eoio.processors.resample_wavelength.processor module

eoio.processors.resample_wavelength.processor

Top-level wavelength resampling processor.

Resamples spectral variables, each on its own wavelength dimension (wavelength_olci, wavelength_floris, … as produced by the FLEX reader), onto one common wavelength grid so they can be combined or compared.

User config example

processors= {

“resample_wavelength”: {

“target_wavelength”: {“min”: 500, “max”: 780, “step”: 1}, “data_vars”: [“floris_apparent_reflectance”, “olci_apparent_reflectance”], “method”: “linear”,

},

}

class eoio.processors.resample_wavelength.processor.ResampleWavelength(params: Dict[str, Any] | None = None, context: Dict[str, Any] | None = None)

Bases: BaseProcessor

Resample spectral variables onto a common wavelength grid.

Each selected variable is interpolated along its own wavelength_* dimension onto the target wavelengths, giving variables that share one wavelength dimension. Target wavelengths outside the range of a variable’s source grid are set to NaN (no extrapolation).

Obsarray uncertainty components (<var>_uncertainty) of resampled variables are resampled with them and re-registered against the resampled variable, keeping their error-correlation definition with the wavelength dimension renamed.

Processor parameters

The following parameters can be provided in the params dict:

param target_wavelength:

Required. Target wavelengths in nm, either a list/array, or a dict {"min", "max", "step"} (max inclusive).

param data_vars:

Optional list of variables to resample. If omitted, all variables with a wavelength_* dimension are resampled.

param method:

Interpolation method ("linear", "nearest", "cubic", …), as for xarray.interp. Default "linear".

param dim_name:

Name of the new common wavelength dimension. Default "wavelength".

param inplace:

If False (default) resampled variables are added with the suffix _resampled (e.g. floris_apparent_reflectance_resampled). If True they replace the originals.

param drop_original:

If True, drop the originals (and their now-unneeded wavelength coordinates) after resampling. No-op when inplace is True. Default False.

param on_missing:

Behaviour if a requested variable is missing or has no wavelength dimension: "error" (default) or "skip".

Notes

• This processor is intended to run after reading.

name = 'resample_wavelength'

run(ds: Dataset) → Dataset

Run the processor on the dataset.

Parameters:

ds – Input dataset.

Returns:

Dataset with spectral variables resampled onto the common wavelength grid.

class eoio.processors.resample_wavelength.processor.ResampleWavelengthConfig(target_wavelength: ndarray, data_vars: Sequence[str] | None = None, method: str = 'linear', dim_name: str = 'wavelength', inplace: bool = False, drop_original: bool = False, on_missing: str = 'error')

Bases: object

Parameters for the resample_wavelength processor.

data_vars: Sequence[str] | None = None

dim_name: str = 'wavelength'

drop_original: bool = False

inplace: bool = False

method: str = 'linear'

on_missing: str = 'error'

target_wavelength: ndarray

eoio.processors.resample_wavelength.processor.WAVELENGTH_DIM_PREFIX = 'wavelength_'

prefix identifying a source wavelength dimension