The pipeline for generating and testing models can be found in main.ipynb. This file generates the tables and figures seen in the paper which can also be found stored in results.tar.gz.

Requirements

• Pipenv
• Python 3.7.3

Running

1. pipenv install && pipenv shell
2. cd paper_download && python pull_papers.py
3. python -m ipykernel install --user --name=card_shark_3_paper
4. jupyter notebook
5. Open main.ipynb, select the kernel > change kernel dropdown menu and select the card_shark_3_paper environment. You can now run through each of the cells to obtain the same results as those stored in results.tar.gz.

card_pubs.csv - Jan 12, 2021

SELECT *
FROM pub
JOIN pub_dbxref as dbx ON pub.pub_id = dbx.pub_id
JOIN dbxref as ref ON ref.dbxref_id = dbx.dbxref_id
ORDER BY pub.pub_id DESC

A total of 131 papers found in card_pubs.csv were found in the papers obtained between 2017-2020 Thus the retrospective venn-diagram and table are relative to 131 papers.