pflm.interp.interp2d

interp2d(x: ndarray, y: ndarray, v: ndarray, x_new: ndarray, y_new: ndarray, method: str = 'linear') → ndarray

Interpolate 2D gridded data using linear or spline interpolation.

Parameters:

xnp.ndarray of shape (n_x,)

X-coordinates of the grid (first axis of v).

ynp.ndarray of shape (n_y,)

Y-coordinates of the grid (second axis of v).

vnp.ndarray of shape (n_x, n_y)

Values on the grid defined by (x, y).

x_newnp.ndarray of shape (m_x,)

Query x-coordinates.

y_newnp.ndarray of shape (m_y,)

Query y-coordinates.

method{“linear”, “spline”}, default=”linear”

Interpolation method.

Returns:

v_newnp.ndarray of shape (m_x, m_y)

Interpolated values evaluated on the mesh defined by (x_new, y_new). The dtype follows x (float32 -> f32 backend; otherwise f64).

Raises:

ValueError

If input dimensionality is invalid, arrays are empty, shape of v does not match (x, y), any input contains NaN, or method is invalid.

See also

interp1d

Interpolate 1D data using linear or spline interpolation.

Notes

• Duplicates in x and y are deduplicated using the first occurrence.

• The underlying C++ implementation expects v in Fortran-like layout with axes swapped; this wrapper transposes/contiguates as needed.

• Backend is selected by dtype of x (float32 -> f32; otherwise f64).