);
[2]针对每个文件名进行操作
String filename = filenames.get(doc ).utf8ToString();