I can answer the "view senses in Wikidata" question, and will let others answer your more broader questions of how Abstract Wikipedia might from a high-level to automatically generate content in other languages based on Lexicographical data in Wikidata.
So...what you just showed in your example is Text format of a general view of how you would like to visualize or view relations between Lexeme and edit or enter new data against those various relations.
This is exactly what we need to show and improve in our Documentation much better visually, and using the Wikidata properties in those visualizations, so that folks know how to wire things up like that manually.
BUT...instead of just in Documentation... we could improve the Lexeme page views themselves!
It's encouraging that there are tools coming online in the Wikidata ecosystem that are improving things so that users might have nice forms with placeholder property elements (not the Wikidata Lexeme pages which lack placeholders for your use case) to easily just fill out. For example, already having Sense, Item for Sense, Synonym, etc. etc. as placeholders on the form ready to fill out for a Lexeme.
In the
SVG for Lexeme Data Model we do a great job showing what is possible. We just need better placeholder forms
in those tools, and I'd say instead of a Tool, a new Gadget or two that can automatically expose those placeholder properties for easier manual entry based on a user's preference of what parts of the Lexeme data model they typically want to work on. Adding synonyms, senses, or sets of properties often used together, etc. as in your use case.
In other words, in the Gadget, actually allowing users to save their setup of the custom Lexeme page views for data entry as they need, which can show what is possible based on the Lexeme data model.
That's what is missing.
Otherwise, everyone has to remember what the Properties #'s are, and constantly retype them into little boxes over and over. (Carpal Tunnel Syndrome comes up rapidly)
Much better to be able to enable a Gadget that has different "modes" of operation, with a few pre-saved templates of Lexeme page views.
