ChatRAG/Data/TextData.cs

using ChatRAG.Data;
using ChatRAG.Models;
using ChatRAG.Services.Contracts;
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 : ITextDataService
    {
        private readonly ITextEmbeddingGenerationService _textEmbeddingGenerationService;
        private readonly TextDataRepository _textDataService;

        public TextData(ITextEmbeddingGenerationService textEmbeddingGenerationService, TextDataRepository textDataService)
        {
            _textEmbeddingGenerationService = textEmbeddingGenerationService;
            _textDataService = textDataService;
        }

        public string ProviderName => "MongoDB";

        // ========================================
        // MÉTODOS ORIGINAIS (já implementados)
        // ========================================

        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, projectId);
        }

        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
                {
                    Id = id ?? Guid.NewGuid().ToString(),
                    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))!;
        }

        // ========================================
        // MÉTODOS NOVOS DA INTERFACE (implementação completa)
        // ========================================

        public async Task<string> SaveDocumentAsync(DocumentInput document)
        {
            var id = document.Id ?? Guid.NewGuid().ToString();
            await SalvarNoMongoDB(id, document.Title, document.Content, document.ProjectId);
            return id;
        }

        public async Task UpdateDocumentAsync(string id, DocumentInput document)
        {
            await SalvarNoMongoDB(id, document.Title, document.Content, document.ProjectId);
        }

        public async Task DeleteDocumentAsync(string id)
        {
            await _textDataService.RemoveAsync(id);
        }

        public async Task<bool> DocumentExistsAsync(string id)
        {
            try
            {
                var doc = await GetById(id);
                return doc != null;
            }
            catch
            {
                return false;
            }
        }

        public async Task<DocumentOutput?> GetDocumentAsync(string id)
        {
            try
            {
                var doc = await GetById(id);
                if (doc == null) return null;

                return new DocumentOutput
                {
                    Id = doc.Id,
                    Title = doc.Titulo,
                    Content = doc.Conteudo,
                    ProjectId = doc.ProjetoId,
                    Embedding = doc.Embedding,
                    CreatedAt = DateTime.UtcNow,
                    UpdatedAt = DateTime.UtcNow,
                    Metadata = new Dictionary<string, object>
                    {
                        ["source"] = "MongoDB",
                        ["has_embedding"] = doc.Embedding != null,
                        ["embedding_size"] = doc.Embedding?.Length ?? 0
                    }
                };
            }
            catch
            {
                return null;
            }
        }

        public async Task<List<DocumentOutput>> GetDocumentsByProjectAsync(string projectId)
        {
            var docs = await GetByPorjectId(projectId);
            return docs.Select(doc => new DocumentOutput
            {
                Id = doc.Id,
                Title = doc.Titulo,
                Content = doc.Conteudo,
                ProjectId = doc.ProjetoId,
                Embedding = doc.Embedding,
                CreatedAt = DateTime.UtcNow,
                UpdatedAt = DateTime.UtcNow,
                Metadata = new Dictionary<string, object>
                {
                    ["source"] = "MongoDB",
                    ["has_embedding"] = doc.Embedding != null,
                    ["embedding_size"] = doc.Embedding?.Length ?? 0
                }
            }).ToList();
        }

        public async Task<int> GetDocumentCountAsync(string? projectId = null)
        {
            if (string.IsNullOrEmpty(projectId))
            {
                var all = await GetAll();
                return all.Count();
            }
            else
            {
                var byProject = await GetByPorjectId(projectId);
                return byProject.Count();
            }
        }

        public async Task<List<string>> SaveDocumentsBatchAsync(List<DocumentInput> documents)
        {
            var ids = new List<string>();

            foreach (var doc in documents)
            {
                var id = await SaveDocumentAsync(doc);
                ids.Add(id);
            }

            return ids;
        }

        public async Task DeleteDocumentsBatchAsync(List<string> ids)
        {
            foreach (var id in ids)
            {
                await DeleteDocumentAsync(id);
            }
        }

        public async Task<Dictionary<string, object>> GetProviderStatsAsync()
        {
            try
            {
                var totalDocs = await GetDocumentCountAsync();
                var allDocs = await GetAll();

                var docsWithEmbedding = allDocs.Count(d => d.Embedding != null && d.Embedding.Length > 0);
                var avgContentLength = allDocs.Any() ? allDocs.Average(d => d.Conteudo?.Length ?? 0) : 0;

                var projectStats = allDocs
                    .GroupBy(d => d.ProjetoId)
                    .ToDictionary(
                        g => g.Key ?? "unknown",
                        g => g.Count()
                    );

                return new Dictionary<string, object>
                {
                    ["provider"] = "MongoDB",
                    ["total_documents"] = totalDocs,
                    ["documents_with_embedding"] = docsWithEmbedding,
                    ["embedding_coverage"] = totalDocs > 0 ? (double)docsWithEmbedding / totalDocs : 0,
                    ["average_content_length"] = Math.Round(avgContentLength, 1),
                    ["projects_count"] = projectStats.Count,
                    ["documents_by_project"] = projectStats,
                    ["health"] = "ok",
                    ["last_check"] = DateTime.UtcNow,
                    ["connection_status"] = "connected"
                };
            }
            catch (Exception ex)
            {
                return new Dictionary<string, object>
                {
                    ["provider"] = "MongoDB",
                    ["health"] = "error",
                    ["error"] = ex.Message,
                    ["last_check"] = DateTime.UtcNow,
                    ["connection_status"] = "error"
                };
            }
        }
    }
}

#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.