Set default field options in index programmatically

[TomBuilder]

I want to index pdf-attachment for full text search.

The index is added with
new Paper_SearchAttachments().Execute(store, store.Conventions);
after initializing the store.

In the Studio it is possible to add "default field options" to set the analyzer, I want to use "Full-Text Search: Yes" and "Indexing: Search". A statement like "Indexes.Add(x => x.Documents, FieldIndexing.Search);" will not compile because the field Documents is not part of the class Paper, it's only for the index.

How can I add the default field options in code?

That's the code:

using Dms.Database.Models.Paper;
using Raven.Client.Documents.Indexes;
using System.Reflection;

namespace Dms.Database.Indexes;

public class Paper_SearchAttachments : AbstractIndexCreationTask<Paper>
{
   public Paper_SearchAttachments()
   {
      Map = papers => from paper in papers
         let pdf = AttachmentsFor(paper).Where(x => x.Name.EndsWith(".pdf"))
            .Select(x => LoadAttachment(paper, x.Name).GetContentAsStream())
         select new
         {
            Documents = pdf.Select(doc => PdfReader2.GetText(doc)),
            TenantId = paper.TenantId
         };

      AdditionalSources = new Dictionary<string, string>
      {
         {
            "PdfReader2",
            @"
            using System.IO;
            using System.Collections.Generic;
            using UglyToad.PdfPig;
            using UglyToad.PdfPig.Content;

            namespace Dms.Database.Indexes
            {
               public static class PdfReader2
               {
                  public static IEnumerable<string> GetText(Stream stream)
                  {
                     var mem = new MemoryStream();
                     stream.CopyTo(mem);
                     mem.Position = 0;

                     var list = new List<string>();

                     using (var document = PdfDocument.Open(mem))
                     {
                        foreach (var page in document.GetPages())
                        {
                           var pageText = page.Text;

                           foreach (var word in page.GetWords())
                           {
                              yield return word.Text;
                           }
                        }
                     }
                  }
               }
            }"
         }
      };

      AdditionalAssemblies = new HashSet<AdditionalAssembly>
      {
         AdditionalAssembly.FromPath(Path.Combine(Path.GetDirectoryName(Assembly.GetExecutingAssembly().Location), @"indexes\Dlls\UglyToad.PdfPig.dll")),
         AdditionalAssembly.FromPath(Path.Combine(Path.GetDirectoryName(Assembly.GetExecutingAssembly().Location), @"indexes\Dlls\UglyToad.PdfPig.Core.dll")),
         AdditionalAssembly.FromPath(Path.Combine(Path.GetDirectoryName(Assembly.GetExecutingAssembly().Location), @"indexes\Dlls\UglyToad.PdfPig.Tokenization.dll")),
         AdditionalAssembly.FromPath(Path.Combine(Path.GetDirectoryName(Assembly.GetExecutingAssembly().Location), @"indexes\Dlls\UglyToad.PdfPig.Tokens.dll")),
         AdditionalAssembly.FromPath(Path.Combine(Path.GetDirectoryName(Assembly.GetExecutingAssembly().Location), @"indexes\Dlls\UglyToad.PdfPig.Fonts.dll"))
      };
   }
}

[mcrio]

See in docs here how you can define the index structure: https://ravendb.net/docs/article-page/5.4/csharp/indexes/creating-and-deploying#example The index can have a completely different structure than the document you are indexing... (may index different documents)

[TomBuilder]

My question is more about https://ravendb.net/docs/article-page/5.4/csharp/indexes/using-analyzers#manipulating-field-indexing-behavior, but the statement "Indexes.Add(x => x.Documents, FieldIndexing.Exact);" will not work because Documents is not part of the entity. In the studio I can use "default field options" to set the search type, but how to do that progammatically?

[mcrio]

You can extend the AbstractIndexCreationTask that has an additional TReduceResult generics paramter.

Something like this:

public class Paper_SearchAttachments : AbstractIndexCreationTask<Paper, Paper_SearchAttachments.Result>
{
    public Paper_SearchAttachments()
    {
        Map = papers => from paper in papers
            let pdf = AttachmentsFor(paper).Where(x => x.Name.EndsWith(".pdf"))
                .Select(x => LoadAttachment(paper, x.Name).GetContentAsStream())
            select new Result
            {
                Documents = pdf.Select(doc => PdfReader2.GetText(doc)),
                TenantId = paper.TenantId
            };

        Index(result => result.Documents, FieldIndexing.Search);
    }

    public class Result
    {
        public List<string> Documents  { get; set; }

        public string TenantID { get; set; }
    }
}

[garayx]

You can add the following code to the index (after the map):

                Index("Documents", FieldIndexing.Exact);
                Index("__all_fields", FieldIndexing.Search);

[TomBuilder]

Thanks! That's solves my problem.