Add AUROC quartet

[emilyriederer]

Hello! Per the kind invitation, I'm sharing a PR to add an "AUROC quartet" (scores and targets that produce same AUROC but materially different performance metrics otherwise).

I added the following files:

• data-raw
  • generate-roctet.py: The initial code is python, so this contains a self-contained virtual env with uv with run instructions for full reproducibility. This writes an output csv
  • generate-roctet.R: This converts the csv to rda
  • The initial csv
• data
  • The final rda
• R
  • auroc_quartet.R: Documents the dataset
• Documentation
  • DESCRIPTION: added to list of contained datasets
  • README.qmd: added example with a few viz.
  • man/auroc_quartet.Rd: Rendered dataset docs

A few outstanding questions:

• I tested README.qmd locally to ensure it works, but I did not commit the new README.md since that blew up the repo with re-cached files and figures. Please let me know if you prefer I push those as well.
• Currently, the dataset contains 4K observations (1K per curve) to be more representative of larger scale predictive data. Please let me know if you have any concerns on the size.