// the find
corteva/geocube
Tool to convert geopandas vector data into rasterized xarray data.
geocube takes geopandas vector data (polygons, lines, points with attributes) and rasterizes it into an xarray Dataset/DataArray, so you can mix vector GIS data into a raster/xarray pipeline. It's for people already in the rasterio/xarray geospatial stack who need vector-to-raster conversion without hand-rolling rasterio.features.rasterize calls themselves.
Test suite includes precomputed .nc comparison files for exact numerical regression checks on rasterization output, not just shape/type assertions. Ships py.typed plus mypy/black/isort/flake8 via pre-commit, which is more engineering discipline than most single-purpose geospatial libraries bother with. Packaged for both PyPI and conda-forge with CI on GitHub Actions and AppVeyor, which matters a lot here since GDAL/fiona/rasterio installs are where this stack usually breaks. Has a real CLI (geocube/cli) in addition to the Python API, so it's usable from a shell pipeline without writing a script.
The README is just badges and links to an external docs site — no code example, no quickstart, you have to leave GitHub to find out what the API looks like. Support is spread across Gitter, GitHub Discussions, and a GIS StackExchange tag, and Gitter is effectively dead across most projects at this point so that link is a dead end in practice. CI badges point to one primary maintainer (snowman2), typical bus-factor risk for a dependency this specific. The core value depends entirely on GDAL/fiona being installed correctly, and nothing in the README addresses that install pain (no conda-first instructions, no container image), which is the actual first blocker for new users.