Enable use of Zarr datasets in wkcuber

[normanrz]

The PR #627 brings initial Zarr support to the Dataset API. However, it would be cool to also be able to use the wkcuber utilities to read/write Zarr datasets. Since the wkcuber largely builds upon the functionality of the Dataset API, in many cases Zarr support should be as easy as adding a --data_format zarr CLI arg. In other cases, e.g. image conversion or metadata guessing, there may be some challenges.
Additionally, I think the new chunk_size and chunks_per_shard parameters should also be exposed via the wkcuber CLI args.