BaseMetric

chronax.metrics.base_metric.BaseMetric

Base class for Chronax metrics, providing standardized input validation and preprocessing of predictions (y_pred) based on the model type (deterministic, stochastic, or hybrid) before delegating the final computation to subclass implementations.

__init__(self, metric_type: str)

Parameter	Type	Default	Description
metric_type	str	-	(undocumented)

__call__(self, y_true: np.ndarray, y_pred: np.ndarray, **kwargs)

Process y_pred based on model type and delegate to _compute method. This method ensures all subclasses automatically get preprocessed y_pred without needing to manually call process_y_pred.

Parameter	Type	Default	Description
y_true	np.ndarray	-	(undocumented)
y_pred	np.ndarray	-	(undocumented)
**kwargs		-	(undocumented)

Returns: The computed metric value. Raises: ValueError (If 'model_type' must be provided in kwargs).

process_y_pred(self, y_true: np.ndarray, y_pred: np.ndarray, model_type: str, point_forecast_statistic: str \| None = None) -> np.ndarray

Process y_pred based on model type, applying necessary transformations.

Parameter	Type	Default	Description
y_true	np.ndarray	-	(undocumented)
y_pred	np.ndarray	-	(undocumented)
model_type	str	-	(undocumented)
point_forecast_statistic	str | None	None	(undocumented)

Returns: np.ndarray (The processed forecast array).