These are two prototypes to explore the HBS Faculty Directory using retrieval augmented generation (RAG).
The backend under app.py is a llamaindex chat engine that queries a vector store in chroma_db
The main guidance-refinement prototype is user-query-form/ which uses a form input to build the RAG query, and then LLM-driven suggested questions for follow up responses.
chat-app/ contains a chat app that takes the input and uses that for queries to the llamaindex chat engine, and directly displays the responses.
Install git-lfs, which is used to store the vector DB. Then clone this repo.
Make a copy of .env-example to .env and add your OpenAI key.
Start a virtual environment with venv, running python -m venv venv and then start it with source venv/bin/activate.
install dependencies pip install -r requirements.txt
In each of the prototype directories user-query-form and chat-app, install the respective app with npm install.
Navigate to the frontend app in user-query-form/ and run npm run build. Start the server with python3 app.py
If instead you are running the chat app, navigate to chat-app/ and npm run build, then start the server with python3 chat_server.py.