2023-11-09

Tidied up the factor description and attributes.

2023-10-17

Added Python restoration commands for some of the HDF5 formats.

2023-10-05

Defer handling of missingness to (future) v1 of the chihaya spec.

2023-10-03

Shifted version numbers into the scoped properties.

This allows us to have multiple versions corresponding to base/subclasses,
especially if they evolve at different rates.

2023-09-28

Avoid redundant runs of the Action on a PR.

2023-05-07

Commented on the need for explicit format specifications.

In the GFF case, it's mostly to avoid guesswork about a missing comment
line. In the BED case, it's a bit more necessary to distinguish between
the different BED variants.

2023-05-03

• Support typical bioinformatics file formats like BAM, BED, etc.
• Removed confusing alternative sparse matrix formats.

2023-04-14

Added date-time as a supported dataframe column type.

2023-02-20

Mandate unique and non-empty names in various fields.

This just makes life easier for downstream consumers, which can now reliably
index on the name without worrying about ambiguity. We also guarantee
non-overlap between the main and alternative experiment names.

2022-10-18

Restored the schema for JSON-based simple lists.