xarray-eopf

An xarray backend implementation for ESA EOPF data products in Zarr format.

Features

After installing this package, user can specify a new xarray backend named "eopf-zarr" to open EOPF sample products. The backend has two modes of operation, default analysis mode and the native mode. Both modes allow

• to open EOPF sample products from the local filesystem or from their original object storage using URLs with both https or s3 protocols;
• to open entire products as xarray.DataTree or xarray.Dataset;
• to open a subgroup as xarray.Dataset. This works with local filesystem or s3-URLs.

The default analysis mode has the aim to represent the EOPF data products in an analysis-ready and convenient way. It provides the following features:

• Open the deeply nested EOPF products as flat xarray.Dataset objects.
• All bands and quality images resampled to a single, user provided resolution, hence, spatial dimensions will be just x and y.
• User-specified resampling by passing spline orders for up-scaling and aggregation methods for downscaling.
• CF-compliant spatial referencing of datasets using a shared grid mapping variable spatial_ref.
• Attach other CF-compliant metadata enhancements such as flag values and meanings for pixel quality information, such as the Sentinel-2 scene classification (variable scl).

The analysis mode is currently implemented Sentinel-2 products only. Support for Sentinel-1 and Sentinel-3 is coming soon.

The native mode does not modify any contents or data, instead it basically delegates to the built-in "zarr" backend.

More information can be found in the package documentation.

Usage

The xarray-eopf package can be installed from PyPI (pip install xarray-eopf) or conda-forge (conda install -c conda-forge xarray-eopf). Now you can open EOPF sample products using xarray by specifying the "eopf-zarr" backend in your Python code:

import xarray as xr

s2_l2a_url = (
    "s3://e05ab01a9d56408d82ac32d69a5aae2a:202504-s02msil2a/15/products/cpm_v256/"
    "S2B_MSIL2A_20250415T142749_N0511_R139_T25WEV_20250415T181516.zarr"
)
s2_l2a_dataset = xr.open_dataset(s2_l2a_url, engine="eopf-zarr", resolution=10)

Development

Setting up a development environment

The recommended Python distribution for development is miniforge which includes conda, mamba, and their dependencies.

git clone https://github.com/EOPF-Sample-Service/xarray-eopf.git
cd xarray-eopf
mamba env create
mamba activate eopf-xr
pip install -ve .

Install the library locally and test

mamba activate eopf-xr
pip install -ve .
pytest

By default, this will run all unit and integration tests. To run only the unit test suite, use:

pytest tests

To run tests and generate a coverage report, use:

pytest --cov xarray_eopf --cov-report html tests

Setting up a documentation environment

mamba activate eopf-xr
pip install .[doc]

Testing documentation changes

mkdocs serve

Deploying documentation changes

mkdocs gh-deploy