mr2.utils.interp

mr2.utils.interp(x: Tensor, xp: Tensor, fp: Tensor, *, left: Tensor | float | None = None, right: Tensor | float | None = None) → Tensor[source]

One-dimensional linear interpolation for monotonically increasing sample points.

Implements numpy.interp

Evaluates the function at the given coordinates x based on the known points (xp, fp). Out-of-bounds values default to fp[0] and fp[-1], matching the default behavior of numpy.interp. Custom left/right fill values can be provided.

Parameters:

• x (Tensor) – The x-coordinates at which to evaluate the interpolated values.

• xp (Tensor) – 1d tensor of x coordinates of data points. The tensor will be sorted internally if needed.

• fp (Tensor) – 1d tensor of y coordinates matching the length of xp.

• left (Tensor | float | None, default: None) – Fill value for x < xp[0]. If None, uses fp[0].

• right (Tensor | float | None, default: None) – Fill value for x > xp[-1]. If None, uses fp[-1].

Returns:

The interpolated values matching the shape of x.