Load pre-trained models from disk

Predict ag commodity futures price movement Args: weather_data: Current weather conditions and forecast

Run NDVI analysis for all 6 commodity regions. Called in main monitoring cycle.

Calculate correlations conditional on volatility regime Args: returns1: First returns series returns

API root endpoint

Fetch all available supply reports

Run historical backtest Test prediction accuracy on past data

Full monitoring cycle across all commodities. Runs every 15 minutes in production. Returns aggregated signals for all active

Seed database with historical freeze events for testing

Store daily market price data

Store daily performance metric

Store USDA crop report

Train all models in the ensemble Args: X_train: Training features y_train: Training targets (price c

Update grove health from satellite imagery