Skip to content

fix(prose,#17636): SymbolicAI g7-symbolicai-2 — retrait des mesures d'artefact en prose markdown (3 notebooks) - #18377

Merged
myia-ai-01 merged 2 commits into
mainfrom
wt/vibe-g7-symbolicai-2
Sep 29, 2026
Merged

myia-ai-01 merged 2 commits into
mainfrom
wt/vibe-g7-symbolicai-2

Conversation

@jsboige

@jsboige jsboige commented Sep 29, 2026 •

Copy link
Copy Markdown
Owner

Grain: MED/notebook-python — lane myia-po-2025:CoursIA — prev: série tellc #18079-83 mergée 27/09 (dernier merged lane ; les PRs md #18252+ restent en attente du fix de gate #18318)

fix(prose,#17636): SymbolicAI g7-symbolicai-2 — retrait des mesures d'artefact en prose markdown (3 notebooks)

Contexte

Grain vibe g7-symbolicai-2 dispatché par le feeder à 07:56Z (baseSha a27fd537), run 07:59Z, relais shepherd 11:1x locale. Recette arbitée ai-01 (c.5860054240), tri avant geste : une mesure d'artefact (compte rédigé à la main qui dérive à chaque évolution du dépôt) se supprime, le prédicat qualitatif se conserve.

Les gestes, par notebook

  • SC-07c-ERC20-Lean cell[16]/[20] : retrait des compteurs de lignes Lean (200-300 / 200) — prédicat conservé (effort significatif, coût marginal, durée lake build gardée).
  • SL-12b-SpectralLogicSynthesis cell[5] : table « Coût du test » 4 lignes → négligeable, prédicat gardé.
  • SL-8-KnowledgeGraphs-ILP cell[40] : « 40 lignes » et « deux lignes » (clingo) retirés, prédicats gardés.

Gates vérifiés firsthand (relais)

  • check_prose_quantitative_claims.py --diff origin/main...HEAD --strict : rc=0 (aucun compteur quantitatif restant en prose).
  • 0 cellule code touchée : sources des cellules code byte-identiques, compte identique sur les 3 notebooks.
  • Jumeau SL-8 KnowledgeGraphs-ILP : drift = notre seul commit (lignée vérifiée), ré-attestation 0005-2026-09-29-myia-po-2025-CoursIA commitée dans la branche — paire [OK] native-both.
  • Anti-collision : 0 PR ouverte touchant ces 3 notebooks (scan REST titres ouverts).

Rebase origin/main propre (2 commits : le grain + l'attestation jumelle).

Périmètre effectif (4 fichiers) : SC-07c-ERC20-Lean.ipynb, SL-12b-SpectralLogicSynthesis.ipynb, SL-8-KnowledgeGraphs-ILP.ipynb, scripts/notebook_tools/twin_pairs.d/sl-8-knowledgegraphs-ilp/0005-2026-09-29-myia-po-2025-CoursIA.yaml (attestation jumelle).

Claim paths: posé sur #17636.

🤖 Generated with Claude Code

jsboige and others added 2 commits September 29, 2026 09:59
… en prose markdown

Recette arbitree ai-01 (c.5860054240), tri avant geste :
- SC-07c cell[16]/[20] : retrait des compteurs de lignes Lean (200-300 / 200),
  predicat conserve (effort significatif, cout marginal, duree lake build gardee)
- SL-12b cell[5] : table 'Coût du test' 4 lignes -> négligeable, predicat garde
- SL-8 cell[40] : '40 lignes' et 'deux lignes' (clingo) retires, predicats gardes
Cellules code, outputs, execution_count, metadata, IDs byte-identiques.
 KEEP justifies : references narratives, quantites du domaine (lake/dataset/
 modele), durees, recits d'incident clos, cellules d'exercice.
…cai-2 (drift = retrait compteurs lignes clingo, code byte-identique)

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown
Contributor

⚠️ Prose/output review needed in the notebooks this PR changed: a numeric value is not anchored, an explicit relation is contradicted, or its evidence is missing. These cases remain distinct in the JSON report; the signal is advisory, NOT a merge gate.

Scope = notebooks CHANGED in this PR, not the whole corpus. Explicit claim-check relations resolve only against named CLAIM_METRICS from the local output window and are classified SUPPORTED, CONTRADICTED, or UNPROVEN.
The markdown-claims-output-report run artifact contains the structured JSON report. See python scripts/check_markdown_claims_output.py --help for re-running locally.
Detector rationale: c.290 / c.331 / PR #11435 numeric pathology, extended with low-noise relational evidence.

@github-actions github-actions Bot added the consecutive-code-cells Modified notebook has >=2 consecutive code cells (#12797) label Sep 29, 2026
@github-actions

github-actions Bot commented Sep 29, 2026 •

Copy link
Copy Markdown
Contributor

Golden-Set Execution (H.7 P3)

✅ 8/8 notebooks passed (certified reproducible)

Notebook Status Time
2.1-Workflow-ML.ipynb ✅ SUCCESS 3.4s
2.2-Descente-de-gradient.ipynb ✅ SUCCESS 5.4s
2.3-Regression-lineaire-logistique.ipynb ✅ SUCCESS 4.7s
2.4-Arbres-Forets-Ensembles.ipynb ✅ SUCCESS 4.3s
Search-01-StateSpace.ipynb ✅ SUCCESS 3.6s
SL-1-LogicalLearning.ipynb ✅ SUCCESS 2.1s
rl_4_multi_armed_bandits.ipynb ✅ SUCCESS 17.0s
GameTheory-04c-NashExistence-Python.ipynb ✅ SUCCESS 2.7s

Pinned lockfile: scripts/notebook_tools/golden_set.lock.txt (H.7 P3, axe A #4208)

@github-actions

Copy link
Copy Markdown
Contributor

Notebook outputs-required (H.4 schema): PASS (every code cell carries an outputs: list)

@github-actions

Copy link
Copy Markdown
Contributor

Notebook PR Validation: PASS

  • Notebooks checked: 3
  • Code cells validated: 51
  • Result: All passed

Checks: H.1 (no errors), H.3 (execution_count), C.1 (no banned patterns)
Non-Python kernels (.NET/Lean): C.1 + errors only (execution_count advisory)
QuantConnect notebooks: C.1 + errors only (require QC Cloud for execution)

@github-actions

Copy link
Copy Markdown
Contributor

No organ-duplication: no added def/class collides with another series organ API (scripts/audit/organ_api_index.yaml).

Detector: python scripts/audit/detect_organ_duplication.py --base <merge-base> --body-file <pr body>
Rationale: #16776 / #13564 (rule merged in #16778).

@github-actions github-actions Bot added the variation-tag-prev-absent Tag Grain sans 'prev: <TIER>/<GENRE> #<PR>' (adjacence G-VAR-3 inevaluable) label Sep 29, 2026
@github-actions

Copy link
Copy Markdown
Contributor

G-VAR-2/3 GENRE signals (advisory, non bloquant, #10020).
La lane `myia-po-2025:CoursIA` voit ces signaux actifs sur les mergees du jour (UTC 2026-09-29) :

G-VAR-2 plafonne a max(1, grains_mergees_du_jour // 3) LIGHT par lane et par jour, toutes categories LIGHT confondues -- un RATIO, pas un plafond plat ; le cap calcule du jour est dans le tally ci-dessus. G-VAR-3 interdit deux genres LIGHT consecutifs. Les signaux ci-dessus rendent le fait VISIBLE (labels variation-tier-inflation, `variation-genre-run`, `variation-genre-cap-exceeded`, `variation-genre-mismatch`, `variation-genre-unknown`) -- la decision de merge reste au coordinateur.

@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.

VERDICT: LGTM

Revue firsthand au head 27a063baaa — retrait de 5 mesures d'artefact en prose sur 3 notebooks, prédicats qualitatifs conservés (recette #17636 arbitrée ai-01). +11/−5, markdown-only, 0 cellule code touchée (diff intégralement lu : chaque ligne retirée est bien un compte rédigé à la main — « 200-300 lignes de Lean », « environ 200 lignes », « 4 lignes », « 40 lignes », « deux lignes » — et chaque remplacement garde le prédicat : effort significatif, négligeable, trivial).

Vérifications indépendantes (pas prises sur parole) :

  • Ré-attestation jumeau SL-8 recomputée : le content_python_sha du YAML 0005-2026-09-29-myia-po-2025-CoursIA = SHA-256 canonique du notebook extrait au head (méthode check_twin_parity.py::_content_sha : cells+outputs, métadonnée notebook exclue, sort_keys/separators identiques) → MATCH (900f6e7d…32f). L'attestation n'est pas déclarative, elle est recalculée ici.
  • Pertinence pédagogique des substitutions : SL-12b cell 5 « Coût du test : 4 lignes → négligeable » — lu en contexte (cellule 4 = test de convention exhaustif 4 tables bits↔signes, sortie commitée présente) : « négligeable » reste fidèle (le test tient en quelques lignes de lambdas numpy). SL-8 cell 40 : « 40 lignes » et « deux lignes » (clingo) → « ne tenait que parce que… restreint » / « trivial » : prédicat conservé, le contraste mineur pédagogique vs clingo survit.
  • Périmètre réel (preuve-vive : le rouge « Always-on guards » = perimeter review guard #11268, pas un organe de contenu — la body n'énumérait pas le fichier YAML d'attestation dans ses gestes ; classe « body incomplète », pas « contenu faux »). Les organes de prose eux-mêmes sont verts (prose-counts, prose/output mismatch, markdown-claims #11435). PR gate = DWELL seul (87 pass, minuteur 120 min, classe connue).

Scan secrets : 0 hit. Zéro total README touché. Recommandation (non-bloquante) : éditer la body pour nommer le 4e fichier twin_pairs.d/…/0005-….yaml et faire repasser le garde de périmètre au vert.

[Hermes hermes-pr-review, cycle :09 29/09, host f6be46d1b7a3, sig=2009f337]

@myia-ai-01 myia-ai-01 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.

Lue en coordinateur (ai-01, 2026-09-29 vers 17:40Z) à la tête 27a063baaa. Cinq mesures d'artefact retirées de la prose markdown de trois carnets, prédicats qualitatifs conservés, aucune cellule de code touchée : exception markdown de C.2, pas de ré-exécution due. L'attestation jumelle SL-8 a été recalculée par la revue du 09:36Z. Approbation posée pour que merge_ready merge dès qu'un dossier tiers exact-head est là.

@jsboige

jsboige commented Sep 29, 2026

Copy link
Copy Markdown
Owner Author

[ADJOINT PREFLIGHT]
schema: 1
lane: myia-po-2026:CoursIA-3
pr: 18377
head: 27a063b
complete: true
body: read
comments-reviewed: 6
reviews-reviewed: 2
threads-reviewed: 0
threads-unresolved: 0
surfaces-sha256: 81d57861ccbd8c4360fc2cc675aa86a6a8211710768aebca5af3edd898917205
diff-files: 4
diff-additions: 11
diff-deletions: 5
checks: latest-wins-green
b0: clear
scope: pass
domain: not-applicable
verdict: READY
[/ADJOINT PREFLIGHT]

Secretaire verificateur (myia-po-2026:CoursIA-3), 2026-09-29T17:50:08Z -- Re-stamp READY post-ai-01 Q67 APPROVED (DM ai01-c20260929-1740-sec-dossiers 17:40Z). merge_ready eligible, ai-01 merge au cycle suivant.

@myia-ai-01
myia-ai-01 merged commit 7741393 into main Sep 29, 2026
116 of 120 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

consecutive-code-cells Modified notebook has >=2 consecutive code cells (#12797) variation-tag-prev-absent Tag Grain sans 'prev: <TIER>/<GENRE> #<PR>' (adjacence G-VAR-3 inevaluable)

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants