eotransform-xarray

What can I use eotransform-xarray for?

The eotransform-xarray package provides common transformations on raster data represented as xarray data structures, following the Transformer protocol of eotransform. This makes them easy to mix and match, and you can quickly chain processing pipelines, using other eotransform protocols. Additionally, processing pipelines constructed from these Transformers, can be automatically applied to the streamed_process function from eotransform, to benefit from I/O hiding.

Getting Started

Installation

pip install eotransform-xarray

Example: streamed processing pipeline

In the following example swath data is resampled, masked and written out as a GeoTIFF stack.

resample = ResampleWithGauss(swath_geometry, raster_geometry, sigma=2e5, neighbours=4, lookup_radius=1e6)
mask = MaskWhere(lambda x: x > 2, np.nan)
squeeze = Squeeze()
with ThreadPoolExecutor(max_workers=3) as ex:
    pipeline = Compose([resample, mask, squeeze])
    streamed_process(input_src, pipeline, SinkToGeoTiff(dst_dir, lambda i, da: f"out_{i}.tif"), ex)

^{snippet source | anchor}

Note, that this example uses eotransform's streamed_process function to hide the I/O operations, using the compute resources more effectively.

Support & Documentation

Dependencies:

eotransform-xarray requires Python 3.8 and has these dependencies:

eotransform>=1.8
xarray
rioxarray
numpy
affine
more_itertools

^{snippet source | anchor}

Citation

If you find this repository useful, please consider giving it a star or a citation:

@software{eotransform_xarray_2023_8002854,
  author       = {Raml, Bernhard},
  title        = {eotransform-xarray},
  month        = jun,
  year         = 2023,
  publisher    = {Zenodo},
  version      = {2.4.1},
  doi          = {10.5281/zenodo.8002854},
  url          = {https://doi.org/10.5281/zenodo.8002854}
}