Semantic Kernel POC

A proof-of-concept chat bot API powered by Microsoft Semantic Kernel with persistent chat history and document memory capabilities.

Features

• Semantic Kernel Integration: Utilizes Microsoft Semantic Kernel for AI-powered conversations
• Persistent Chat History: SQLite database storage for chat sessions and messages
• Document Memory: Kernel Memory integration for document context and retrieval
• RESTful API: Clean API endpoints for chat operations
• Swagger Documentation: Built-in API documentation and testing interface

Technology Stack

• .NET 9.0: Latest .NET framework
• ASP.NET Core: Web API framework
• Entity Framework Core: Database ORM with SQLite
• Microsoft Semantic Kernel: AI orchestration framework
• Kernel Memory: Document processing and retrieval
• Swagger/OpenAPI: API documentation

API Endpoints

Chat History

• GET /api/chathistory - Retrieve chat history
• POST /api/chathistory - Create new chat messages

Documents

• GET /api/documents - List available documents
• POST /api/documents - Upload and process documents

Memory

• GET /api/memory - Query memory store
• POST /api/memory - Add content to memory

Configuration

• GET /api/config - Get system configuration

Getting Started

Prerequisites

• .NET 9.0 SDK
• OpenAI API key (for Semantic Kernel)

Setup

1. Clone the repository

   git clone <repository-url>
   cd Semantic-Kernel-POC

2. Configure settings Create appsettings.Development.json:

   {
     "OpenAI": {
       "ApiKey": "your-openai-api-key",
       "ModelId": "gpt-4"
     },
     "SemanticKernel": {
       "ServiceType": "OpenAI"
     },
     "Database": {
       "ConnectionString": "Data Source=chatbot.db"
     }
   }

3. Run the application

   dotnet run

4. Access Swagger UI Navigate to https://localhost:5001/swagger to explore the API

Project Structure

├── Controllers/           # API controllers
│   ├── ChatHistoryController.cs
│   ├── ConfigController.cs
│   ├── DocumentsController.cs
│   └── MemoryController.cs
├── Data/                  # Database context
├── Models/                # Data models and configurations
├── Services/              # Business logic services
├── Documents/             # Sample documents
└── Program.cs             # Application entry point

Configuration Options

OpenAI Configuration

• ApiKey: Your OpenAI API key
• ModelId: AI model to use (e.g., "gpt-4", "gpt-3.5-turbo")

Database Configuration

• ConnectionString: SQLite database connection string

Semantic Kernel Configuration

• ServiceType: AI service provider ("OpenAI")

Development Workflow

This project follows strict development workflows defined in CLAUDE.md:

• Design-first approach avoiding over-engineering
• Task-driven development with TODO tracking
• One task at a time execution
• Comprehensive testing and cleanup procedures

Sample Documents

The Documents/ folder contains sample text files that can be used to test document processing and memory features.

Contributing

1. Follow the workflow rules in CLAUDE.md
2. Create feature branches using format: feature/task{number}-{description}
3. Make meaningful commits for each step
4. Ensure all tests pass before submitting PRs

License

This is a proof-of-concept project for educational and demonstration purposes.