Repository navigation
Add: interpretation cells for Search-4 LocalSearch-Csharp (N-Queens benchmark, exos) - #14037
Merged
Merged
Conversation
…enchmark, exos) Enrichissement markdown-only : 5 cellules d'interpretation, chacune ancree aux outputs commites (encodage par permutation -> 6 conflits diagonaux initiaux ; benchmark HC 7/15 vs SA/Tabu 15/15 sur 8-Reines, 4/10 vs 10/10 sur 20-Reines) ou a la structure explicitement citee. Code byte-identique (13/13 cellules). Densite 947 -> 1266 c/cell (seuil 1200). Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>
Attestation (post-commit, blobs HEAD) de la paire search-4-localsearch apres l'enrichissement markdown-only du cote C# (PR de contenu associee). Seul le csharp_sha/content_csharp_sha bougent ; content_python_sha intact -> aucun drift du cote Python. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>
Contributor
Contributor
Golden-Set Execution (H.7 P3)✅ 8/8 notebooks passed (certified reproducible)
Pinned lockfile: |
Contributor
Bash Syntax Advisory — shebang / executable-bit warningsSee the |
Contributor
Notebook PR Validation: PASS
Checks: H.1 (no errors), H.3 (execution_count), C.1 (no banned patterns) |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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: MED/notebook-dotnet — lane myia-po-2027:CoursIA — prev: MED/notebook-python #14009 (substance distincte : autre famille, C# et non Python, lectures à dominante méthodologie-du-voisinage, pas méthodologie-de-mesure)
Sujet
Search-4-LocalSearch-Csharp.ipynbsortait à 947 c/cell (seuil 1200). Ajout de 5 cellules markdown d'interprétation, chacune ancrée aux outputs committés ou à la structure du code explicitement citée. Le notebook porte déjà des interprétations aux cellules [3],[5],[7],[9],[11],[14],[16],[18],[20] ; les insertions couvrent les zones de code sans interprétation adjacente :reines[i] = colonne de la reine de la ligne i= permutation → une reine par ligne/colonne par construction, donc seuls les conflits diagonaux restent (la démo rend 6 conflits initiaux) ; objectifConflits= compteur de diagonales, 0 = solution ; espacen!(pasn^n)BenchmarkCS0219(itérateursitHC/itSA/itTScomptés jamais utilisés) = reliquat, code non modifiébreakdès que le coût passe sous leseuil; la vraie calibration est qualité visée vs budget de relances ; stubpass→ à compléterValidation
MyIA.AI.Notebooks/Search/Part1-Foundations/Search-4-LocalSearch-Csharp.ipynb(enrichissement markdown) +scripts/notebook_tools/twin_pairs.d/search-4-localsearch.yaml(rebaseline twin-parity, commit dédié).execution_count+metadatacomparés JSON par index contreorigin/main) — enrichissement markdown-only, exempt de re-exécution (C.2/C.3 exception modifs uniquement markdown).pedagogy_density.py947 → 1266 c/cell (seuil 1200 atteint) ; 29 → 34 cellules ; 5 insertions, 0 deletions.detect_consecutive_code_cells.py0 run ·detect_markdown_rendering.py0 violation ·detect_code_in_markdown_cells.py0 nouveau vs baseline (2).scan_cell_ordering.py1 finding PRÉEXISTANT :SECTION_GAP—## 6. Exercicessaute de 3 à 6 sous le parent<root>. Vérifié sur une copie témoin deorigin/main(/tmp/s4_base.ipynb, cellule QC MomentumStrategy: FIXED (Sharpe 0.459) #22) : le défaut existait avant cette PR et n'est pas causé par elle. Relevé, non traité (hors scope — règle 3 : signaler, pas toucher).Search/Part1-Foundations(search-4-localsearch.yaml),parity_level: semantic— enrichissement markdown-only d'un seul côté + rebaseline (attestation lue sur HEAD blobs, post-commit), workflow documenté comme légitime dans les audits de la paire.notebook-dotnet), chemin disjoint de feat(notebooks): App-7-Wordle 8 lectures ancrees outputs - densite 1009->1347 c/cell, code byte-identique #14009 (Applications/CSP), feat(notebooks,#13410): App-19-WFC 10 lectures ancrees outputs - densite 705->1373 c/cell, code byte-identique #13980 (Applications/Hybrid) et refactor(search,#13777): isoler les face-a-face MGS vs mealpy en sous-serie dediee #13959 (Part4-Metaheuristics).