emarco177/langgraph-course — reverse-engineered prompt

I want a hands on Python repo for learning how to build AI agents with LangGraph. Make it feel like a course companion where I can run a few separate examples, ideally as branches or clearly separated projects, and each one should actually work end to end.

Please include an advanced RAG example that can load data, retrieve context, grade whether the docs are relevant, optionally do web search, and route adaptively before generating an answer. I also want a classic ReAct style agent, plus reflection and reflexion style agents that can critique and improve their own output. The flows should show multi step graph logic, conditionals, and passing state between nodes.

Keep setup simple with a main entry point, environment variables for OpenAI and optional web search and tracing, and clear instructions so I can clone it, install dependencies, and run each example. If it makes sense, structure it so the history can be followed lesson by lesson. Look up current LangGraph docs online if you need to.