feat(search,#10382): CSP-7 C# section 6 Hierarchical CSP — Configuration PC sur Choco, accord cross-moteurs CP-SAT - #12741
Open
jsboige wants to merge 10 commits into
Open
feat(search,#10382): CSP-7 C# section 6 Hierarchical CSP — Configuration PC sur Choco, accord cross-moteurs CP-SAT#12741jsboige wants to merge 10 commits into
jsboige wants to merge 10 commits into
Conversation
… + SoftCsp<T>, accord framework vs Choco) Fermeture du gap Semiring du registre twin csp-7-soft : miroir .NET de la section 2 du jumeau Python (Bistarelli-Montanari-Rossi 1997). Interface generique ISemiring<T> (Bottom/Top/Combine/Project) + 3 instances (Boolean/Fuzzy/Weighted), classe SoftCsp<T> (Evaluate = Project depuis Top, SolveBruteForce = ordre a+b=b), demo vacances floues 9 combinaisons, et accord numerique exact framework vs moteur Choco (Rome/Juillet 0.8 = 0.8). Cellules executees via harness console net9.0 + IKVM 8.15.0 (recette #4711, recette memoire console-project + transplant), outputs reels transplantes. Exercices renumerotes 4 -> 6 ; intro et conclusion mises a jour (table de parite + verdict SOTA). Gap residuel : Hierarchical. Co-Authored-By: Claude-Code <noreply@anthropic.com>
…ing (gap residuel : Hierarchical uniquement) Co-Authored-By: Claude-Code <noreply@anthropic.com>
…ec-sequence) Re-exec reelle des 3 cellules exercices dans un harness console (sorties byte-identiques aux outputs committes, verifiees par script), puis renumbering transplant pattern #12159 : Semiring A/B/C = 10/11/12, exercices = 13/14/15. check_exec_ratchet local : CLEAN->CLEAN, 0 regression. corrige le rouge Exec-sequence ratchet de la CI (#12616). Co-Authored-By: Claude-Code <noreply@anthropic.com>
…e exec contigue (blob 1481c446)
…2026-08-23 a 4 espaces (l'indentation 6 cassee par --update rendait le YAML unparseable -> crash des 3 gardes twin/scripts) Co-Authored-By: Claude-Code <noreply@anthropic.com>
…guration PC sur Choco (accord cross-moteurs CP-SAT)
Fil Hierarchical CSP (Mittal-Falkenhainer) du jumeau Python porte cote C# :
MANDATORY (socket, alimentation) postees, STRONG/WEAK reifiees via ifThenElse
(equivalent OnlyEnforceIf de CP-SAT), objectif lexicographique linearise
scalar({strong,weak},{M,1}). Exemple resolu Configuration PC (meme catalogue,
meme M=5 que le Python) + Exercice 4 station de travail 1200 EUR (stub C.1).
Execution reelle (console .NET 9 + IKVM 8.15.0 + Choco 4.10.17, recette #4711) :
meme optimum que CP-SAT — i5-12400/MSI B660M/RTX 3060/CV550/NZXT, 815 EUR,
score 145, strong 0, weak 3, objectif 3, 121 ms. Conclusion : ligne de parite
+ verdict mis a jour (configuration PC passe dans les problemes partages).
Tranche 2 de la paire apres Semiring #12616 (branche empilee dessus).
Voir #10382 (registre twin), See #4711.
Grain: DEEP/notebook-dotnet — lane myia-po-2024:CoursIA — prev: DEEP/lean #12343
Co-Authored-By: Claude-Code <noreply@anthropic.com>
…rchical — SOTA-OK + parity semantic Rebaseline apres eec97c8 (csharp_sha 020dfe9, content cf7d94d6) + reevaluation du verdict : bridge_verdict RECOVERABLE-LOCAL -> SOTA-OK, parity_level surface -> semantic — dernier gap nomme (fil Hierarchical) ferme, accords numeriques x2 (Semiring #12616 Rome 0.8 ; Hierarchical Choco<->CP-SAT objectif 3, meme config 815 EUR). Entree --update 2026-08-24 reindentee 4/6 -> 2/4 espaces (bug #12617, meme correctif manuel que 63639a4). Entrees known_differences non datees devenues fausses : superscription datee appended, prose d'origine preservee. Checker : 157 paires OK, DRIFT=0. Co-Authored-By: Claude-Code <noreply@anthropic.com>
…tion section 6 (ratchet exec-sequence) Renumbering transplant pattern #12159 / precedent 26a25a3 : les 2 cellules section 6 (inserees avant la section Exercices) prennent 13/14, les stubs exercices passent 13/14/15 -> 15/16/17. Permutation pure des labels execution_count, sources et outputs byte-identiques (outputs = execution reelle console .NET 9 + IKVM 8.15.0 + Choco 4.10.17 du commit eec97c8). check_exec_sequence : UNORDERED -> CLEAN (1..17). Co-Authored-By: Claude-Code <noreply@anthropic.com>
Owner
Author
|
Log d'execution reelle du runner console .NET 9 (IKVM 8.15.0 + Choco 4.10.17, recette #4711) — les sources executees sont celles des cellules commitees (transplant) : |
jsboige
commented
Aug 24, 2026
jsboige
left a comment
jsboige
left a comment
Owner
Author
There was a problem hiding this comment.
[Hermes] — DEEP review a141f2a9 (dédup : 0 review cluster sur ce SHA).
Vérifications effectuées sur le notebook committé (fetch raw au SHA de tête, pas le diff seul) :
- Hygiène source : 38 cellules (17 code), jointures
execution_count1→17 contigus, 0 null. - Authenticité outputs : les lignes citées dans le body sont présentes verbatim dans les outputs committés (
Prix total : 815 EUR,Score objectif : 5 x 0 (strong) + 3 (weak) = 3, bilan TDP 315W / alim 550W, accord cross-moteurs CP-SAT objectif 3).ifThenElse/setObjective/elementbien dans les sources. - Discipline C.1 : 0 pattern interdit (
raise NotImplementedError/assert False/1/0) — stub étudiant propre. - Registre twin-parity : delta YAML lu en entier — superscription datée qui préserve la prose d'origine,
parity_level surface→semanticetRECOVERABLE-LOCAL→SOTA-OKmotivés par les 3 tranches (#12159/#12616/celle-ci), réindentation 2/4 conforme au correctif #12617. - Security scan : 0 match (
HF_TOKEN|API_KEY|BEARER|PASSWORD|SECRET|TOKEN\s*=).
Score objectif : … = 3 (121 ms) mais (121 ms) n'apparaît dans aucun output du notebook committé (timings réels : 103/93/9/3/1 ms). L'extrait présenté comme verbatim ne l'est pas à ce timing près — vraisemblablement la mesure du run scratchpad console-runner plutôt que la sortie transplantée. Cosmétique, à ignorer ou corriger au prochain passage sur le fichier.
Verdict : PASS avec la réserve ci-dessus. (Contrainte token : COMMENT only, self-review cap.)
Contributor
|
No new ASCII flowchart introductions in modified notebooks (vs merge-base). Inherited flowcharts are tolerated by design (frozen-inheritance #11840). Detector: |
Contributor
MD hierarchy drift -- a141f2aCette PR augmente le compte de defauts de rendu markdown Corriger (ex. |
Contributor
|
✅ No render volume delta signal (notebooks modified vs merge-base preserve >= 50% of base rendered output per MIME family). See |
Contributor
Golden-Set Execution (H.7 P3)✅ 8/8 notebooks passed (certified reproducible)
Pinned lockfile: |
Contributor
|
No new degraded-mode confessions in modified notebooks (vs merge-base). Inherited confessions are tolerated by design (frozen-inheritance). Detector: |
Contributor
|
✅ No fragmented stream outputs detected. See |
Contributor
github-actions
Bot
added
pedagogy-density-below-threshold
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Grain: DEEP/notebook-dotnet — lane myia-po-2024:CoursIA — prev: DEEP/lean #12343
Summary
Section 6 Hierarchical CSP (Mittal & Falkenhainer 1990 — niveaux MANDATORY > STRONG > WEAK) ajoutee au twin C# de CSP-7-Soft, miroir exact de l'exemple resolu Configuration PC du jumeau Python (meme catalogue, meme budget 820 EUR, meme M = 5). Notebook 33 -> 38 cellules :
M*strong + weak, M > max cout weak) ;arithmcompatibilite socket + alimentation >= TDP+80), STRONG/WEAK reifiees parifThenElse(bool, cstr, cstr)= equivalent exact du doubleOnlyEnforceIf(b)/OnlyEnforceIf(b.Not())de CP-SAT, attributs viamodel.element, objectifscalar({strong, weak}, {M, 1})+setObjective(false, hierCost), bouclewhile(solver.solve())capturant le best ;// TODO etudiant, aucunraise) ;## 6. Exercices->## 7.; ligne tableau + verdict de la conclusion patchees.Empilement
PR empilee sur #12616 (branche
feature/csp7-semiring-section) : ai-01 decide l'ordre de merge — merger #12616 d'abord, puis celle-ci (base = main, le diff cumule inclut donc la section 5 Semiring).Execution reelle (H.1 / D.4)
check_exec_sequence.py: CLEAN (1..17) — renumbering transplant pattern feat(csp-7,#10382): section Fuzzy CSP via Choco dans le twin C# #12159 (precedent 26a25a3, commit 3492a9d), permutation pure des labels, sources et outputs byte-identiques.raise NotImplementedError/assert False/1/0: aucun, verifie par script) ; 17 cellules code, toutesexecution_count != null+ outputs.Registre twin-parity (53d8414 + a141f2a)
Reevaluation :
bridge_verdictRECOVERABLE-LOCAL -> SOTA-OK,parity_levelsurface -> semantic — dernier gap nomme (fil Hierarchical) ferme ; accords numeriques x2 (Semiring #12616 : Rome 0.8 ; Hierarchical : Choco<->CP-SAT objectif 3, meme config 815 EUR). Entreesknown_differencesnon datees devenues fausses : superscription datee appended, prose d'origine preservee. Entrees--updatereindentees 4/6 -> 2/4 espaces (bug #12617, meme correctif manuel que 63639a4). Checker : 157 paires OK, DRIFT=0.Verdict SOTA
SOTA-OK — vrai moteur Choco 4.10.17 via IKVM 8.15.0, la sortie commitee est sa vraie sortie.
See #10382 (issue-parapluie, tranche Hierarchical — livraison partielle, pas de Closes)