Question

我正在尝试通过错别字错误对文档进行简单查询。

我已使用BooleanQuery将其存档，并为每个字段添加了FuzzyQuery。但是现在我的问题是得分。

据我了解，评分是根据每个字段的所有FuzzyQuery的总和来计算的。但是我只想保留最好的一个，而不要做这个加法。

例如，如果我搜索“ manticore”，我想从字段名称中获得最佳结果。

查询

SearchEngine(final String indexStoragePath) throws IOException {
    MMapDirectory mMapDirectory = new MMapDirectory(Paths.get(indexStoragePath));
    searcherManager = new SearcherManager(mMapDirectory, new SearcherFactory());
    indexSearcher = searcherManager.acquire();
}

List<String> getMatchingIds(final String searchQuery, int maxSearch) throws IOException {
    BooleanQuery.Builder builder = new BooleanQuery.Builder();

    for (Map.Entry<String, Integer> input : FIELDS.entrySet()) {
        FuzzyQuery query = new FuzzyQuery(new Term(input.getKey(), searchQuery), input.getValue());
        builder.add(query, BooleanClause.Occur.SHOULD);
    }

    BooleanQuery resultQuery = builder.build();

    TopDocs docs = indexSearcher.search(resultQuery, maxSearch);
    ScoreDoc[] hits = docs.scoreDocs;

    List<String> result = Lists.newArrayList();
    for (ScoreDoc hit : hits) {
        Document document = getDocument(hit);
        result.add(document.get(ID));
    }

    return result;
}

已编入索引的

Indexer(final String indexStoragePath) throws IOException {
    MMapDirectory mMapDirectory = new MMapDirectory(Paths.get(indexStoragePath));
    StandardAnalyzer analyzer = new StandardAnalyzer();
    IndexWriterConfig config = new IndexWriterConfig(analyzer);
    config.setOpenMode(IndexWriterConfig.OpenMode.CREATE_OR_APPEND);

    MultiTermQuery.TopTermsScoringBooleanQueryRewrite
    config.setSimilarity(new BooleanSimilarity());

    writer = new IndexWriter(mMapDirectory, config);
    writer.commit();
}

如何修改此评分系统？

lucene-多字段查询-基于最佳字段的评分

0 个答案: