Question

我正在学习如何使用Lucene Java API，并为概念验证设置了一个小例子。每当我发送一个查询时，它会给我一个结果：无论查询是什么，字典中的第一项。

我的词典中有89834个术语，这里是前5个。

Field name: "drug"
"cysteine"
"glycine"
"dihydroxyacetone"
"glycerone"
"arginine"
...

搜索“精氨酸”会返回以下内容：

Found 1 results
"cysteine"

我投入的任何东西都会让半胱氨酸成为唯一的结果。

这是代码。

def luceneTest(cxn: RepositoryConnection)
{
    //build the index
    val analyzer: Analyzer = new StandardAnalyzer()
    val indexPath: Path = Paths.get("lucene/model1")
    val directory: Directory = FSDirectory.open(indexPath)
    val config: IndexWriterConfig = new IndexWriterConfig(analyzer)
    val iwriter: IndexWriter = new IndexWriter(directory, config)
    val doc = addCSVtoLuceneIndex("lucene_dictionary.csv")
    iwriter.addDocument(doc)
    iwriter.close()

    //query the index
    val ireader: DirectoryReader = DirectoryReader.open(directory)
    val isearcher: IndexSearcher = new IndexSearcher(ireader)
    val parser: QueryParser = new QueryParser("drug", analyzer)
    val query: Query = parser.parse("arginine")
    val hits: Array[ScoreDoc] = isearcher.search(query, 10).scoreDocs
    logger.info("found " + hits.size + " results.")
    for (a <- hits)
    {
        val hitdoc: Document = isearcher.doc(a.doc)
        logger.info(hitdoc.get("drug"))
    }
    ireader.close()
    directory.close()
}

def addCSVtoLuceneIndex(dictionary: String): Document =
    {
        val doc: Document = new Document()
        val br: BufferedReader = new BufferedReader(new FileReader(dictionary))
        var index = 1
        for (b <- 1 to 83984)
        {
            var line = br.readLine()
            var strArray: Array[String] = line.split(",")
            var strToAdd = ""
            for (a <- 1 to strArray.length - 1) strToAdd += strArray(a)
            doc.add(new Field("drug", strToAdd, TextField.TYPE_STORED))
            //logger.info("added " + strToAdd)
        }
        doc
    }

Answer 1

您在addCSVtoLuceneIndex中迭代文件，但将所有内容放在一个lucene文档中。假设您希望每行都有文档。

你确定要把文件中的每一行都放在没有逗号的地方吗？

Lucene API Query始终返回第一个字典结果

1 个答案: