Custom Python package for the Advanced ML AT2 assignment.
Includes:
- Data fetching (Open-Meteo API)
- Feature engineering (classification + regression)
- Model training utilities (XGBoost, CatBoost)
- Metrics, persistence, and plotting helpers
The FastAPI app is in a separate repo:
• / → Overview + endpoints.
• /health/ → Health check.
• /predict/rain/?date=YYYY-MM-DD → Predict rain in 7 days.
• /predict/precipitation/fall/?date=YYYY-MM-DD → Predict 3-day precipitation sum.
⸻
Git Issues • CRLF/LF warnings: harmless, Windows auto-converts line endings. • “remote contains work that you do not have locally”: run
git pull origin main --allow-unrelated-histories git push -u origin main --force
• Error “File already exists” → bump version in pyproject.toml.
[tool.poetry] name = "at2weather" version = "0.1.2"
• Error “Requires-Python >=3.10,<3.11” → adjust Python version in Poetry config to match your interpreter.
Render Deployment • Build command:
pip install -r requirements.txt
• Start command:
uvicorn app.main:app --host 0.0.0.0 --port 10000
• Add at2weather inside repo so Render doesn’t miss the package.
⸻