2

我为此搜索了很多，但直到现在我都找不到任何解决方案。我有一个与邻近搜索相结合的大型查询。我需要找出查询中结果的确切位置在哪里。例如查询的一部分是 "hospital"~2 "readmissio"~2 。现在 Lucene 检索到正确的文档，其中一个包含正确的值“hospital re-admission”，但是我如何在文档中的“re-admission”查询中突出显示“readmissio”。有一些像 levenshtein 这样的解决方法，但由于检索量很大，这是不切实际的，我希望 lucene 中有一些解决方案可以从查询中找出检索到的数据的位置？请指教。

    string newQueryString = "";
    string querystring = "To determine whether high performing hospitals with low 30 day risk standardized hospital readmissio rates have a lower proportion of readmission";
    string[] tokens = NLP.Tokenize(querystring);
    for (int i = 0; i < tokens.Length; i++)
    {
        string token = tokens[i];
        token = "\"" + token + "\"" + "~" + 2;

        // add token to new string expression
        newQueryString = newQueryString + " " + token;
    }

    query = MultiFieldQueryParser.Parse(Lucene.Net.Util.Version.LUCENE_CURRENT,
    new string[] { newQueryString }
    , new string[] { "TERM" },
    selectedAnalyzer);



    TopDocs tp = indexSearcher.Search(query, 1);
    int max = tp.TotalHits;

    List<Lucene.Net.Documents.Document> ids = new List<Lucene.Net.Documents.Document>();
    TopScoreDocCollector collector = TopScoreDocCollector.Create(max+1, true);
    indexSearcher.Search(query, collector);
    ScoreDoc[] hits = collector.TopDocs().ScoreDocs;
    for (int i = 0; i < hits.Length; i++)
    {

        int docId = hits[i].Doc;
        Lucene.Net.Documents.Document doc = indexSearcher.Doc(docId);
        string concept = doc.GetFieldable("TERM").StringValue;


        string[] contentTerms = NLP.Tokenize(querystring);
        ITermFreqVector tfvector = reader.GetTermFreqVector(docId, "TERM");
        TermPositionVector tpvector = (TermPositionVector)tfvector;

        for (int k = 0; k < contentTerms.Length; k++)
        {
            string[] terms = tfvector.GetTerms();
            int termidx = tfvector.IndexOf(contentTerms[k].Trim()); /// How to have readmission/readmissions here????? 
            int[] termposx = tpvector.GetTermPositions(termidx);
            TermVectorOffsetInfo[] tvoffsetinfo = tpvector.GetOffsets(termidx);

            int offsetStart = 0;
            int offsetEnd = 0;
            List<Tokens> conceptList = new List<Tokens>();
            for (int j = 0; j < tvoffsetinfo.Length; j++)
            {
                offsetStart = tvoffsetinfo[j].StartOffset;
                offsetEnd = tvoffsetinfo[j].EndOffset;
                string key = concept.Substring(offsetStart, offsetEnd - offsetStart);
                //////////////// Code continue////////////////
            }

        }
    }

因为它显示了如何在查询字符串中突出显示“重新入院”概念的“redamissi”部分。？？？或者在另一个例子中，如果我的查询中有重新接纳，我如何突出显示它以获取“重新接纳”或“接纳”的检索值？

4

0 回答 0