healpix_convolution.pad

healpix_convolution.pad(cell_ids, *, grid_info, ring, mode='constant', constant_value=0, end_value=0, reflect_type='even')

pad an array

Parameters:

• cell_ids (array-like) – The cell ids.

• grid_info (xdggs.DGGSInfo) – The grid parameters.

• ring (int) – The pad width in rings around the input domain. Must be 0 or positive.

• mode (str, default: "constant") – The padding mode. Can be one of:

  • “constant”: fill the padded cells with constant_value.

  • “linear_ramp”: linearly interpolate the padded cells from the edge of the array to end_value. For ring 1, this is the same as mode="constant"

  • “edge”: fill the padded cells with the values at the edge of the array.

  • “reflect”: pad with the reflected values.

  • “mean”: pad with the mean of the neighbours.

  • “minimum”: pad with the minimum of the neighbours.

  • “maximum”: pad with the maximum of the neighbours.

• constant_value (scalar, default: 0) – The constant value used in constant mode.

• end_value (scalar, default: 0) – The othermost value to interpolate to. Only used in linear ramp mode.

• reflect_type ({"even", "odd"}, default: "even") – The reflect type. Only used in reflect mode.

• initial (scalar, default: 0) – The identity to use in maximum / minimum mode.

Returns:

padding_object (healpix_convolution.Padding) – The padding object. Can be used to apply the same padding operation for different arrays with the same geometry.

Warning

This assumes cell_ids is sorted, and will give wrong results if it is not.