ForecastInputDataset#

class openstef_core.datasets.ForecastInputDataset(data: DataFrame, sample_interval: timedelta = timedelta(minutes=15), forecast_start: datetime | None = None, *, horizon_column: str = 'horizon', available_at_column: str = 'available_at', sample_weight_column: str = 'sample_weight', target_column: str = 'load') → None[source]#

Bases: TimeSeriesDataset

Time series dataset for forecasting with validated target column.

Used for training and prediction data where a specific target column must exist. The target column represents the value being forecasted.

Invariants

Target column must exist in the dataset
Inherits all TimeSeriesDataset guarantees (sorted timestamps, consistent intervals)

Attrs:: target_column: Name of the target column to forecast. sample_weight_column: Name of the column containing sample weights. forecast_start: Optional timestamp indicating when the forecast period starts.

Example

>>> import pandas as pd
>>> from datetime import timedelta
>>> data = pd.DataFrame({
...     'load': [100, 120, 110],
...     'temperature': [20, 22, 21],
...     'weights': [1.0, 0.5, 1.0],
... }, index=pd.date_range('2025-01-01', periods=3, freq='h'))
>>> dataset = ForecastInputDataset(
...     data=data,
...     sample_interval=timedelta(hours=1),
...     target_column='load',
...     sample_weight_column='weights',
... )
>>> dataset.target_column
'load'
>>> dataset.sample_weight_column
'weights'
>>> len(dataset.target_series)
3
>>> len(dataset.sample_weight_series)
3