A thin wrapper for llama-cpp-python to be called from a comfyui node

LangChain integration with Ollama LLMs to build a simple Retrieval-Augmented Generation (RAG) pipeline: embeds text via Hugging Face, stores in FAISS, retrieves relevant context, and invokes Ollama (llava or bakllava) for both text and image‐based queries.