Transformations

Spectral indices provide transformations to create data processing pipelines on remote sensing images.

class spectral_indices.transformations.Aggregation(dim: str = 'time', skipna: bool = False)

Base class for aggregations along a dimension (sum, mean etc..).

Parameters:

• dim (str) – Dimensions to perform aggregation on. Default to ‘time’.

• skipna (bool) – To perform aggregation with nan or not. Default to False.

class spectral_indices.transformations.ApplyMasking(maskings: ~typing.List[~spectral_indices.transformations.mask.Masking | ~typing.Dict[str, ~typing.Any]] = <factory>)

Wraps multiple masking operations.

planify(array: DataArray) → DataArray

Apply masking to data.

Parameters:

array (DataArray) – Array to mask that contain band masks.

Returns:

• Lazy masked array without band masks.

Return type:

DataArray

class spectral_indices.transformations.Masking(logical_string: str = <factory>, mask: ~spectral_indices.sources.masks.Mask = <factory>)

Wrap masking metadata.

Parameters:

• logical_string (str) – A string that represent the logical to apply within a xarray.where() method. Use ‘mask’ to represent the mask array. i.e: “mask > 5”.

• mask – (Mask): Corresponding Mask class.

class spectral_indices.transformations.Max(dim: str = 'time', skipna: bool = False)

planify(array: DataArray)

Lazy transformation on xarray. This function plan some modifications that are added to dask graph.

class spectral_indices.transformations.Mean(dim: str = 'time', skipna: bool = False)

planify(array: DataArray)

Lazy transformation on xarray. This function plan some modifications that are added to dask graph.

class spectral_indices.transformations.Min(dim: str = 'time', skipna: bool = False)

planify(array: DataArray)

Lazy transformation on xarray. This function plan some modifications that are added to dask graph.

class spectral_indices.transformations.SaveRaster(out_path: str, nodata: float = -1000)

Save xarray to a tiff file supporting dask chunk parralle save.

Parameters:

• out_path (str) – Path to save raster.

• nodata (float) – Value to fill nodata. Default to -1000

planify(array: DataArray, lock: Any) → DataArray

Planify the array saving on a pipeline.

class spectral_indices.transformations.Sum(dim: str = 'time', skipna: bool = False)

planify(array: DataArray)

Lazy transformation on xarray. This function plan some modifications that are added to dask graph.

class spectral_indices.transformations.TemporalInterpolation(method: str = 'linear', gap: timedelta | timedelta64 | Timedelta = None)

Perform temporal interpolation on xarray.

Parameters:

• method (str) – Method to perform interpolation. See https://docs.xarray.dev/en/stable/generated/xarray.DataArray.interpolate_na.html to see all available options. Default to ‘linear’.

• gap (Union[timedelta, timedelta64, Timedelta) – Maximum length of nan gap to fill. Default to None (no limit).

planify(array: DataArray) → DataArray

Lazy transformation on xarray. This function plan some modifications that are added to dask graph.

class spectral_indices.transformations.Transformation

Base class for transformations on xarrays.

classmethod from_dict(transformation_dict: Dict[str, Any]) → Transformation

Build transformation from attributes dict.

Parameters:

trasnformation_dict (Dict[str, Any]) – Dict of trasnformation attributes.

Returns:

• Return an instance of transformation.

Return type:

Transformation

property name: str

Return class name.

planify(array: DataArray) → DataArray

Lazy transformation on xarray. This function plan some modifications that are added to dask graph.

to_dict() → Dict[str, Any]

Return attributes as dict.