[BetterRed]

Assuming the annotations are in their own column, then a PI could wrangle them into separate 'sidecar' files, e.g. into TheBook - TheAuthor.annot files, and add/replace the .annot files to the calibre library as an extra 'format file' (so that they don't get stranded).

Maybe even, the sidecar create/replace could be provided as an option in the Annotations plugin - I don't use the plugin so I can't guess as to the feasibility of doing that.

Once the sidecar files are available they could be searched externally by a search tool that shows previews of the matching file contents with the search terms highlighted (Windows Search and X1 spring to mind). And, if required, the file list from the search could be could be dropped into the Drop Search Results plugiin, which will Mark the books - from which you could create/update a Reading List.

The advantage of using an external search tool, is that you could simultaneously search in annotations, and the book itself, and in email content, and in media metadata, and in web pages... The corollary is, if annotations are only available from the calibre database then from a practical perspective they only available via software which is aware of the structure of the calibre database, and they cannot be viewed alongside data that's not in the the calibre database.

At a pinch just maybe you could search the metadata OPF's with an external tool, likely to be ugly though.

BR