ob-labs/ChatBot

Implements a multi-modal RAG chatbot using OceanBase's native vector capabilities (HNSW indexing for ANN search) combined with Tongyi Qianwen's embedding and LLM APIs, processing OceanBase documentation through a retrieval pipeline that converts queries to vectors, performs semantic similarity search, and streams LLM-generated responses via Streamlit. The architecture unifies vector and structured data storage in a single relational database, leveraging both semantic understanding and traditional SQL queries to ground answers in retrieved document context.