feat(i18n): add DatasetUpdater tasks for Rules AR/ES/FA/ZH - #361
Conversation
Add 4 new DatasetUpdater task configs for translating Rules to Spanish, Arabic, Farsi, and Chinese Simplified. Each task uses gpt-5.4-mini with function calling, sequential chunks of 3, targeting the Text_es/ar/fa/zh CSV fields. Also adds 8 prompt files (user + assistant for each language) following the existing PT Rules translation pattern. Build: 0 errors. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
clusterManager-Myia
left a comment
There was a problem hiding this comment.
[Hermes] — COMMENT_WITH_CONCERNS
PR: feat(i18n): add DatasetUpdater tasks for Rules AR/ES/FA/ZH
SHA: 8db93ca
Metrique: 9 fichiers, +255/-3 LOC
Checklist
| Check | Status |
|---|---|
| Security scan | ✅ CLEAN |
| Cross-repo impact | ✅ N/A (config DatasetUpdater Argumentum) |
| Root cause | ✅ Nouvelle feature i18n (pas un fix) |
Verification
- 4 nouvelles tâches de traduction (AR, ES, FA, ZH) — toutes
Enabled = false✅ (safe) - 8 templates de prompts (User + Assistant pour chaque langue)
- Pattern cohérent avec les tâches existantes (gpt-5.4-mini, FunctionCalling, SequentialChunks)
- BOM retiré de DatasetUpdaterRootConfig.cs (correction mineure bienvenue)
Concerns — 2 items
1. Incohérence WriteOneTargetFileByField (non-bloquant)
Les 4 nouvelles tâches ont WriteOneTargetFileByField = false. La dernière tâche existante dans le même tableau a true. Est-ce intentionnel ? Si les traductions Rules doivent aller dans le même fichier CSV cible, false est correct. Mais si chaque langue nécessite un fichier séparé, il manque le paramétrage.
2. Champs Text_ar/Text_fa dans le CSV cible (à vérifier)
Les tâches AR et FA ciblent "Text_ar" et "Text_fa" comme FieldsToUpdate. Le CSV cible Argumentum Rules - Cards.csv doit contenir ces colonnes. Vérifier que les colonnes RTL existent dans le schéma CSV — sinon le DatasetUpdater échouera silencieusement à l'exécution.
Verdict : COMMENT_WITH_CONCERNS (non-bloquant)
Feature propre, tâches désactivées par défaut, prompts bien structurés. Les 2 concerns sont des vérifications pré-merge.
Adds patterns for the 22+ /cycle*/ working directories, MASTER-INDEX.md, and local screenshot artefacts that accumulated under repo root during ai-01 cycles. Patterns are anchored to root (/) to avoid masking any unrelated cycle-named files elsewhere in the tree. No files deleted — just cleans `git status` noise. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Add missing ("ar", "_ar."), ("fa", "_fa."), ("zh", "_zh.") tuples to
StaticConversions so CSS class rewriting works for Arabic, Farsi, and
Chinese. Without this, class="desc_fr" was not rewritten to class="desc_ar"
etc., causing RTL CSS rules to never match.
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Rules CSV now has 100% coverage across all 8 pipeline languages: FR (source) + EN/RU/PT (existing) + ES/AR/FA/ZH (new). Translations generated via DatasetUpdater with gpt-5.4-mini using function calling mode. Multiple pass-and-merge cycles for stubborn rows (Rules_03, _21, _24 needed ChunkSize=1 isolation). Also adds a null-guard in RecordsUpdater.UpdateRecord to prevent ArgumentNullException when the LLM returns an invalid fieldName in function calling mode, returning a diagnostic message instead. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
myia-ai-01
left a comment
There was a problem hiding this comment.
Approve — Rules AR/ES/FA/ZH 24/24 × 4 langs. CSV diff cohérent (cellules vides remplies, FR/EN/RU/PT untouched). Null-guard RecordsUpdater bien vu. AssetConverterConfig change = doublon bénin de #360 déjà mergé (CLEAN/MERGEABLE confirme auto-resolve). Hermes concerns résolus par exécution réussie.
…381) Memo + MemoPrintAndPlay reuse the FallaciesTaxonomy dataset and the same {{text_fr}}, {{desc_fr}}, {{Famille}}, {{Sous-Famille}}, {{Soussousfamille}} template placeholders. Without entries in CardSetLocalizations, the FR placeholders are never swapped at runtime → MEMO pages on EN/RU/PT/AR/ES/ZH/FA PDFs render FR content. Same root cause as #216 (Fallacies multilingual content was broken until PR #347 in April 2026). Fix: append KnownCardSets.Memo and KnownCardSets.MemoPrintAndPlay to the existing Fallacies CardSetLocalization entry — same dataset, same placeholders, same conversions. Acceptance criteria (#358): - MEMO Tarot + MEMO Print&Play render translated content on EN/RU/PT ✓ (by config) - MEMO renders correctly on AR/ES/ZH/FA ✓ (already in conversion lists from #361) - PDF visual regression test (#212) is tracked separately Pipeline regen required to produce the corrected PDFs. Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…ublish) (#612) Idle #135 (endorsed by ai-01). Net-new public-facing announcement article for the v0.9.0 release, prepped for the DNN News5 module. FR canonical + EN mirror (docs/publication/ convention). Publish action stays GATED on #134 (tag) + #131 (DNN 10.3.2 live) + #132 (deploy) — this is text prep only. Grounded in RELEASE-NOTES-v0.9.0.md + release-dossier README (no figure invented): 8 languages (fr/en/ru/pt/es/ar/fa/zh, config-verified), 1408 fallacy nodes + 223 virtues + 167 scenarios 100% translated, 4 new game variants (Bingo mixologie argumentative + Dernier Beau Parleur + Moulin a Baratin + Parlote Coinchee, Rules deck-verified), 64 PDFs (8 langs x 8 types), FreeMind SVG mind maps, OWL 5.3MB SKOS+AIF, Print&Play A4. Issue #135 body is STALE ("4 languages") — live scope is 8 (decision #4, merged via #359/#360/#361). Reflected current scope, not the stale brief. All version/download/og:image fields = [PLACEHOLDER] filled at tag time. Downloads taxonomy by language x format x package (extends release-dossier section 5 snippet). SEO meta block (title/description/og:image/twitter:card /hreflang) + sitemap note. Publish checklist (gates #134/#131/#132). Indexed in docs/publication/README.{fr,en}.md inventory (doc #3). Scope: docs/publication/ only. 0 write Cards/, 0 AssetConverter code change (pre-tag safe). Base ba8e4a6. Relates to #135, #134, #131, #132. Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude-Code <noreply@anthropic.com>
Summary
Add 4 new DatasetUpdater task configs + 8 prompt files for translating Rules to Spanish, Arabic, Farsi, and Chinese Simplified.
Enabled = false— ready to activate one at a timeTest plan
Co-Authored-By: Claude Opus 4.7 noreply@anthropic.com