v0.2.0

Release v0.2.0 (2022/10/27)

Features

• Add PyTorch dataloader @jhamman (#25)
• Add TensorFlow dataloader @maxrjones (#73)
• Let TensorFlow accessor and data loader handle either xarray.DataArray or xarray.Dataset inputs @norlandrhagen (#107)
• Let torch accessor and dataloader handle either xarray.DataArray or xarray.Dataset inputs @weiji14 (#85)

Enhancement

• Lazily generate batches @tjvandal (#112)
• Use .sizes instead of .dims for xr.Dataset/xr.DataArray compatibility @weiji14 (#71)

Documentation

• Add support for community contributions @maxrjones (#79)
• Add demo notebook @tcchiao (#31)
• Add pypi badge @raybellwaves (#27)
• Rely on xarray-contrib code of conduct @maxrjones (#90)
• Use pangeo_sphinx_book_theme and mock imports @maxrjones (#62)

Maintenance

• Add GitHub workflow for PyPI releases @maxrjones (#108)
• Drop support for Python 3.7 @maxrjones (#81)
• Setup benchmarks @maxrjones (#64)
• Setup release drafter workflow @norlandrhagen (#95)
• Specify build backend and project metadata in pyproject.toml @maxrjones (#100)
• Switch LICENSE file to Apache (matches setup.py/pyproject.toml) @jhamman (#110)
• Test against expected batch in test_batch_3d_2d_input @maxrjones (#66)
• Use importlib.metadata rather than pkg_resources for version metadata @norlandrhagen (#97)
• Use informative error message for optional dependency imports @maxrjones (#118)
• Use exceptions rather than assert statements for generator @maxrjones (#63)
• 100% test coverage @jhamman (#45)