Skip to content

Eight explainers are still listed as known exceptions in the structure check #1130

Open

Description

@yakew7

Where: scripts/check_explainer_structure.py (KNOWN_EXCEPTIONS), explainers/neural-networks.md, sampling-bias.md, differential-privacy.md, disparate-impact.md, fairness-metric-conflicts.md, shap-values.md, how-ai-detects-patterns.md, predictive-parity.md.

Current gap: the structure check lets each of these eight pages lack one thing CONTRIBUTING.md asks of every explainer: a Detection Code section (neural-networks, disparate-impact, fairness-metric-conflicts, shap-values), a Related Concepts section (sampling-bias, which has "Relationship to Other Concepts"), a Concrete Example (differential-privacy), or the standard section order (how-ai-detects-patterns, predictive-parity). The check fails if an exception becomes stale, so each fix also removes its entry.

Repro: sed -n '/KNOWN_EXCEPTIONS/,/^}/p' scripts/check_explainer_structure.py

Suggested approach: fix the pages one by one: add a short runnable detection block with real output to the four that lack it, rename or add the heading where it is only named differently, add a concrete example from one of the audits to differential-privacy, and reorder the two with the wrong order. Use only numbers the code prints, and rebuild.

Done when: KNOWN_EXCEPTIONS is empty (or deleted), make lint passes, and the generated pages are rebuilt.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    documentationImprovements or additions to documentationgood first issueGood for newcomers

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions