Skip to content

chore(catalog): scheduled auto-regenerate (long-lived PR) - #18035

Merged
myia-ai-01 merged 1 commit into
mainfrom
chore/catalog-refresh-pending
Sep 27, 2026
Merged

myia-ai-01 merged 1 commit into
mainfrom
chore/catalog-refresh-pending

Conversation

@github-actions

Copy link
Copy Markdown
Contributor

Purpose

Long-lived PR refreshed daily by catalog-cron.yml (schedule cron 03:37 UTC) to keep
COURSE_CATALOG.generated.{json,md} + CATALOG-STATUS markers + the curriculum (docs/curriculum/*.md)
and health dashboard (docs/archive/reference/HEALTH_DASHBOARD.md) canonical on main.

Maintainer review only — no direct push to main (issue #10136). The bot opens
or updates this PR; merging it refreshes main through the normal PR gate.

What to review

  • Catalog totals (notebook_count, per-family counts) drift only if new notebooks landed on main since the last merge.
  • README <!-- CATALOG-STATUS:START -->…:END --> markers reflect the freshly-regenerated catalog.
  • Curriculum pages under docs/curriculum/ and docs/archive/reference/HEALTH_DASHBOARD.md stay consistent.

Close policy

When the catalog stops drifting (the workflow runs ::notice title=Catalog::main catalog already canonical -- no change.),
this PR will be auto-closed by catalog-pr-close.yml to keep the queue clean. Subsequent
drift reopens it under the same branch.

See .claude/rules/catalog-pr-hygiene.md and #2433, #10136.

Grain: MED/refactor -- lane myia-ai-01:CoursIA

Refresh COURSE_CATALOG.generated.{json,md} + CATALOG-STATUS markers.

Catalog is owned by automation; agents must not regenerate it on feature branches (see .claude/rules/catalog-pr-hygiene.md, #2433, #10136).
@github-actions github-actions Bot added automation catalog-refresh markdown-table-syntax Table syntax defect in changed files (CODE_SPAN_PIPE, NO_SEP, ...). Advisory. See #10097. paragraph-length Paragraph > 2000 chars (wall-of-text, #15405). Resorb before merge. labels Sep 27, 2026

@clusterManager-Myia clusterManager-Myia left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

[NanoClaw] structural review (artefact généré : recount mécanique base↔head, pas de lecture du diff brut)

VERDICT: LGTM contenu (vérifié : recount mécanique du catalogue) — en COMMENT : garde APPROVE non satisfaite ce tour (gh pr checks exit 1, jambe DWELL = minuteur 120 min, levée ~12:07Z).

Vérifications firsthand :

  1. Recount JSON base↔head (2 352 952 → 2 363 821 o) : 1306 → 1312 entrées (+6/−0, zéro retrait). Les 6 ajouts existent tous sur main (contents API, 49-141 Ko chacun) : Complexity-03c Zoo-Navigation, PT_17 laya_proper_rewards_toy, ML 2.11f Comparaison-Optimisation-Convexe, Langlands 01-formes-modulaires-sl2z-hecke, Lean-24b Confiance-Preuves-Native, Tweety-02f Modal-Zoo-Lean-Python. Deltas par série cohérents : Complexity 8→9, GenAI 246→247, ML 119→120, SymbolicAI 304→307.
  2. 61 entrées existantes modifiées sur des champs de télémétrie d'exécution uniquement (last_success_sha ×61, executed_at ×61, resource_cost ×61, issue_pr_associee ×61, last_validation ×57, cells_total/cells_markdown ×25, scientific_review_detail ×15) — aucune mutation sémantique de masse, titres et statuts stables.
  3. Cohérence JSON↔MD : Total notebooks: 1312 dans le .md régénéré = compte JSON exact ; les marqueurs CATALOG-STATUS des 8 README touchés portent les +3/−3 attendus.
  4. Checks au head 06effd7c : 28/28 checks de contenu passent ; seul rouge = PR gate, dont l'annotation est explicite : [pr-gate] DWELL -- plancher 120 min, reste 113 min ; écoulé à 12:07:00Z. Rien a corriger dans le code : cette jambe est un minuteur — un échec de minuteur, pas un verdict de contenu. Un APPROVE éventuel attendrait la gate rejouée verte après 12:07Z.

Rien détecté côté risque : aucun retrait d'entrée, aucun champ à secret, aucune entrée fantôme.

— [NanoClaw] (myia-ai-01) [09:50Z] · Review mécanique du généré, statique depuis ai-01

@github-actions

Copy link
Copy Markdown
Contributor Author

Path-collision (organ #13359/#13615)

Cette PR #18035 (chore(catalog): scheduled auto-regenerate (long-lived PR)) touche au moins un chemin de fichier aussi modifie par d'autres PRs ouvertes. Risque de double-livraison (meme fichier livre deux fois, 2x le travail et 2x les runs CI). Advisory : parfois legitime (tranches coordonnees, partition paths: explicite, PRs empilees exclues) -- l'organe rend visible, il ne bloque pas.

Le verdict terminal (#15578) signale qu'un cote de la paire est deja sur main. L'organe mesure un recouvrement de chemins ; il ne compare pas le contenu des deux livraisons, donc il ne conclut PAS a une redondance (#15768) : deux PRs peuvent toucher le meme fichier pour des raisons disjointes. L'arbitrage reste a la lane ou au coordinateur.

@jsboige

jsboige commented Sep 27, 2026

Copy link
Copy Markdown
Owner

[ADJOINT PREFLIGHT]
schema: 1
lane: myia-po-2026:CoursIA-2
pr: 18035
head: 06effd7
complete: true
body: read
comments-reviewed: 1
reviews-reviewed: 1
threads-reviewed: 0
threads-unresolved: 0
surfaces-sha256: 833a15b3eb571e7b4d2212df0588cb9c322b44b35999b7bb7ce7bb73ac83c3b6
diff-files: 13
diff-additions: 1302
diff-deletions: 916
checks: latest-wins-green
b0: clear
scope: pass
domain: pass
verdict: READY
[/ADJOINT PREFLIGHT]

Lecture firsthand

  • Body : PR longue-duree rafraichie quotidiennement par catalog-cron.yml (cron 03:37 UTC) maintenant COURSE_CATALOG.generated.{json,md} + marqueurs CATALOG-STATUS + curriculum (docs/curriculum/*.md) + dashboard sante (docs/archive/reference/HEALTH_DASHBOARD.md) canoniques sur main. Maintainer review only, no direct push (issue ci: le contexte requis 'PR gate' rejette tout push bot sur main -- catalog-cron mort, option A de #10133 non viable #10136).
  • Comments : 1 (bot H.4 PASS).
  • Reviews : 1 (clusterManager-Myia COMMENTED 2026-09-27T09:49:05Z, prefixe LGTM contenu verbatim -- recount mecanique base↔head : 1306 -> 1312 entrees (+6/-0), coherence JSON↔MD, 28/28 checks de contenu passent ; APPROVE non satisfait ce tour car DWELL = minuteur 120 min en cours, levee ~12:07Z).
  • Threads inline : 0 non resolu.
  • Diff : 13 fichiers +1302/-916 -- regen catalogue + 8 README mis a jour (CATALOG-STATUS +3/-3 attendus) + curriculum + dashboard.
  • Checks (latest-wins) : PR gate success 12:18:31Z (DWELL expire, gate rejouee verte), perimeter-review-guard success 09:49Z, prose-counts success, fast-lane 8/8, CodeQL success. 28/28 checks de contenu OK.
  • B.0 : pas de Hermes CONCERNS, pas de CHANGES_REQUESTED. LGTM contenu.
  • Scope vs body : coherent.

Verdict

READY.

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

Labels

automation catalog-refresh markdown-table-syntax Table syntax defect in changed files (CODE_SPAN_PIPE, NO_SEP, ...). Advisory. See #10097. paragraph-length Paragraph > 2000 chars (wall-of-text, #15405). Resorb before merge.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants