Bienvenue dans le depot CoursIA, plateforme educative complete pour l'apprentissage de l'intelligence artificielle en C# et Python.
- Introduction
- Cartographie complete
- Series de notebooks
- Configuration et API Keys
- Kernels Jupyter
- Outils externes
- Mise en route
- Infrastructure Docker
- Scripts et validation
- Outils Claude Code
- Contribution
Ce depot contient 255+ notebooks Jupyter interactifs couvrant :
- IA Symbolique : Logiques formelles, argumentation, verification formelle (Lean 4, Tweety, Z3)
- Probabilites : Inference bayesienne, modeles graphiques (Infer.NET)
- Theorie des jeux : Nash, jeux evolutionnaires, cooperatifs, CFR, OpenSpiel
- Machine Learning : ML.NET, algorithmes genetiques
- IA Generative : OpenAI, LLMs, generation d'images (DALL-E, FLUX, Qwen, SD3.5)
- Trading Algorithmique : QuantConnect LEAN, ML/DL/RL pour strategies de trading
Les notebooks sont en C# (.NET Interactive), Python et Lean 4, avec une documentation pedagogique complete.
CoursIA/
├── MyIA.AI.Notebooks/ # 255+ notebooks interactifs
│ ├── GenAI/ # IA Generative (55+ notebooks)
│ │ ├── 00-GenAI-Environment/# Setup et configuration (6 notebooks)
│ │ ├── Image/ # Generation d'images (19 notebooks)
│ │ │ ├── 01-Foundation/ # DALL-E 3, GPT-5, Forge
│ │ │ ├── 02-Advanced/ # Qwen, FLUX, SD3.5, Z-Image
│ │ │ ├── 03-Orchestration/# Multi-modeles, workflows
│ │ │ └── 04-Applications/ # Production, contenu educatif
│ │ ├── Texte/ # LLMs et generation texte (10 notebooks)
│ │ ├── SemanticKernel/ # Microsoft Semantic Kernel (14 notebooks)
│ │ └── Vibe-Coding/ # Claude Code et Roo Code tutorials
│ │
│ ├── SymbolicAI/ # IA Symbolique (47+ notebooks)
│ │ ├── Tweety/ # TweetyProject - 10 notebooks
│ │ ├── Lean/ # Lean 4 - 10 notebooks
│ │ ├── Argument_Analysis/ # Analyse argumentative - 6 notebooks
│ │ └── Planners/ # Fast-Downward, PDDL
│ │
│ ├── GameTheory/ # Theorie des jeux (26 notebooks)
│ │ ├── GameTheory-1 to 17 # Notebooks principaux
│ │ └── *b, *c variants # Lean + Python side tracks
│ │
│ ├── Probas/Infer/ # Infer.NET - 20 notebooks
│ ├── Sudoku/ # Resolution Sudoku (11 notebooks)
│ ├── Search/ # Recherche et optimisation (5 notebooks)
│ ├── ML/ # Machine Learning (14 notebooks)
│ ├── RL/ # Reinforcement Learning (3 notebooks)
│ ├── QuantConnect/ # Trading algorithmique + AI (27 notebooks Python)
│ ├── IIT/ # PyPhi - Information integree (1 notebook)
│ ├── Probas/ # Probabilites (22 notebooks)
│ ├── EPF/ # Devoirs etudiants (4 notebooks)
│ └── Config/ # Configuration API (settings.json)
│
├── .claude/ # Configuration Claude Code
│ ├── agents/ # 10 agents specialises
│ └── commands/ # 6 skills (commandes slash)
│
├── scripts/ # Scripts utilitaires
│ ├── verify_notebooks.py # Verification multi-famille
│ ├── extract_notebook_skeleton.py # Extraction structure
│ └── genai-stack/ # Validation GenAI
│
├── docker-configurations/ # Infrastructure Docker GPU
│ ├── services/
│ │ ├── comfyui-qwen/ # ComfyUI + Qwen Image Edit
│ │ ├── orchestrator/ # Multi-services (FLUX, SD3.5)
│ │ ├── vllm-zimage/ # Z-Image/Lumina
│ │ └── forge-turbo/ # Forge SD
│ └── shared/ # Modeles et cache partages
│
├── GradeBookApp/ # Systeme de notation etudiants
├── notebook-infrastructure/ # Papermill automation
└── MyIA.AI.Shared/ # Bibliotheque C# partagee
| Categorie | Notebooks | Kernels | Duree estimee | API requise |
|---|---|---|---|---|
| SymbolicAI | 47+ | Python, Lean 4 | ~25h | OpenAI (optionnel) |
| GameTheory | 26 | Python, Lean 4 | ~18h30 | OpenAI (optionnel) |
| Infer.NET | 20 | .NET C# | ~17h | - |
| GenAI | 55+ | Python | ~25h | OpenAI/Anthropic |
| Sudoku | 11 | C#, Python | ~2h | - |
| Search | 5 | C#, Python | ~1h10 | - |
| ML | 14 | C#, Python | ~4h | - |
| RL | 3 | Python | ~2h | - |
| QuantConnect | 27 | Python | ~30h | QuantConnect (gratuit) |
| IIT | 1 | Python | ~1h30 | - |
| Total | 255+ | Mixed | ~130h | - |
47+ notebooks couvrant les logiques formelles, l'argumentation computationnelle et la verification formelle.
| Serie | Notebooks | Contenu | Prerequis | README |
|---|---|---|---|---|
| Tweety | 10 | TweetyProject, logiques PL/FOL/DL, argumentation Dung, ASPIC+ | JDK 17+ (auto) | README |
| Lean | 10 | Lean 4, types dependants, tactiques, Mathlib, LLM integration | WSL, elan | README |
| Argument_Analysis | 6 | Analyse argumentative multi-agents avec Semantic Kernel | OpenAI API | README |
| Planners | 1 | Fast-Downward, planification PDDL | Python | README |
| Autres | 14+ | Z3, OR-Tools, RDF.NET | Varies | - |
Notebooks Tweety (detail) :
| Notebook | Contenu | Outils externes |
|---|---|---|
| Tweety-1-Setup | JDK, JPype, libs | - |
| Tweety-2-Basic-Logics | PL, FOL, SAT4J, pySAT | pySAT |
| Tweety-3-Advanced-Logics | DL, Modal, QBF | SPASS (admin req.) |
| Tweety-4-Belief-Revision | CrMas, MUS, MaxSAT | MARCO |
| Tweety-5-Abstract-Argumentation | Dung, semantiques | - |
| Tweety-6-Structured-Argumentation | ASPIC+, DeLP, ASP | Clingo |
| Tweety-7a-Extended-Frameworks | ADF, Bipolar, WAF | - |
| Tweety-7b-Ranking-Probabilistic | Ranking semantics | - |
| Tweety-8-Agent-Dialogues | Protocoles de dialogue | - |
| Tweety-9-Preferences | Voting, social choice | - |
26 notebooks (17 principaux + 9 side tracks) combinant Python et Lean 4.
| Partie | Notebooks | Contenu | Kernel |
|---|---|---|---|
| Fondations | 1-6 | Forme normale, Nash, minimax, jeux evolutionnaires | Python |
| Jeux dynamiques | 7-12 | Forme extensive, backward induction, jeux bayesiens | Python |
| Avances | 13-17 | CFR, jeux differentiels, cooperatifs, mechanism design, MARL | Python + OpenSpiel |
| Side tracks b | 2b, 4b, 8b, 15b, 16b | Formalisations Lean 4 | Lean 4 (WSL) |
| Side tracks c | 2c, 4c, 8c, 15c, 16c | Approfondissements Python | Python |
22 notebooks couvrant l'inference bayesienne avec Infer.NET (C#) et Pyro (Python).
| Section | Notebooks | Kernel | Contenu |
|---|---|---|---|
| Racine | 2 | Python/C# | Infer-101 (intro), Pyro_RSA_Hyperbole (pragmatique) |
| Infer/ 1-13 | 13 | C# | Fondamentaux, modeles classiques, debugging |
| Infer/ 14-20 | 7 | C# | Theorie de la decision bayesienne |
11 notebooks (7 C#, 4 Python) illustrant differentes approches algorithmiques.
| Approche | Notebooks | Technologies | Kernel |
|---|---|---|---|
| Backtracking | 1, Python | MRV, recherche exhaustive | C#, Python |
| Genetique | 2, Python | GeneticSharp, PyGAD | C#, Python |
| Contraintes | 3, Python | OR-Tools CP/SAT/MIP | C#, Python |
| SMT | 4, Python | Z3, bitvectors | C#, Python |
| Couverture exacte | 5, Python | Dancing Links (DLX) | C#, Python |
| Probabiliste | 6 | Infer.NET | C# |
Note : Les notebooks C# utilisent #!import et necessitent une execution cellule par cellule (Papermill incompatible).
5 notebooks sur les algorithmes de recherche et les metaheuristiques.
| Notebook | Kernel | Contenu |
|---|---|---|
| CSPs_Intro | Python | Programmation par contraintes, AC-3, N-Queens, Min-Conflicts |
| Exploration | Python | BFS, DFS, A*, Hill Climbing, Simulated Annealing |
| GeneticSharp-EdgeDetection | C# | Detection de bords avec GeneticSharp |
| Portfolio_Optimization | C# | Optimisation de portefeuille financier |
| PyGad-EdgeDetection | Python | Detection de bords avec PyGAD |
55+ notebooks organises en plusieurs sous-domaines.
| Sous-domaine | Notebooks | Contenu | Services requis |
|---|---|---|---|
| 00-Environment | 6 | Setup, Docker, API, validation, deploiement local | - |
| Image/ | 19 | Generation d'images (4 niveaux) | OpenAI/Docker GPU |
| Texte/ | 10 | OpenAI, Prompts, Structured Outputs, RAG, Reasoning, Production | OpenAI API |
| SemanticKernel/ | 14 | SK Fundamentals a MCP, NotebookMaker, templates | OpenAI API |
| Vibe-Coding/ | 5+ | Notebooks CLI Claude Code + ateliers Roo Code | Claude/Roo |
Structure Image/ :
| Niveau | Contenu |
|---|---|
| 01-Foundation | DALL-E 3, GPT-5, Forge SD-XL Turbo, Qwen |
| 02-Advanced | Qwen Image Edit 2509, FLUX, SD 3.5, Z-Image/Lumina |
| 03-Orchestration | Comparaison multi-modeles, workflows, optimisation |
| 04-Applications | Contenu educatif, workflows creatifs, production |
1 notebook sur PyPhi et la theorie de l'information integree.
| Notebook | Contenu | Duree |
|---|---|---|
| Intro_to_PyPhi | TPM, Phi, CES, Causation actuelle, Macro-subsystemes | ~90 min |
14 notebooks couvrant ML.NET (C#) et Python Data Science avec agents IA.
| Section | Notebooks | Contenu |
|---|---|---|
| ML.NET | 5 | Introduction, Features, Entrainement, AutoML, Evaluation |
| Python Foundations | 2 | NumPy, Pandas |
| AI Agents Workshop | 7 | RFP Analysis, CV Screening, Data Wrangling, First Agent |
3 notebooks sur Stable Baselines3 et l'apprentissage par renforcement.
| Notebook | Contenu | Duree |
|---|---|---|
| stable_baseline_1 | Introduction PPO, CartPole | ~30 min |
| stable_baseline_2 | Wrappers, sauvegarde, callbacks | ~40 min |
| stable_baseline_3 | HER, goal-conditioned RL, SAC/DDPG | ~45 min |
27 notebooks Python sur le trading algorithmique avec QuantConnect LEAN, incluant ML/DL/RL/LLM.
| Phase | Notebooks | Contenu |
|---|---|---|
| Fondations LEAN | 01-04 | Setup, Platform Fundamentals, Data Management, Research |
| Universe & Assets | 05-08 | Universe Selection, Options, Futures/Forex, Multi-Asset |
| Trading Avance | 09-12 | Order Types, Risk Management, Indicators, Backtesting |
| Algorithm Framework | 13-15 | Alpha Models, Portfolio Construction, Optimization |
| Data Alternatives | 16-17 | Alternative Data, Sentiment Analysis |
| ML/DL/AI | 18-27 | Features, Classification, Regression, Deep Learning, RL, LLM |
Caracteristiques :
- Cloud-first (QuantConnect free tier)
- 9 notebooks dedies ML/DL/RL/LLM
- Production-ready (deployment live)
README QuantConnect | Getting Started
| Famille | Fichier .env | Variables cles |
|---|---|---|
| GenAI | MyIA.AI.Notebooks/GenAI/.env |
OPENAI_API_KEY, ANTHROPIC_API_KEY, COMFYUI_API_TOKEN |
| Argument_Analysis | MyIA.AI.Notebooks/SymbolicAI/Argument_Analysis/.env |
OPENAI_API_KEY, GLOBAL_LLM_SERVICE, BATCH_MODE |
| Lean | MyIA.AI.Notebooks/SymbolicAI/Lean/.env |
OPENAI_API_KEY, GITHUB_TOKEN, LEAN_VERSION |
| GameTheory | MyIA.AI.Notebooks/GameTheory/.env |
BATCH_MODE, OPENSPIEL_NUM_THREADS |
| QuantConnect | MyIA.AI.Notebooks/QuantConnect/.env |
QC_USER_ID, QC_API_TOKEN, QC_ORG_ID |
| C# Notebooks | MyIA.AI.Notebooks/Config/settings.json |
apikey, model, type (openai/azure) |
| Docker ComfyUI | docker-configurations/services/comfyui-qwen/.env |
CIVITAI_TOKEN, HF_TOKEN, COMFYUI_BEARER_TOKEN |
# GenAI
cp MyIA.AI.Notebooks/GenAI/.env.example MyIA.AI.Notebooks/GenAI/.env
# Editer et ajouter: OPENAI_API_KEY, ANTHROPIC_API_KEY
# Argument Analysis
cp MyIA.AI.Notebooks/SymbolicAI/Argument_Analysis/.env.example MyIA.AI.Notebooks/SymbolicAI/Argument_Analysis/.env
# Editer et ajouter: OPENAI_API_KEY
# Lean (pour notebooks 7-10)
cp MyIA.AI.Notebooks/SymbolicAI/Lean/.env.example MyIA.AI.Notebooks/SymbolicAI/Lean/.env
# Editer et ajouter: OPENAI_API_KEY, GITHUB_TOKEN
# GameTheory
cp MyIA.AI.Notebooks/GameTheory/.env.example MyIA.AI.Notebooks/GameTheory/.env
# C# Notebooks
cp MyIA.AI.Notebooks/Config/settings.json.openai-example MyIA.AI.Notebooks/Config/settings.json
# Editer et ajouter: apikey
# Docker ComfyUI
cp docker-configurations/services/comfyui-qwen/.env.example docker-configurations/services/comfyui-qwen/.env
# Editer et ajouter: CIVITAI_TOKEN, HF_TOKENGenAI (.env.template) :
# API principale
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
OPENROUTER_API_KEY=sk-or-... # Alternative multi-modeles
# Services Docker
COMFYUI_API_URL=http://localhost:8188
COMFYUI_API_TOKEN=...
# Configuration
DEFAULT_VISION_MODEL=gpt-5-mini
GENAI_TIMEOUT_SECONDS=120
GENAI_MAX_RETRIES=3Lean (.env.example) :
# LLM Integration (notebooks 7-10)
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
# LeanDojo (notebook 10)
GITHUB_TOKEN=ghp_...
# Lean configuration
LEAN_VERSION=4.3.0
LEAN_TIMEOUT=30| Famille | Kernel | Installation |
|---|---|---|
| Python notebooks | python3 |
Conda mcp-jupyter-py310 |
| C# notebooks | .net-csharp |
dotnet tool install -g Microsoft.dotnet-interactive |
| Lean 4 | lean4_jupyter |
Via elan (WSL uniquement) |
Python :
python -m venv venv
venv\Scripts\activate # Windows
pip install jupyter ipykernel
python -m ipykernel install --user --name=coursia --display-name "Python (CoursIA)".NET Interactive :
dotnet tool install -g Microsoft.dotnet-interactive
dotnet interactive jupyter installLean 4 (WSL uniquement) :
# Dans WSL
curl https://raw.githubusercontent.com/leanprover/elan/master/elan-init.sh -sSf | sh
pip install lean4_jupyter
python -m lean4_jupyter.install| Probleme | Impact | Contournement |
|---|---|---|
Papermill + #!import |
Notebooks C# avec imports bloquent | Execution cellule par cellule |
| Lean sur Windows | signal.SIGPIPE non supporte | Utiliser WSL |
| Cold start .NET | Premier demarrage 30-60s | Relancer apres timeout |
| Outil | Version | Notebooks | Installation |
|---|---|---|---|
| Z3 SMT Solver | 4.12+ | Sudoku, SymbolicAI, Search | pip install z3-solver |
| OR-Tools | 9.8+ | Sudoku, Search, SymbolicAI | pip install ortools |
| Tweety | 1.28 | Tweety series | Auto-telecharge (JARs) |
| JDK | 17+ | Tweety, Argument_Analysis | Auto-telecharge (Zulu portable) |
| PyPhi | 1.2+ | IIT | pip install pyphi |
| Lean 4 | 4.3+ | Lean, GameTheory | Via elan (WSL) |
| Mathlib4 | Latest | Lean 6+ | Auto avec lake |
| OpenSpiel | 1.4+ | GameTheory 13-15 | pip install open_spiel |
| Infer.NET | 0.4+ | Probas/Infer | Via NuGet |
| Clingo | 5.6+ | Tweety-6 | Installation manuelle |
| pySAT | 1.8+ | Tweety-2 | pip install python-sat |
| Outil | Usage | Note |
|---|---|---|
| SPASS | Logique modale (Tweety-3) | Requiert droits admin Windows |
| EProver | FOL prover | Linux uniquement |
| MARCO | MUS enumeration | Avec Z3 |
| Fast-Downward | Planification PDDL | Auto-compilation |
- Python 3.10+ avec pip
- .NET 9.0 SDK (pour notebooks C#)
- Visual Studio Code avec extensions Python, Jupyter, .NET Interactive
- WSL (recommande pour Lean et certains outils)
- Docker + GPU (optionnel, pour GenAI avance)
# 1. Cloner le depot
git clone https://github.com/jsboige/CoursIA.git
cd CoursIA
# 2. Environnement Python
python -m venv venv
venv\Scripts\activate # Windows
pip install jupyter openai anthropic python-dotenv
# 3. Kernel Python
python -m ipykernel install --user --name=coursia --display-name "Python (CoursIA)"
# 4. Packages .NET
dotnet restore MyIA.CoursIA.sln
# 5. Configuration API (choisir selon besoins)
cp MyIA.AI.Notebooks/Config/settings.json.openai-example MyIA.AI.Notebooks/Config/settings.json
cp MyIA.AI.Notebooks/GenAI/.env.example MyIA.AI.Notebooks/GenAI/.env
# Editer les fichiers et ajouter les cles APISudoku/Search (aucune config requise) :
pip install z3-solver ortools numpy matplotlibTweety (JDK auto-telecharge) :
pip install jpype1 python-sat
# Executer Tweety-1-Setup.ipynb pour telecharger JDK et JARsLean (WSL requis) :
# Dans WSL
curl https://raw.githubusercontent.com/leanprover/elan/master/elan-init.sh -sSf | sh
pip install lean4_jupyter openai anthropic
python -m lean4_jupyter.install
# Valider: python MyIA.AI.Notebooks/SymbolicAI/Lean/scripts/validate_lean_setup.pyGameTheory :
pip install numpy scipy matplotlib nashpy open_spiel networkx
cp MyIA.AI.Notebooks/GameTheory/.env.example MyIA.AI.Notebooks/GameTheory/.envGenAI (Docker GPU recommande) :
pip install -r MyIA.AI.Notebooks/GenAI/requirements.txt
cp MyIA.AI.Notebooks/GenAI/.env.example MyIA.AI.Notebooks/GenAI/.env
# Editer .env avec API keys- ML - Introduction au Machine Learning (ML.NET et Python)
- Sudoku - Algorithmes de resolution (backtracking, contraintes)
- Search - Recherche et optimisation
- RL - Reinforcement Learning avec Stable Baselines3
- SymbolicAI/Tweety - Logiques formelles et argumentation
- Probas - Inference bayesienne (Infer.NET, Pyro)
- IIT - Theorie de l'information integree (PyPhi)
- GameTheory - Theorie des jeux
- SymbolicAI/Lean - Verification formelle (WSL requis)
- GenAI - IA generative (API keys requises)
| Service | Port | GPU | VRAM | Description |
|---|---|---|---|---|
| ComfyUI-Qwen | 8188 | Oui | ~29GB | Qwen Image Edit 2509 |
| FLUX.1-dev | 8189 | Oui | ~10GB | Text-to-image |
| Stable Diffusion 3.5 | 8190 | Oui | ~12GB | Image generation |
| Z-Image/Lumina | 8001 | Oui | ~10GB | Lumina-Next-SFT |
| Orchestrator | 8090 | Non | - | Service management |
# ComfyUI seul
cd docker-configurations/services/comfyui-qwen
cp .env.example .env
docker-compose up -d
# Multi-services (orchestrator)
cd docker-configurations/services/orchestrator
docker-compose up -dVariables .env requises :
# Tokens pour telecharger les modeles
CIVITAI_TOKEN=...
HF_TOKEN=...
# Authentification ComfyUI
COMFYUI_BEARER_TOKEN=...
COMFYUI_USERNAME=admin
COMFYUI_PASSWORD=...
# GPU
GPU_DEVICE_ID=0
CUDA_VISIBLE_DEVICES=0docker-configurations/
├── shared/
│ ├── models/ # Cache modeles (~50GB+)
│ ├── cache/ # HuggingFace cache
│ └── outputs/ # Images generees
└── .secrets/ # Tokens (read-only)
| Script | Chemin | Usage |
|---|---|---|
notebook_tools.py |
scripts/ |
Outil consolide : skeleton, validate, analyze, check-env |
notebook_helpers.py |
scripts/ |
Helpers pour manipulation notebooks et iteration |
extract_notebook_skeleton.py |
scripts/ |
Extraction structure pour README |
validate_notebooks.py |
scripts/genai-stack/ |
Validation GenAI via Papermill |
validate_stack.py |
scripts/genai-stack/ |
Validation ecosysteme complet |
check_vram.py |
scripts/genai-stack/ |
Verification VRAM disponible |
validate_lean_setup.py |
MyIA.AI.Notebooks/SymbolicAI/Lean/scripts/ |
Validation environnement Lean |
test_notebooks.py |
MyIA.AI.Notebooks/Probas/Infer/scripts/ |
Tests Infer.NET |
Note : Les anciens scripts de maintenance sont archives dans scripts/archive/.
# Outil consolide notebook_tools.py (recommande)
python scripts/notebook_tools.py skeleton MyIA.AI.Notebooks/Sudoku --output markdown
python scripts/notebook_tools.py validate MyIA.AI.Notebooks/Sudoku --quick
python scripts/notebook_tools.py analyze MyIA.AI.Notebooks/Sudoku
python scripts/notebook_tools.py check-env Sudoku
# Helpers pour manipulation de cellules
python scripts/notebook_helpers.py list notebook.ipynb
python scripts/notebook_helpers.py analyze notebook.ipynb
python scripts/notebook_helpers.py get-source notebook.ipynb 5
python scripts/notebook_helpers.py get-output notebook.ipynb 5
# Extraction structure (alternatif)
python scripts/extract_notebook_skeleton.py MyIA.AI.Notebooks/Sudoku --output markdown
# Validation stack GenAI
python scripts/genai-stack/validate_stack.py
# Validation Lean
python MyIA.AI.Notebooks/SymbolicAI/Lean/scripts/validate_lean_setup.py --wslLe workflow .github/workflows/notebook-validation.yml valide automatiquement :
- Format des notebooks (JSON valide)
- Syntaxe Python/C#
- Execution de base (timeout 60s)
| Commande | Description |
|---|---|
/verify-notebooks [target] |
Verifier et tester les notebooks |
/enrich-notebooks [target] |
Enrichir avec du contenu pedagogique |
/cleanup-notebooks [target] |
Nettoyer et reorganiser le markdown |
/build-notebook [topic] |
Construire un nouveau notebook from scratch |
/execute-notebook [path] |
Executer un notebook via MCP Jupyter |
/validate-genai |
Valider le stack GenAI complet |
Options :
--quick: Structure uniquement (pas d'execution)--fix: Correction automatique des erreurs--python-only/--dotnet-only: Filtrer par kernel--consecutive: Focus sur cellules de code consecutives--iterate: Iteration sur cellules jusqu'a objectif atteint--dry-run: Lister sans modifier
Exemples :
/verify-notebooks Sudoku --quick
/enrich-notebooks Infer --consecutive
/cleanup-notebooks Tweety --dry-run| Agent | Fichier | Mission |
|---|---|---|
notebook-enricher |
.claude/agents/notebook-enricher.md |
Enrichissement pedagogique |
infer-notebook-enricher |
.claude/agents/infer-notebook-enricher.md |
Specialisation Infer.NET |
notebook-cleaner |
.claude/agents/notebook-cleaner.md |
Nettoyage markdown |
notebook-cell-iterator |
.claude/agents/notebook-cell-iterator.md |
Iteration sur cellules |
readme-updater |
.claude/agents/readme-updater.md |
Mise a jour README |
notebook-designer |
.claude/agents/notebook-designer.md |
Conception de notebooks |
notebook-executor |
.claude/agents/notebook-executor.md |
Execution de notebooks |
notebook-iterative-builder |
.claude/agents/notebook-iterative-builder.md |
Construction iterative |
notebook-validator |
.claude/agents/notebook-validator.md |
Validation de notebooks |
readme-hierarchy-auditor |
.claude/agents/readme-hierarchy-auditor.md |
Audit et maintenance hierarchie README |
Claude Code dispose d'un MCP pour executer les notebooks :
| Categorie | Outils |
|---|---|
| Lecture/Ecriture | read_notebook, write_notebook, create_notebook |
| Cellules | read_cells, add_cell, update_cell, remove_cell |
| Kernels | list_kernels, manage_kernel (start/stop/restart) |
| Execution | execute_on_kernel, execute_notebook |
| Jobs async | manage_async_job (status, logs, cancel) |
- Fork le depot
- Creer une branche (
git checkout -b feature/nouvelle-fonctionnalite) - Commit (
git commit -m 'Add: nouvelle fonctionnalite') - Push (
git push origin feature/nouvelle-fonctionnalite) - Ouvrir une Pull Request
- Pas d'emojis dans le code et les fichiers generes
- PEP 8 pour Python, conventions standard pour C#
- Branches :
type/nom-court(ex:feature/notebook-transformers) - Commits :
Type: description(ex:Add: notebook sur les Transformers)
Chaque famille de notebooks doit avoir un .env.example documentant :
- Les variables requises vs optionnelles
- Le format attendu (API key, URL, boolean)
- Les valeurs par defaut
Ce projet est sous licence MIT - voir LICENSE.
Repository: https://github.com/jsboige/CoursIA