multitask.base.agg_predictor

multitask.base.agg_predictor(results, targets, weights)

Aggregate per-target prediction packages into a weighted forecast.

Combines future predictions, training predictions, and training actuals from per-target prediction packages into an aggregated package compatible with downstream consumers. This is a module-level convenience function; the same logic is available as BaseTask.agg_predictor.

Parameters

Name Type Description Default
results Dict[str, Dict[str, Any]] Mapping of target name to prediction package (as returned by build_prediction_package). required
targets List[str] Ordered list of target names to aggregate. required
weights List[float] Per-target aggregation weights aligned with targets. required

Returns

Name Type Description
Dict[str, Any] Aggregated prediction package with keys train_actual,
Dict[str, Any] train_pred, future_pred, future_actual,
Dict[str, Any] metrics_train, metrics_future, metrics_future_one_day,
Dict[str, Any] validation_passed, and (when present in all sources)
Dict[str, Any] test_actual.

Examples

import numpy as np
import pandas as pd
from spotforecast2_safe.multitask.base import agg_predictor

rng = np.random.default_rng(0)
idx_train = pd.date_range("2023-01-01", periods=100, freq="h", tz="UTC")
idx_future = pd.date_range("2023-01-05 04:00", periods=6, freq="h", tz="UTC")

def _pkg(train_val, future_val):
    return {
        "train_actual": pd.Series(np.full(100, train_val), index=idx_train),
        "train_pred": pd.Series(np.full(100, train_val * 0.99), index=idx_train),
        "future_pred": pd.Series(np.full(6, future_val), index=idx_future),
        "future_actual": pd.Series(dtype="float64"),
    }

results = {"wind": _pkg(100.0, 110.0), "solar": _pkg(200.0, 210.0)}
agg = agg_predictor(results, targets=["wind", "solar"], weights=[0.5, 0.5])
print(f"future_pred (weighted mean): {agg['future_pred'].iloc[0]:.1f}")
assert set(agg.keys()) >= {"train_actual", "train_pred", "future_pred", "validation_passed"}
future_pred (weighted mean): 160.0