In tutorials, we currently extract metadata based on a standardized format. For guides, with only one currently available, there isn't any such standard. And further, there's some metadata types that aren't currently extracted for guides: namely keywords. This is a good opportunity to design a metadata system for guides that is straightforward for authors to use, and provides a rich set of metadata that can be reliably extracted by Librarian.
One possibility is to use standard HTML meta headers in the JupyterBook output. Another possibility is to include a JSON sidecar file with each JupyterBook that's linked via a <link> header.
Note a related issue, #14, to automatically capture package keywords. I think that an explicity metadata file could still be useful for types of metadata such as authors and scientific domain/task keywords that may be tricky to extract from the JupyterBook HTML
In tutorials, we currently extract metadata based on a standardized format. For guides, with only one currently available, there isn't any such standard. And further, there's some metadata types that aren't currently extracted for guides: namely keywords. This is a good opportunity to design a metadata system for guides that is straightforward for authors to use, and provides a rich set of metadata that can be reliably extracted by Librarian.
One possibility is to use standard HTML meta headers in the JupyterBook output. Another possibility is to include a JSON sidecar file with each JupyterBook that's linked via a
<link>
header.Note a related issue, #14, to automatically capture package keywords. I think that an explicity metadata file could still be useful for types of metadata such as authors and scientific domain/task keywords that may be tricky to extract from the JupyterBook HTML