MCP AI POC

# Architecture Overview ## Project Structure This is a Python proof-of-concept application that demonstrates integration with OpenAI's API using their modern Python SDK (>=1.0.0). ### Core Components ``` src/mcp_poc/ ├── app.py # Main application logic └── __init__.py # Package initialization src/tests/ └── test_app.py # Unit tests for the application ``` ## Key Classes and Functions ### `OpenAIClient` - **Purpose**: Manages OpenAI API client instantiation and configuration - **Pattern**: Singleton-like pattern with lazy initialization - **Key Method**: `get_client()` - Returns configured OpenAI client instance ### `generate_completion(messages, model)` - **Purpose**: Generate text completions using OpenAI's chat API - **Parameters**: - `messages`: List of conversation messages - `model`: OpenAI model to use (default: "gpt-4o") - **Returns**: Generated text response ### `select_model()` - **Purpose**: Interactive model selection from available OpenAI models - **Behavior**: Fetches and displays available GPT models, allows user selection - **Fallback**: Defaults to "gpt-4o" if selection fails ### `chat_loop()` - **Purpose**: Main interactive chat interface - **Features**: - Model selection at startup - Conversation history management - Graceful error handling - Exit commands ("quit", "exit") ## Dependencies - **openai**: Official OpenAI Python SDK for API interactions - **pytest**: Testing framework (dev dependency) ## Configuration - Requires `OPENAI_API_KEY` environment variable - No additional configuration files needed