[dxnnystyles]

Summary

checks existing benchmark output before writing. subset runs that would remove other audits, or replace incomplete or inconsistent output, now stop with a clear error. compatible reruns remain supported.

Type

• Audit
• Explainer
• Bug fix
• Other

---

How I Tested

• python -m pytest tests/test_results_provenance.py -q (15 passed).
• python -m pytest tests/ -q -m "not slow" (1,244 passed, 1 skipped, 23 deselected).
• ruff check faircode/benchmark.py faircode/cli.py tests/test_results_provenance.py tests/test_benchmark.py and git diff --check passed.
• make lint stops at existing closed issue references in ROADMAP.md (JSON string cells "NA" and "" are missing in the web engine and ordinary values in Python #1036, XLSX date cells are labelled 2020-01-01 00:00:00 by Python and by their serial number (43831) by the web engine #1039, Numeric column headers keep their type in Python (the integer 2020) and are text in the web engine #1045, An empty header cell is named Unnamed: 1 by pandas and an empty string by the web engine #1046, Two long group labels that share their first 29 characters still print identically after the ellipsis fix #1038).
• ruff format --check reports these files would be reformatted; the upstream versions also fail that check.

Linked issue

Closes #1151

[vercel]

@dxnnystyles is attempting to deploy a commit to the yashkewlani2020-gmailcom's projects Team on Vercel.

A member of the Team first needs to authorize it.

[github-actions]

@yakew7 @ahmdkaml - new PR touching a path you own, please review.