Found a bug in handling relative dataset paths resulting in a `ValueError` saying path elements are missing `=` signs.
0.3.0
A complete new approach dropping pyarrow and implementing Hive flavor partitioning completely in Python.
0.2.5
Fixes read_partition in another way to support even more schemes
0.2.4
Bugfix the read_partition implementation for remote files where scheme is required in the url.
0.2.3
We open up the dependency range to allow for a polars upgrade for dependent packages. We also open up pyarrow, but plan on removing the pyarrow dependency for version 2.0 to allow for schema upgrades in datasets.