spectrochempy.interpolate

interpolate(dataset, dim=None, dims=None, coord=None, method='linear', fill_value=np.nan, assume_sorted=False, inplace=False)[source]

Interpolate dataset onto new coordinates.

Parameters:

  • dataset (NDDataset) – Dataset to interpolate.

  • dim (str or int, optional) – Single dimension name or index.

  • dims (list of str or int, optional) – List of dimensions to interpolate. Higher priority than dim.

  • coord (Coord, np.ndarray, NDDataset, or dict, optional) – Target coordinates. If dict, maps dim -> coord. For single dim: can be Coord, ndarray, or NDDataset. For multiple dims: dict {dim: coord} or sequential application.

  • method (str, optional, default=’linear’) – Interpolation method: ‘linear’ or ‘pchip’.

  • fill_value (any, optional, default=np.nan) – Value for points outside the original range.

  • assume_sorted (bool, optional, default=False) – If True, skip monotonicity checks.

  • inplace (bool, optional, default=False) – If True, modify the dataset in place.

Returns:

NDDataset – Dataset with interpolated data and updated coordinates.

Raises:

  • TypeError – If coordinate is not numeric.

  • ValueError – If coordinates are not strictly monotonic.

Notes

  • Labels are NOT interpolated and are reset after interpolation.

  • For multiple coordinates per dimension, all are interpolated consistently.

  • Secondary coordinates are interpolated numerically. If they represent analytical transformations of the primary coordinate (e.g., wavelength = 1/wavenumber), the result may be approximate - consider recomputing them analytically after interpolation.

  • Unit conversion is performed if needed before interpolation.

  • Sequential interpolation is applied for multiple dimensions (not true N-D).