Version 1.4.0

Changelog:

• (codebreaking) Complete overhaul of eolearn.coregistration. See documentation for details.
• (codebreaking) Removed non-working HVPlot backend for eolearn.visualization.
• (codebreaking) The SpatialResizeTask had a bug when resizing w.r.t resolution. The issue was fixed and the signature of the task was redesigned to better avoid mistakes. See documentation for details.
• (codebreaking) The EOPatch methods get_features and get_feature_list were recombined into a new get_features method. The method get_time_series was removed. See documentation for details.
• (codebreaking) Removed unsound use_int_coords option in eolearn.ml_tools.sampling.random_point_in_triangle.
• Added ability to specify query in execute method of MeteoblueTask.
• SentinelHubInputTask no longer saves redundant data into meta-features.
• Module eolearn.core.utils.types was moved to eolearn.core.types. Old one will be removed in the future.
• Switched opencv-contrib-python-headless requirement to opencv-python-headless
• Added type annotations to most of the code base.
• Various improvements to tests and code.

Version 1.3.1

Changelog:

• Sentinel Hub IO tasks now support a custom timestamp filtration via timestamp_filter parameter, contributed by @ColinMoldenhauer.
• MergeFeatureTask now supports the axis parameter.
• Fix minor issues with the coregistration module.
• Prepare for future removal of sentinelhub.os_utils.
• Fix type annotations after mypy update.
• Improvements to tests and various minor changes.

Version 1.3.0

Changelog:

• (codebreaking) Adapted Sentinel Hub tasks to sentinelhub-py 3.8.0 which switched to Catalog 1.0.0.
• (codebreaking) Removed support for loading pickled objects in EOPatches (deprecated since version 1.0.0).
• (codebreaking) Various improvements of FeatureIO class. Only affects direct use of class.
• Added type annotations to majority of eolearn.core. The types are now exposed via py.typed file, which enables use of mypy. Added type-checking to CI for the core module.
• Numpy-array based features can now save and load object populated arrays.
• Improved documentation building, fixed links to GitHub.
• Improved test coverage.
• Added pre-commit hooks to repository for easier development.
• Various minor improvements.

Version 1.2.1

Changelog:

• Corrected the default for no_data_value in ImportFromTiffTask and ExportToTiffTask to None. The previous default of 0 was a poor choice in many scenarios. The switch might alter behavior in existing code.
• Changed the way SpatialResizeTask accepts parameters for the final image size. Now supports resizing by using resolution.
• Added ExplodeBandsTask that explodes a multi-band feature into multiple features.
• Exposed resampling parameters in Sentinel Hub tasks and included a geometry execution parameter.
• Reworked internal classes FeatureIO and _FeatureDict to improve types and maintainability.
• Fixed y-axis orientation of MeteoblueRasterTask.
• FilterTimeSeriesTask adjusted to work with multiprocessing.
• EOPatch plotting no longer anti-aliases by default (removes issues with phantom values in mask plots)
• Improved documentation building, fixing a few broken links.

Version 1.2.0

Changelog:

• Improved handling of filesystem objects:
  • introduced utility functions pickle_fs and unpickle_fs into eo-learn-core,
  • updated all IO tasks to fully support filesystem as an __init__ parameter,
  • updated EOExecutor to support filesystem propagation to worker processes.
• Official support for Python 3.10.
• Moved eolearn.coregistration.ThunderRegistrationTask into an extension of eo-learn-coregistration package because it doesn't support Python 3.10.
• Updated functionality of eolearn.features.SimpleFilterTask. It now returns a new EOPatch and doesn't raise an error if all time slices would be filtered out.
• Larger updates of eolearn.io.ExportToTiffTask. It doesn't allow undefined folder parameter anymore, but it has better support for filesystem objects.
• Added eolearn.core.utils.raster.fast_nanpercentile utility function and moved constant_pad function into the same module.
• Suppressed a warning when saving an EOPatch with empty vector features.
• Updated code-style checkers by including flake8 and checks for Jupyter notebooks.
• Various style improvements in example notebooks, code, and tests.

Version 1.1.1

Changelog:

• Fixed a bug in eolearn.io.ImportFromTiffTask where a bounding box from the image wasn't added to the EOPatch.
• Increased minimal version of Pillow dependency in eolearn.features.

Version 1.1.0

Changelog:

• Large improvements of parallelization in EOExecutor. Introduced the eolearn.core.utils.parallelize module, featuring tools for different parallelization modes.
• Added support for session sharing in SentinelHubInputTask, SentinelHubEvalscriptTask and SentinelHubDemTask by adding a session_loader parameter. Session sharing of sentinelhub-py is explained here.
• Added SpatialResizeTask to eolearn.features.feature_manipulation for spatially resizing EOPatch features.
• Improved how ImportFromTiffTask reads from remote filesystems.
• Switched to non-structural hashing of EONode class to avoid massive slowdowns in large workflows.
• Improved procedure for building documentation and displaying of type annotations.
• Various minor improvements.

Version 1.0.2

Changelog:

• Added workaround for an issue introduced by fs==2.4.16.
• Executor progress bar improved for use-cases with many EOPatches.
• LoadTask and SaveTask can now handle empty queries (by setting eopatch_folder=None).
• Minor improvements in code and documentation.

Version 1.0.1

Changelog:

• Fixed an issue where vector features with empty dataframes couldn't be saved to a Geopackage.
• Memory improvement in EOPatch merging procedure.
• Added support for aws_session_token, contributed by @theirix.
• Fixed an issue in ImportFromTiffTask.
• Fixed a packaging issue where some new subpackage extensions didn't work in the version 1.0.0 that was released to PyPI.
• eo-learn abstract package from now on requires fixed versions of eo-learn subpackages.
• Applied isort formatting on the entire package.
• Minor improvements in code and documentation.

Version 0.10.2

This version fixes the dependencies, so that pip install "eo-learn<1" works correctly.