How hung-3008/kgchat-03 works
Project type and purpose
• This is a backend library/service module for managing and interacting with multiple Large Language Model (LLM) providers.
• It is organized as a plugin‐style system where each provider (OpenAI, Gemini, Ollama, etc.) implements a common interface and can be registered dynamically.
File structure
• Top‐level directory “backend/llm”:
• base/ – Abstract classes and interfaces (LLMClient, Config, Factory)
• factory/ – A registry and factory that selects and instantiates the correct client
• providers/ – One subfolder per provider, each with its own Client and Config classes
• utils/ – Shared utilities (exceptions, rate limiter, retry handler)
• No frontend or database components; the core functionality is API calls to external LLM services.
README for additional insights
• (Empty in this case, so rely entirely on code structure.)