SNU RAG Chatbot

RAG based chatbot to deal with school rules related questions of Seoul National University.

Done as a part of Creative Integrated Design2 Class.

UI and Demo

How to use

1. Install dependencies

pip install -r requirements.txt

2. Install Milvus standalone based on your environment. Details are given here.

3. Prepare the database

cd snu_milvus

python init_db.py -n “snu_milvus_final”
python insert_embedding.py -n “snu_milvus_final” -I “output_final.csv” -b “4” -d “cuda”

4. Move and run llm_inference.ipynb in your own colab environment

5. Run the backend

   • You must set replace COLAB_URI of rag_backend/rag/views.py based on the ngrok url that you received from the llm_inference.ipynb

cd rag_backend

python manage.py runserver

6. Run the frontend

cd rag_frontend

npm install
npm start