TCM.InAppAuth 1.0.7

TCM.InAppAuth

A shared library for AI integration using LM Studio in the TCM application suite.

Features

• LM Studio Integration: Connect to LM Studio server for AI chat completions
• Prompt Management: Store and retrieve AI prompts from the InAppAuth database
• Easy Integration: Simple extension methods for service registration

Installation

Add a project reference to TCM.InAppAuth in your project:

<ProjectReference Include="..\..\..\shareds\TCM.InAppAuth\TCM.InAppAuth.csproj" />

Configuration

appsettings.json

Add the following configuration to your appsettings.json:

{
  "ConnectionStrings": {
    "InAppAuthConnection": "Server=your-server;Database=InAppAuth;User Id=your-user;Password=your-password;TrustServerCertificate=True"
  },
  "LMStudio": {
    "BaseUrl": "http://localhost:1234",
    "DefaultModel": "your-default-model-id",
    "DefaultMaxTokens": 2048,
    "DefaultTemperature": 0.7,
    "TimeoutSeconds": 120,
    "ApiKey": ""
  }
}

Service Registration

In your Program.cs or Startup.cs:

using TCM.InAppAuth.Extensions;

// Option 1: Add all TCM.InAppAuth services including database
builder.Services.AddTCMAIWithDatabase(builder.Configuration);

// Option 2: Add only LM Studio services (without database)
builder.Services.AddTCMAI(builder.Configuration);

// Option 3: Add only database context (without LM Studio services)
builder.Services.AddInAppAuthDbContext(builder.Configuration);

Usage

Basic Chat Completion

using TCM.InAppAuth.Services;
using TCM.InAppAuth.Models;

public class MyService
{
    private readonly ILMStudioService _lmStudioService;

    public MyService(ILMStudioService lmStudioService)
    {
        _lmStudioService = lmStudioService;
    }

    public async Task<string> GetAIResponse(string userMessage)
    {
        var result = await _lmStudioService.SendMessageAsync(
            message: userMessage,
            systemMessage: "You are a helpful assistant."
        );

        if (result.Success)
        {
            return result.Content;
        }
        
        throw new Exception(result.ErrorMessage);
    }
}

Using Stored Prompts

// First, add a prompt to the InAppAuth database
// Then use it by promptCode:

var result = await _lmStudioService.SendPromptAsync(
    promptCode: "QUALITY_CHECK",
    userInput: "Check this product description for quality issues."
);

Advanced Chat Completion

var request = new ChatCompletionRequest
{
    Model = "your-model-id",
    Messages = new List<ChatMessage>
    {
        ChatMessage.System("You are an expert in manufacturing."),
        ChatMessage.User("What are best practices for quality control?")
    },
    MaxTokens = 1000,
    Temperature = 0.5m
};

var result = await _lmStudioService.ChatCompletionAsync(request);

Managing Prompts

using TCM.InAppAuth.Repositories;
using TCM.InAppAuth.Entities;

public class PromptService
{
    private readonly IAIPromptRepository _promptRepository;

    public PromptService(IAIPromptRepository promptRepository)
    {
        _promptRepository = promptRepository;
    }

    public async Task CreatePrompt()
    {
        var prompt = new AIPrompt
        {
            PromptCode = "PRODUCTION_ANALYSIS",
            Prompt = "Analyze the following production data and provide insights:",
            ModelId = "your-model-id",
            Description = "Analyze production data",
            SystemMessage = "You are a production analyst expert.",
            MaxTokens = 2000,
            Temperature = 0.7m
        };

        await _promptRepository.AddAsync(prompt);
    }
}

Database Schema

The InAppAuth database contains the following table:

AIPrompt

Project Structure

TCM.InAppAuth/
├── Contexts/
│   └── InAppAuthDbContext.cs          # Database context for InAppAuth
├── Entities/
│   └── AIPrompt.cs               # Prompt entity
├── EntityConfigurations/
│   └── AIPromptConfiguration.cs  # EF Core configuration
├── Extensions/
│   └── ServiceCollectionExtensions.cs  # DI extensions
├── Models/
│   ├── AICompletionResult.cs     # Completion result model
│   ├── ChatCompletionRequest.cs  # API request model
│   ├── ChatCompletionResponse.cs # API response model
│   └── ChatMessage.cs            # Chat message model
├── Repositories/
│   ├── AIPromptRepository.cs     # Repository implementation
│   └── IAIPromptRepository.cs    # Repository interface
├── Services/
│   ├── ILMStudioService.cs       # Service interface
│   └── LMStudioService.cs        # Service implementation
├── Settings/
│   └── LMStudioSettings.cs       # Configuration settings
└── TCM.InAppAuth.csproj           # Project file

Environment Variables

For non-development environments, the following environment variables are used:

• InAppAuthConnection: Connection string for the InAppAuth database
• ASPNETCORE_ENVIRONMENT: Environment name (Development, Staging, Production)