如何使用 ElasticSearch DSL C# 降低与特定字段和值匹配的文档的分数

Question

When I carry out a search I look for documents with a lang code of 'en' and 'gen', im trying to ensure that the 'en' ones appear before the 'gen' ones do. So after looking around, I found this Boosting that might fit the bill as I don't want to completely exclude the 'gen' results..just drop the score a tad.

Here's what i have so far.. any help would be greatly appreciated (still learning about ElasticSearch).

SearchDescriptor<ExpandoObject> searchDescriptor = new SearchDescriptor<ExpandoObject>()
            .Index(!searchModel.IndexesToSearch.Any() ? Indices.All : Indices.Index(searchModel.IndexesToSearch))
            .From(searchModel.IndexOfLastResult)
            .Size(searchModel.AmountOfResults)
            .MinScore(searchModel.MinnimumResultScore)              
            .SearchType(SearchType.DfsQueryThenFetch);
            
            IPromise<IList<ISort>> SortDescriptor(SortDescriptor<ExpandoObject> s) => s.Field(searchModel.SortByField, searchModel.SortByOrder);
List<Func<QueryContainerDescriptor<ExpandoObject>, QueryContainer>> filters = new List<Func<QueryContainerDescriptor<ExpandoObject>, QueryContainer>>();
                 
                foreach (var filter in searchModel.FilterValues)
                {                        
                    filters.Add(fq => fq.Terms(t => t.Field(filter.FieldToSearch.Trim()).Terms(filter.Values)));
                } 
                  searchDescriptor
                    .Query(q => q
                        //.Boosting(b => b
                        //    .Name("language")
                        //    .Negative(qq => qq
                        //        .MatchAll(m => m
                        //            .Name("gen")))
                        //    .NegativeBoost(0.2)
                        .Bool(b => b
                                .Should(s => s 
                                    .MultiMatch(m => m
                                        .Fields( string.Join(',', searchModel.FieldsToSearch))
                                        .Query(searchModel.SearchTerm)))
                                .Filter(filters)
                        )).Sort(SortDescriptor);

This query appears as:

{
  "from": 0,
  "min_score": 0.1,
  "query": {
    "bool": {
      "filter": [
        {
          "terms": {
            "language": [
              "en",
              "gen"
            ]
          }
        }
      ],
      "should": [
        {
          "multi_match": {
            "fields": [
              "title"
            ],
            "query": "foo"
          }
        }
      ]
    }
  },
  "size": 100,
  "sort": [
    {
      "_score": {
        "order": "desc"
      }
    }
  ]
}

Answer 1

You can try using negative boosting query to demote certain documents without excluding them from search results

https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-boosting-query.html#query-dsl-boosting-query

Another approach (I am writing the query, you can replace the same logic in C++):

You can also apply less weightage to the docs which match lang code as 'gen'

{
    "filter": {
         "bool": {
             "must":{"match": {"language": "gen}}
           }
     },
     "weight": 0.50
}