Changes

rag-api/Services/Contracts/IAiClient.cs

@@ -0,0 +1,7 @@
+namespace Api.Services.Contracts;
+
+public interface IAiClient
+{
+    Task<float[]> CreateEmbeddingAsync(string input, CancellationToken ct);
+    Task<string> CreateChatCompletionAsync(string systemPrompt, string userPrompt, decimal temperature, CancellationToken ct);
+}

rag-api/Services/Contracts/IDocumentClassifier.cs

@@ -0,0 +1,8 @@
+using Api.Services.Contracts.Models;
+
+namespace Api.Services.Contracts;
+
+public interface IDocumentClassifier
+{
+    Task<DocumentClassification> ClassifyAsync(string text, string? providedType, string? providedTitle, CancellationToken ct);
+}

rag-api/Services/Contracts/IRagRepository.cs

@@ -0,0 +1,16 @@
+using Api.Services.Contracts.Models;
+
+namespace Api.Services.Contracts;
+
+public interface IRagRepository
+{
+    Task InitializeAsync(CancellationToken ct);
+    Task<RagDocumentRecord?> GetDocumentByTextHashAsync(string textHash, string? sourceUrl, CancellationToken ct);
+    Task<RagDocumentRecord?> GetDocumentByIdAsync(string id, CancellationToken ct);
+    Task SaveDocumentAsync(RagDocumentRecord document, IReadOnlyList<RagChunkRecord> chunks, CancellationToken ct);
+    Task<IReadOnlyList<SearchCandidateChunk>> SearchChunksAsync(float[] queryEmbedding, IReadOnlyList<string>? targetTypes, int topK, CancellationToken ct);
+    Task<float[]?> GetEmbeddingAsync(string cacheKey, CancellationToken ct);
+    Task SaveEmbeddingAsync(string cacheKey, string model, string textHash, float[] vector, CancellationToken ct);
+    Task<string?> GetChatCompletionAsync(string cacheKey, CancellationToken ct);
+    Task SaveChatCompletionAsync(string cacheKey, string model, decimal temperature, string responseText, CancellationToken ct);
+}

rag-api/Services/Contracts/IRagService.cs

@@ -0,0 +1,13 @@
+using Api.Requests;
+using Api.Responses;
+using Api.Services.Contracts.Models;
+
+namespace Api.Services.Contracts;
+
+public interface IRagService
+{
+    Task<IndexDocumentResponse> IndexTextAsync(IndexDocumentRequest request, CancellationToken ct);
+    Task<IndexDocumentResponse> IndexPdfAsync(IFormFile file, string? documentType, string? title, string? sourceUrl, CancellationToken ct);
+    Task<SearchResponse> SearchAsync(SearchRequest request, CancellationToken ct);
+    Task<RagDocumentDetails?> GetDocumentAsync(string documentId, CancellationToken ct);
+}

rag-api/Services/Contracts/ITextChunker.cs

@@ -0,0 +1,6 @@
+namespace Api.Services.Contracts;
+
+public interface ITextChunker
+{
+    IReadOnlyList<string> Chunk(string text, int chunkSize, int overlap);
+}

rag-api/Services/Contracts/ITextExtractor.cs

@@ -0,0 +1,7 @@
+namespace Api.Services.Contracts;
+
+public interface ITextExtractor
+{
+    Task<string> ExtractPdfAsync(Stream stream, CancellationToken ct);
+    string Normalize(string value);
+}

rag-api/Services/Contracts/Models/DocumentClassification.cs

@@ -0,0 +1,10 @@
+namespace Api.Services.Contracts.Models
+{
+    public sealed class DocumentClassification
+    {
+        public required string DocumentType { get; init; }
+        public double Confidence { get; init; }
+        public required string Title { get; init; }
+        public Dictionary<string, string> Metadata { get; init; } = [];
+    }
+}

rag-api/Services/Contracts/Models/RagChunkRecord.cs

@@ -0,0 +1,11 @@
+namespace Api.Services.Contracts.Models
+{
+    public sealed class RagChunkRecord
+    {
+        public required string Id { get; init; }
+        public required string DocumentId { get; init; }
+        public int ChunkIndex { get; init; }
+        public required string Text { get; init; }
+        public required float[] Embedding { get; init; }
+    }
+}

rag-api/Services/Contracts/Models/RagDocumentDetails.cs

@@ -0,0 +1,13 @@
+namespace Api.Services.Contracts.Models
+{
+    public sealed class RagDocumentDetails
+    {
+        public required string Id { get; init; }
+        public required string DocumentType { get; init; }
+        public required string Title { get; init; }
+        public string? SourceUrl { get; init; }
+        public required string Text { get; init; }
+        public required string TextHash { get; init; }
+        public DateTimeOffset CreatedAt { get; init; }
+    }
+}

rag-api/Services/Contracts/Models/RagDocumentRecord.cs

@@ -0,0 +1,15 @@
+namespace Api.Services.Contracts.Models
+{
+    public sealed class RagDocumentRecord
+    {
+        public required string Id { get; init; }
+        public required string DocumentType { get; init; }
+        public required string Title { get; init; }
+        public string? SourceUrl { get; init; }
+        public required string Text { get; init; }
+        public required string TextHash { get; init; }
+        public double TypeConfidence { get; init; }
+        public string MetadataJson { get; init; } = "{}";
+        public DateTimeOffset CreatedAt { get; init; }
+    }
+}

rag-api/Services/Contracts/Models/SearchCandidateChunk.cs

@@ -0,0 +1,9 @@
+namespace Api.Services.Contracts.Models
+{
+    public sealed class SearchCandidateChunk
+    {
+        public required RagDocumentRecord Document { get; init; }
+        public required RagChunkRecord Chunk { get; init; }
+        public double Score { get; init; }
+    }
+}