Lucene .Net SetBoost on Field不影响结果 [英] Lucene .Net SetBoost on Field is not affecting results

问题描述

在将字段添加到文档时，我编写了以下代码来测试SetBoost方法.

I wrote the following code to test the SetBoost method when adding a field to a document.

using System;
using System.Collections.Generic;
using System.IO;
using System.Linq;
using System.Text;
using Lucene;
using Lucene.Net;
using Lucene.Net.Analysis;
using Lucene.Net.Analysis.Standard;
using Lucene.Net.Documents;
using Lucene.Net.Index;
using Lucene.Net.QueryParsers;
using Lucene.Net.Search;
using Lucene.Net.Store;
using Directory = Lucene.Net.Store.Directory;
using Version = Lucene.Net.Util.Version;

namespace LuceneTest
{
    public class LuceneTest
    {
        static void Main(string[] args)
        {
            var product1 = new Document();
            product1.Add(new Field("Id", "1", Field.Store.YES, Field.Index.NOT_ANALYZED));
            var title1 = new Field("title", "Special One", Field.Store.YES, Field.Index.ANALYZED);
            title1.SetBoost(2f);
            product1.Add(title1);
            product1.Add(new Field("synopsis", "special synopsis", Field.Store.YES, Field.Index.ANALYZED));

            var product2 = new Document();
            product2.Add(new Field("Id", "2", Field.Store.YES, Field.Index.NOT_ANALYZED));
            var title2 = new Field("title", "Special Two", Field.Store.YES, Field.Index.ANALYZED);
            title2.SetBoost(2f);
            product2.Add(title2);
            product2.Add(new Field("synopsis", "special synopsis", Field.Store.YES, Field.Index.ANALYZED));

            var product3 = new Document();
            product3.Add(new Field("Id", "3", Field.Store.YES, Field.Index.NOT_ANALYZED));
            var title3 = new Field("title", "Normal One", Field.Store.YES, Field.Index.ANALYZED);
            title3.SetBoost(2f);
            product3.Add(title3);
            product3.Add(new Field("synopsis", "special synopsis", Field.Store.YES, Field.Index.ANALYZED));

            var product4 = new Document();
            product4.Add(new Field("Id", "4", Field.Store.YES, Field.Index.NOT_ANALYZED));
            var title4 = new Field("title", "Normal Two", Field.Store.YES, Field.Index.ANALYZED);
            title4.SetBoost(2f);
            product4.Add(title4);
            product4.Add(new Field("synopsis", "special synopsis", Field.Store.YES, Field.Index.ANALYZED));

            var product5 = new Document();
            product5.Add(new Field("Id", "5", Field.Store.YES, Field.Index.NOT_ANALYZED));
            var title5 = new Field("title", "Special Three", Field.Store.YES, Field.Index.ANALYZED);
            title5.SetBoost(2f);
            product5.Add(title5);
            product5.Add(new Field("synopsis", "normal synopsis", Field.Store.YES, Field.Index.ANALYZED));

            Directory directory = FSDirectory.Open(new DirectoryInfo(Environment.CurrentDirectory + "\\Lucene"));
            Analyzer analyzer = new StandardAnalyzer();
            var writer = new IndexWriter(directory, analyzer, true, IndexWriter.MaxFieldLength.UNLIMITED);

            writer.AddDocument(product1);
            writer.AddDocument(product2);
            writer.AddDocument(product3);
            writer.AddDocument(product4);
            writer.AddDocument(product5);
            writer.Optimize();
            writer.Close();

            Console.WriteLine("searching...");
            var indexReader = IndexReader.Open(directory, true);
            var indexSearcher = new IndexSearcher(indexReader);

            var booleanQuery1 = new BooleanQuery();
            booleanQuery1.Add(new BooleanClause(new PrefixQuery(new Term("title", "special")), BooleanClause.Occur.SHOULD));
            booleanQuery1.Add(new BooleanClause(new PrefixQuery(new Term("synopsis", "special")), BooleanClause.Occur.SHOULD));

            var booleanQuery2 = new BooleanQuery();
            booleanQuery2.Add(new BooleanClause((Query)booleanQuery1, BooleanClause.Occur.MUST));
            TopDocs results = indexSearcher.Search(booleanQuery2, (Filter)null, 200);
            var hits = results.ScoreDocs;

            foreach (var hit in hits)
            {
                var document = indexSearcher.Doc(hit.doc);
                Console.WriteLine(document.Get("Id") + " " + document.Get("title") + " " + hit.score);
            }
            Console.WriteLine("done...");
            Console.ReadLine();
        }

    }   

}

我正在使用Lucene 2.9.4.1.版本.我在标题字段上设置了boost.当我在标题和提要字段中搜索特殊"一词时，我希望产品1、2和5在顶部，但是我得到以下信息:

I am using Lucene version 2.9.4.1. I set boost on the title field. I would expect products 1, 2, and 5 to be at the top when I search for the term "special" in the title and synopsis fields, but instead I get the following:

searching...
1 Special One 1.414214
2 Special Two 1.414214
3 Normal One 0.3535534
4 Normal Two 0.3535534
5 Special Three 0.3535534
done...

产品5与产品3和4的得分相同，即使其标题中带有特殊"一词，也没有提要.

Product 5 has the same score as products 3 and 4 even though it has the term "special" in its title, just not the synopsis.

任何帮助或想法都将不胜感激. 谢谢

Any help or thoughts would be appreciated. Thanks

推荐答案

我认为问题在于您正在使用PrefixQueries.前缀查询被重写为常数计分查询.您可以自己设置重写方法，例如:

I believe the issue is that you are using PrefixQueries. Prefix Query gets rewritten to a constant scoring query. You can set the rewrite method yourself, like:

PrefixQuery pquery = new PrefixQuery(new Term("title", "special"));
pquery.setRewriteMethod(MultiTermQuery.SCORING_BOOLEAN_QUERY_REWRITE);

或者您可以尝试仅使用TermQuery而不是PrefixQuery.无论哪种方式，您都应该看到场级增强生效.

Or you could try just using TermQuery instead of PrefixQuery. Either way you should see the field level boost take effect.

哦，请注意，如果您想了解为什么要对结果进行评分，请查看

Oh, one note, if you wish to understand why results are being scored the way they are, you should take a look at Searcher.explain. Scoring gets complicated, and this is a very handy tool for understanding and tuning it.

这篇关于Lucene .Net SetBoost on Field不影响结果的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！