[Oracle AI Developer Hub] Add notebook: Agentic RAG with LangGraph and hybrid search

## Description Add a new notebook demonstrating agentic RAG using Oracle Database 26AI as the sole data backend - no separate vector database or search engine required. ## What the notebook covers - Oracle 26AI Free running locally via Docker (no cloud account needed) - Document ingestion with langchain-oracledb OracleVS and HuggingFace local embeddings - Three retrieval modes: - Vector similarity search using HNSW indexes and VECTOR_DISTANCE() - Keyword search using Oracle Text - Hybrid search combining both in a single SQL CTE - LangGraph create_react_agent wired to the vector store as a tool - Full tool call visibility showing the agent's reasoning chain ## Why this is valuable - First notebook in the repo using LangGraph create_react_agent with Oracle - Hybrid search via a single SQL CTE is a cleaner pattern than multi-step approaches - Agent tool call visibility is useful for debugging and demos - Fully self-contained: all sample data inline, no external files needed #