NEST Library and query issues

[yurehspellflinger (yureh)]

Good day,
I was busy with NEST on elasticsearch, but I have alot of troubling making it work.

I have this code for creating the index

     var typeName = typeof(SEARCHTYPE).Name.ToLower();
     var client = new ElasticClient(Settings);
                client.DeleteIndex(typeName,x=>x.Index(typeName));
                var ax = client.CreateIndex(typeName, c => c
                           .Index(typeName)
                                           .Mappings(m => m.Map<SEARCHTYPE>(d => d.AutoMap())));
            }

and my search is one string with multiple words which should get the best result based on this type. the type has the following fields

title: string
text:string
keywords:string(commaseparated)
identifiers:string(commaseparated)

I want the best result for my search
the priority is as follows:keywords,title identifiers,text

var searches = searchString.Split(' ').Where(x => x.Length > 2);
            var searchItems = new List<string>();
            var synonyms = new List<string>(); //todo
            searchItems.Add(searchString);
            foreach (var searchItem in searches)
            {
                searchItems.Add(searchItem);
                searchItems.AddRange(languageItems.Where(x => x.StartsWith(searchItem)));
            }

            if (!searchItems.Any()) return new ReadOnlyCollection<SearchSuggestion>(new List<SearchSuggestion>());


//Okay, I admit... I need help
            var search=Client.Search<SearchSuggestion>(x => x
                .Query(q => q
                    .Terms(b =>b
                        .Terms(searchItems).Boost(1.6)))                        
                                ); 
            return search.Documents;
        }

Could anybody please help me?
I am having alot of troubles with this...

thank you in advance.

Yureh

[forloop (Russ Cam)]

You should take a look at the Definitive Guide, particularly:

• Mapping and Analysis
• Search in Depth
• Dealing with Human Language

As a very general approach you could

1. map keywords and identifiers as a collection of types in your POCO e.g. IEnumerable<string>
2. analyze these fields at index time in a way that satisfies your problem domain
3. analyze the query input at search time with a full text query to find matches against indexed documents. The analysis applied at search time does not necessarily need to be the same as that applied at index time.