residuals

SuperLSS.residuals(

X: object | EagerFrame,

y: NDArray,

*,

kind: Literal['pit', 'quantile'] = 'quantile',

sample_weight: NDArray | None = None,

offsets: Mapping[str, NDArray] | None = None,

seed: int = 42,

) → NDArray[float64]

Return one residual per row: "quantile" or the raw "pit" value.

This is residual_set() with everything but the array discarded. A correct family makes the PIT values uniform and the quantile residuals standard normal, so these are the values every distributional diagnostic in the suite is drawn from.