Can the Huggingface croissant API endpoint read croissant.json metadata created by this tool?

[cboettig]

Is it possible to get the croissant metadata provided by the HuggingFace Datasets API, e.g. at enpoint https://huggingface.co/api/datasets/{USER}/{REPO}/croissant to reflect metadata from a croissant.json, such as might be generated from this tool?

I am unable to find documentation or examples for this either in this repo or on huggingface datasets documentation.
I have tried simply pushing a croissant.json file to a huggingface datasets repo, but so far the croissant metadata generated automatically by huggingface and served by its API seems only to reflect what is automatically extracted from the README. It's really great to see HuggingFace Datasets embrace Croissant metadata and make datasets discoverable in schema.org platforms like Google Dataset Search, but it quite defeats the point when that metadata does not reflect the more detailed metadata provided in a json-ld file / mlcroissant tool...