Skip to content

audit: add D1.3 non-destructive data-quality reports#29

Merged
sunpole merged 25 commits into
mainfrom
audit/d1.3-data-quality-v23.8.10
Jul 23, 2026
Merged

audit: add D1.3 non-destructive data-quality reports#29
sunpole merged 25 commits into
mainfrom
audit/d1.3-data-quality-v23.8.10

Conversation

@sunpole

@sunpole sunpole commented Jul 23, 2026

Copy link
Copy Markdown
Owner

Цель

Завершить D1.3: выполнить воспроизводимый неразрушающий аудит качества зарегистрированных source/current datasets без исправления или перезаписи данных.

Реализовано

  • scripts/audit-data-quality.mjs — permanent read-only audit с generation и --check;
  • docs/DATA_QUALITY_AUDIT.md — severity model, structural/hygiene/ambiguity rules, determinism и limitations;
  • reports/data-quality-audit.json — deterministic machine-readable report;
  • reports/data-quality-audit.md — deterministic human-readable report;
  • permanent GitHub Actions check report freshness and structural gate;
  • synchronized maintained documentation;
  • factual uNews patchnote и новый real Chromium document-render exact report page.

Фактические результаты

  • logical datasets: source-divinity-code-en, ru-current-v1;
  • records: 4 086 в каждом;
  • IDs: unique, ordered and aligned 1–4086;
  • preserved-field differences: 0 for id, symbol, description, source, date_added;
  • changed-field counts: aliases 4 083, notes 4 086, tags 4 086;
  • structural gate: PASS;
  • errors: 0;
  • warnings: 0;
  • human-review instances: 5 022 в пяти aggregated groups.

5 022 — не доказанные errors. Findings пересекаются; shared aliases могут быть intentional, а empty source notes могут отражать source structure. Audit ничего не исправляет.

Aggregated review groups

  • current alias collisions: 854;
  • current alias-to-primary matches: 1 145;
  • source alias collisions: 693;
  • source alias-to-primary matches: 1 145;
  • empty source notes: 1 185.

Защищённые области

Changed-file list не содержит:

  • data/bd2.json;
  • data/db.json;
  • data/divinity_code_ru.json;
  • data/report.txt;
  • runtime/PWA/package metadata;
  • versions/;
  • _archive/.

uNews

Patchnote: news/2026-07-23-udream-v23-8-10-data-quality-audit.md

Image: news/2026-07-23-udream-v23-8-10-data-quality-audit.png

Image source commit: cd6aa539a418861108d58f2206050291642e7fcb

Capture UTC: 2026-07-23T07:09:49Z

Final branch state

Final handoff commit before review: 22fe94598e8de3a4c72ac76ffb9265625249c3eb.

После merge

  • закрыть Issue №28;
  • вернуть main/WORK_STATUS.md в READY;
  • следующая задача — D1.4 two-book product architecture.

@sunpole
sunpole marked this pull request as ready for review July 23, 2026 07:16
@sunpole
sunpole merged commit d4c100c into main Jul 23, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant