使 Lucene.net Highlighter 仅显示来自特定字段的匹配项

Question

如果我有一个包含这些内容的 fullText 字段

In 2014 and 2015 the results were ... [more] ... and Sony are developing ... [more]

并查询

+loadTime:[2014 TO 2015] +fullText:sony

荧光笔选择 2014 年和 2015 年 作为最佳片段。如何让荧光笔忽略来自查询的 loadTime 部分的匹配项并使用来自搜索的 fullText 部分的匹配项？我想查看 ... sony ... 片段，即使它的分数低于（刚刚发生的）匹配全文的日期部分。

我的代码：

ScoreDoc[] hits = [create search];
IFormatter formatter = new SimpleHTMLFormatter("<b>", "</b>");
QueryScorer scorer = new QueryScorer(query, );
Highlighter highlighter = new Highlighter(formatter, scorer);

for (int i = 0; i < hits.Length; i++)
{
    int docId = hits[i].Doc;
    float score = hits[i].Score;
    Document doc = search.Doc(docId);

    string fragments = string.Empty;
    if (collectFragments)
    {
        TokenStream stream = _analyzer.TokenStream("", new StringReader(doc.Get(AppConstants.Fields.FullText)));
        fragments = highlighter.GetBestFragments(stream, doc.Get(AppConstants.Fields.FullText), 2, "...");
    }

    ...
}

Answer 1

表达式“+loadTime:[2014 TO 2015] +fullText:sony”似乎意味着您想要匹配 loadTime 介于 2014 年和 2015 年之间并且在全文中包含 Sony 的文档。 抱歉，我阅读了 Lucene In Action(3.1.2 Parsing a user-entered query expression: QueryParser) 并查看了 queryparsersyntax.html，但没有找到编写像您这样的查询表达式的方法。最接近的是

loadTime:[2014 TO 2015] AND fullText:sony

可能因为版本问题，我的是Lucene 3.4.0。 而解决问题的方法可能是QueryScore

/**
 * @param query Query to use for highlighting
 * @param field Field to highlight - pass null to ignore fields
*/
public QueryScorer(Query query, String field) {
  init(query, field, null, true);
}

在您的代码中，我看到您未填写该参数。我在我的 Intellij 上尝试了 Scala 并且成功了，下面是代码。

def singleFieldHighlighter = {
    val textToDivide = "In 2016 and 2017 the results were ... [more] ... and Sony are developing ... [more]"
    val tokenStream = new StandardAnalyzer(Version.LUCENE_30).tokenStream("fullText", new StringReader(textToDivide));
    val searchString = "loadTime:[2014 TO 2015] AND fullText:sony"
    val parser = new QueryParser(version, "fullText", new WhitespaceAnalyzer(version))
    val parsedQuery = parser.parse(searchString)

    val scorer = new QueryScorer(parsedQuery, "fullText")
    val formatter = new SimpleHTMLFormatter("<span class='highlight'>", "</span>")
    val highlighter = new Highlighter(formatter, scorer)
    highlighter.setTextFragmenter(new SimpleSpanFragmenter(scorer))
    highlighter.getBestFragments(tokenStream, textToDivide, 3, "...")
}


// The result
// In 2014 and 2015 the results were ... [more] ... and <span class='highlight'>Sony</span> are developing ... [more]

希望对你有帮助，如果没有，我建议你阅读我上面提到的 Lucene In Action 一章来弄清楚。