Feature request
It would be convinent to consider dataset/ directory as a consistent parquet dataset, with hive partitioning and no data duplication. Having some extra data there may cause some problems with parquet-dataset readers which are not aware of HATS, e.g. Hugging Face datasets library, see #615 for the context.
P.S. Actually HATS standard section 3.1 says that this file should be in the catalog root:
https://www.ivoa.net/documents/Notes/HATS/20250822/NOTE-hats-ivoa-1.0-20250822.html
Feature request
It would be convinent to consider
dataset/directory as a consistent parquet dataset, with hive partitioning and no data duplication. Having some extra data there may cause some problems with parquet-dataset readers which are not aware of HATS, e.g. Hugging Face datasets library, see #615 for the context.P.S. Actually HATS standard section 3.1 says that this file should be in the catalog root:
https://www.ivoa.net/documents/Notes/HATS/20250822/NOTE-hats-ivoa-1.0-20250822.html