fix: check_generated_files_current.py now rebuilds from source, not f…

[propcgamer20-png]

…rom HEAD

It only ever diffed the working tree against 'git show HEAD:' - so editing a source .md without running 'make build-explainers' afterward still reported "up to date", as long as nothing was committed yet. Confirmed exactly the issue's repro (git worktree add + append a line to explainers/demographic-parity.md) now correctly fails with 'content differs from a fresh regeneration' where it used to pass.

_fresh_build_dir() now does the real regeneration: a git worktree checked out to HEAD (shares this repo's object database/history, so git-log-derived dates resolve identically to a build done in ROOT itself), with every currently-tracked file's working-tree content copied on top before running build_explainers.py, generate_og_images.py, and freeze_paper_results.mirror_for_mcp() inside it - the same sequence build-explainers.yml runs, now run by this script itself instead of assumed to have already happened. The worktree is cleaned up via ExitStack regardless of what happens after.

Updated tests/test_check_generated_files_current.py to monkeypatch _fresh_build_dir directly (pre-populated stub directory) instead of faking 'git show', per the file's own stated testing philosophy of avoiding a real git repo in unit tests. Added a regression test for the exact scenario in the bug report.

Verification: reproduced the issue's exact repro against the unmodified script (confirmed it passed when it shouldn't), then against the fixed script (confirmed it now fails with the right reason). All 9 tests/test_check_generated_files_current.py tests pass; full suite is 402 passed, 1 skipped, 4 failed - the 4 failures are tests/test_xlsx_edge_cases.py, confirmed identical on unmodified main (SheetJS CDN returns HTTP 403 in this sandbox), unrelated to this change.

Fixes #502

Summary

Type

• Audit
• Explainer
• Bug fix
• Other

---

Audit checklist

• I opened or linked a corresponding issue first
• The folder is named after the domain, not the dataset
• unfair.py includes protected attributes and prints the required output format
• fair.py removes protected attributes and identified proxy variables
• Both scripts use random_state=42 and an 80/20 train/test split
• Proxy variables were actually tested, not just guessed
• unfair.png and fair.png are included as PNG screenshots
• The dataset is public and accessible without login or payment
• The dataset file is included, or DATA.md is included if the file is too large
• README.md includes the new results row and audit section
• A notebook was added if the audit benefits from one

Before fairness gap:

After fairness gap:

Reduction:

Protected attribute(s):

Proxy variables dropped:

---

Explainer checklist

• The file is in explainers/ and uses lowercase hyphenated naming
• It includes a plain-language definition
• It uses a real example from this repo or a documented real-world case
• It includes runnable Python detection or measurement code
• It acknowledges limitations or trade-offs
• It links to related explainers or repo projects
• It includes 2-3 primary sources
• The Explainers table in README.md was updated

---

Linked issue

Closes #

[vercel]

@propcgamer20-png 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.