cd server/db
python 001_add_session_tables.py
python 002_add_interaction_table.py
- Download these 6 data files from https://uofi.app.box.com/s/m8lwwym09tay04669idbscphrsmt6w1e
- all_papers_order.json
- compressed_array.npz
- docs_per_user.json
- nn_model.pkl
- normalized_user_ratings.json
- quick_lookup.csv
- Download arxiv-metadata-oai-snapshot.json from https://www.kaggle.com/datasets/Cornell-University/arxiv
- First, download archive.zip, then extract to get the json file.
- Place all the previous downloaded files in db folder
python create_fresh_db.py
# See Note1
cd ..
# go to server folder- Install python if pip doesn't work
pip install fastapi faiss-cpu sentence_transformers uvicorn
python server.py
# Running the backend server
cd ../ui
- Install nodejs if npm doesn't work
npm install
npm start
# Running the web server
After set up: Run UI, run server. Interact via webpage at http://localhost:3000
run: create_fresh_db.py If you want to make the full DB (2M papers), set EARLY_STOP = -1 Else: set it to whatever number of papers you want to process. If you had made the DB earlier, to update it with the new changes, simply run 001_add_session_tables.py to add the new tables. Otherwise, if you do not have a previous DB, create_fresh_db.py should be all you need to set up.