A command to populate/update gh-pages for a dataset

[yarikoptic]

Might actually be the one creating or updating GitHub action which would update gh-pages for that dataset when pushed

[jsheunis]

Could you give an example to illustrate what you mean, since I'm having trouble understanding exactly what is suggested here.

Would this be e.g. if there is a datalad dataset that has a corresponding record in an existing catalog, and you want an updated metadata record to be generated and then rendered in the catalog if the datalad dataset itself is updated?

[yarikoptic]

target use-case: I have a dataset for which I extracted metadata (or may be even not?! that would be cool if action automatically extracted it), I push dataset to github, and GitHub action does everything necessary so that gh-pages for the dataset serves a datalad-catalog website for that dataset. create-sibling-github could gain (somehow) an option --with-gh-pages=datalad-catalog (not unlike we have create-sibling --with-ui which creates http://datasets.datalad.org/ like website) to deploy such an action upon creation of the github repo.

[jsheunis]

Thanks for clarifying, this makes a lot of sense. I have done something like this before here: https://github.com/jsheunis/automaticat-demo/blob/main/.github/workflows/generate_cat_entry.yml. This workflow checks the diff for every push to main, and if there's a new subdataset added or if any files changed, then it extracts metadata for those datasets/files, translates the records to the catalog schema, and adds it to the catalog which is contained in the gh-pages branch.

Your use case is slightly different in that it operates from a single dataset which is not assumed to be part of a super-sub-dataset nested structure. But this actually makes the process a bit simpler.

I like the idea of adding the option to create-sibling-github.