0

我尝试使用 lucene 4.3.1 实现基于索引的文本搜索。代码如下。我使用 au NGramTokenyzer 创建了索引，因为我想找到对于 FuzzyQuery 来说太远的搜索结果。我的解决方案有两个问题。首先是我不明白为什么它会找到一些东西，而另一些却没有。例如，如果我寻找“Buter”、“utter”或“Bute”，它会找到“Butter”，但如果我寻找“Btter”则没有结果。我的实现是否有错误，我应该做些什么不同？此外，我希望它总是为每个查询提供（例如）10 个结果。这是否可以通过我的代码实现，或者我需要更改哪些内容才能获得这 10 个结果？

这是代码： 

public LuceneIndex() throws IOException{
    File dir = new File(indexDirectoryPath);
    index = FSDirectory.open(dir);
    analyzer = new NGramAnalyzer();
    config = new IndexWriterConfig(luceneVersion, analyzer);
    indexWriter = new IndexWriter(index, config);
    reader = DirectoryReader.open(FSDirectory.open(dir));
    searcher = new IndexSearcher(reader);
    queryParser = new QueryParser(luceneVersion, "label", new NGramAnalyzer());


}

/**
 *  building the index 
 * @param graph
 * @throws IOException
 */
public void makeIndex(MyGraph graph) throws IOException {
    FieldType fieldType = new FieldType();
    fieldType.setTokenized(true);

    //read the items that should be indexed
    ArrayList<String> DbList = Helper.readListFromFileDb(indexFilePath);
    for (String word : DbList) {
        Document doc = new Document();
        doc.add(new TextField("label", word, Field.Store.YES));
        indexWriter.addDocument(doc);
    }

    indexWriter.close();

}


public void searchIndexWithQueryParser(String searchString, int numberOfResults) throws IOException, ParseException {
System.out.println("Searching for '" + searchString + "' using QueryParser");

Query query = queryParser.parse(searchString);

System.out.println(query.toString());


TopDocs results = searcher.search(query, numberOfResults);
ScoreDoc[] hits = results.scoreDocs;

//just to see some output...
int i = 0;

Document doc = searcher.doc(hits[i].doc);
String label = doc.get("label");

System.out.println(label);
}

编辑：NGramAnalyzer 的代码

public class NGramAnalyzer extends Analyzer {

int minGram = 2;
int maxGram = 2;



@Override
protected TokenStreamComponents createComponents(String fieldName, Reader reader) {

    Tokenizer source = new NGramTokenizer(reader, minGram, maxGram);
    CharArraySet charArraySet = StopFilter.makeStopSet(Version.LUCENE_43,
            FoodProductBlackList.blackList, true);
    TokenStream filter = new StopFilter(Version.LUCENE_43, source, charArraySet);

    return new TokenStreamComponents(source, filter);
}
}
4

0 回答 0