Lucene计数匹配项

Question

我在Lucene_35的字段中搜索。我想知道我的术语中有多少单词与该领域相匹配。 例如，我的领域是“JavaServer Faces（JSF）是一个基于Java的Web应用程序框架，旨在简化基于Web的用户界面的开发集成。”，我的查询词是的 “java/JSF /框架/ doesnotexist”和希望导致3，因为只有的 “java”，“JSF”和“框架”出现在现场。 这里是我下面一个简单的例子：

public void explain(String document, String queryExpr) throws Exception { 

     StandardAnalyzer analyzer = new StandardAnalyzer(Version.LUCENE_35); 
     Directory index = new RAMDirectory(); 
     IndexWriterConfig config = new IndexWriterConfig(Version.LUCENE_35, analyzer); 
     IndexWriter w = new IndexWriter(index, config); 
     addDoc(w, document); 
     w.close(); 
     String queryExpression = queryExpr; 
     Query q = new QueryParser(Version.LUCENE_35, "title", analyzer).parse(queryExpression); 

     System.out.println("Query: " + queryExpression); 
     IndexReader reader = IndexReader.open(index); 
     IndexSearcher searcher = new IndexSearcher(reader); 
     TopDocs topDocs = searcher.search(q, 10); 
     for (int i = 0; i < topDocs.totalHits; i++) { 
      ScoreDoc match = topDocs.scoreDocs[i]; 
      System.out.println("match.score: " + match.score); 
      Explanation explanation = searcher.explain(q, match.doc); //#1 
      System.out.println("----------"); 
      Document doc = searcher.doc(match.doc); 
      System.out.println(doc.get("title")); 
      System.out.println(explanation.toString()); 
     } 
     searcher.close(); 
    } 

与上述参数的输出是：

0.021505041 = (MATCH) product of: 
    0.028673388 = (MATCH) sum of: 
    0.0064956956 = (MATCH) weight(title:java in 0), product of: 
     0.2709602 = queryWeight(title:java), product of: 
     0.30685282 = idf(docFreq=1, maxDocs=1) 
     0.8830299 = queryNorm 

....

 0.033902764 = (MATCH) fieldWeight(title:framework in 0), product of: 
     1.4142135 = tf(termFreq(title:framework)=2) 
     0.30685282 = idf(docFreq=1, maxDocs=1) 
     0.078125 = fieldNorm(field=title, doc=0) 
    0.75 = coord(3/4) 

我想这3/4作为结果。

问候！

Answer 1

您可以通过覆盖Lucene的DefaultSimilarity用下面的方法定义，实现这一目标： - > state.getBoost（）

TF（频率） -

• computeNorm（场，状态）>频率== 0 ？ 0：1个
• IDF（docFreq，numDocs） - > 1
• 坐标（重叠，maxOverlap） - > 1/maxOverlap
• queryNorm（sumOfQuareWeights） - > 1

以这种方式，最终文档的分数结尾是匹配因子（1/maxOverlap）乘以匹配项的数量。

Directory dir = new RAMDirectory(); 

Similarity similarity = new DefaultSimilarity() { 
    @Override 
    public float computeNorm(String fld, FieldInvertState state) { 
    return state.getBoost(); 
    } 

    @Override 
    public float coord(int overlap, int maxOverlap) { 
    return 1f/maxOverlap; 
    } 

    @Override 
    public float idf(int docFreq, int numDocs) { 
    return 1f; 
    } 

    @Override 
    public float queryNorm(float sumOfSquaredWeights) { 
    return 1f; 
    } 

    @Override 
    public float tf(float freq) { 
    return freq == 0f ? 0f : 1f; 
    } 
}; 
IndexWriterConfig iwConf = new IndexWriterConfig(Version.LUCENE_35, 
    new WhitespaceAnalyzer(Version.LUCENE_35)); 
iwConf.setSimilarity(similarity); 
IndexWriter iw = new IndexWriter(dir, iwConf); 
Document doc = new Document(); 
Field field = new Field("text", "", Store.YES, Index.ANALYZED); 
doc.add(field); 
for (String value : Arrays.asList("a b c", "c d", "a b d", "a c d")) { 
    field.setValue(value); 
    iw.addDocument(doc); 
} 
iw.commit(); 
iw.close(); 

IndexReader ir = IndexReader.open(dir); 
IndexSearcher searcher = new IndexSearcher(ir); 
searcher.setSimilarity(similarity); 
BooleanQuery q = new BooleanQuery(); 
q.add(new TermQuery(new Term("text", "a")), Occur.SHOULD); 
q.add(new TermQuery(new Term("text", "b")), Occur.SHOULD); 
q.add(new TermQuery(new Term("text", "d")), Occur.SHOULD); 

TopDocs topDocs = searcher.search(q, 100); 
System.out.println(topDocs.totalHits + " results"); 
ScoreDoc[] scoreDocs = topDocs.scoreDocs; 
for (int i = 0; i < scoreDocs.length; ++i) { 
    int docId = scoreDocs[i].doc; 
    float score = scoreDocs[i].score; 
    System.out.println(ir.document(docId).get("text") + " -> " + score); 
    System.out.println(searcher.explain(q, docId)); 
} 
ir.close();