ChatRAG/Data/TextData.cs

using ChatRAG.Data;
using ChatRAG.Models;
using Microsoft.SemanticKernel;
using Microsoft.SemanticKernel.Embeddings;
using System.Text;

#pragma warning disable SKEXP0001 // Type is for evaluation purposes only and is subject to change or removal in future updates. Suppress this diagnostic to proceed.

namespace ChatApi.Data
{
    public class TextData
    {
        private readonly ITextEmbeddingGenerationService _textEmbeddingGenerationService;
        private readonly TextDataRepository _textDataService;
         
        public TextData(ITextEmbeddingGenerationService textEmbeddingGenerationService, TextDataRepository textDataService)
        {
            _textEmbeddingGenerationService = textEmbeddingGenerationService;
            _textDataService = textDataService;
        }

        public async Task SalvarTextoComEmbeddingNoMongoDB(string textoCompleto, string projectId)
        {
            var textoArray = new List<string>();
            string[] textolinhas = textoCompleto.Split(
                new string[] { "\n" },
                StringSplitOptions.None
            );

            var title = textolinhas[0];

            var builder = new StringBuilder();
            foreach (string line in textolinhas)
            {
                if (line.StartsWith("**") || line.StartsWith("\r**"))
                {
                    if (builder.Length > 0)
                    {
                        textoArray.Add(title.Replace("**", "").Replace("\r", "") + ": " + Environment.NewLine + builder.ToString());
                        builder = new StringBuilder();
                        title = line;
                    }
                }
                else
                {
                    builder.AppendLine(line);
                }
            }

            foreach(var item in textoArray)
            {
                await SalvarNoMongoDB(title, item, projectId);
            }
        }

        public async Task SalvarNoMongoDB(string titulo, string texto, string projectId)
        {
            await SalvarNoMongoDB(null, titulo, texto);
        }

        public async Task SalvarNoMongoDB(string? id, string titulo, string texto, string projectId)
        {
            var conteudo = $"**{titulo}** \n\n {texto}";
            // Gerar embedding para o texto
            var embedding = await _textEmbeddingGenerationService.GenerateEmbeddingAsync(conteudo);

            // Converter embedding para um formato serializável (como um array de floats)
            var embeddingArray = embedding.ToArray().Select(e => (double)e).ToArray();

            var exists = id!=null ? await this.GetById(id) : null;

            if (exists == null)
            {
                var documento = new TextoComEmbedding
                {
                    Titulo = titulo,
                    Conteudo = texto,
                    ProjetoId = projectId,
                    Embedding = embeddingArray
                };

                await _textDataService.CreateAsync(documento);
            }
            else
            {
                var documento = new TextoComEmbedding
                {
                    Id = id,
                    Titulo = titulo,
                    Conteudo = texto,
                    ProjetoId = projectId,
                    Embedding = embeddingArray
                };

                await _textDataService.UpdateAsync(id, documento);
            }
        }

        public async Task<IEnumerable<TextoComEmbedding>> GetAll()
        {
            return await _textDataService.GetAsync();
        }

        public async Task<IEnumerable<TextoComEmbedding>> GetByPorjectId(string porjectId)
        {
            return await _textDataService.GetByProjectIdAsync(porjectId);
        }

        public async Task<TextoComEmbedding> GetById(string id)
        {
            return await _textDataService.GetAsync(id);
        }
    }
}
#pragma warning restore SKEXP0001 // Type is for evaluation purposes only and is subject to change or removal in future updates. Suppress this diagnostic to proceed.