---

name: python-pro description: Write idiomatic Python code with advanced features like decorators, generators, and async/await. Optimizes performance, implements design patterns, and ensures comprehensive testing. Use for ML training, analytics tools, performance profiling, or any Python heavy lifting. metadata: short-description: Advanced Python patterns

Python Pro - Advanced Python Patterns

When to Use This Skill

Invoke python-pro for:

• CatBoost model training (ml/train_models.py, feature_builder.py)
• Analytics tools optimization (async batching, caching)
• Performance profiling (bottleneck identification)
• Advanced Python patterns (decorators, generators, context managers)
• Heavy data processing on campaign datasets
• Executing ML designs from @sama-2.0

Use fastapi-production-patterns instead for:

• API endpoints, routing, middleware
• Pydantic validation, request/response models
• CORS configuration, authentication middleware
• FastAPI-specific patterns (dependency injection at API layer)

Clear Boundary:

---

Executable Scripts

Run these scripts directly for profiling and debugging:

Profile a Function

Copypython scripts/profile_function.py app.ml.feature_builder build_features
python scripts/profile_function.py app.agents.analyst gather_evidence_async --args '{"candidate": {"list_id": "GM_30D"}}'

Compare Two Implementations

Copypython scripts/benchmark_compare.py app.tools.v1:analyze app.tools.v2:analyze_async --runs 10

Check Memory Usage

Copypython scripts/memory_check.py app.ml.feature_builder build_all_features --args '{"n_campaigns": 1000}'

---

{{PROJECT_NAME}} ML System

Use references/mission_inbox_ml.md for ML-specific documentation:

• Model locations (ml/models/*.cbm)
• Training commands (python ml/train_models.py)
• Feature list (110 features from FeatureBuilder)
• How MLPredictor serves predictions to Analyst Agent
• EPC lookup (historical, NOT ML)
• Database tables used for training

---

Core Patterns and Examples

Use references/patterns.md for detailed code patterns and examples across:

• Decorators (caching, timing, retries, validation)
• Generators (lazy feature building, chunking, async generators)
• Async/concurrency (batching, sync-to-async, semaphores)
• Profiling (cProfile, line_profiler, memory_profiler, benchmarking)
• Type hints and static analysis (TypedDict, Protocol, Generic, mypy/ruff/black)
• Testing (fixtures, parametrization, async tests, mocking)
• Design patterns (strategy/factory for ML and tool creation)
• Quick reference cheat sheet

---

Usage Guidance

• Prefer clear, typed interfaces for analytics and ML modules.
• Favor async batching when tool calls are independent.
• Profile before optimizing; keep hotspots visible.