My main question about this is whether it can be dynamically/incrementally updated within the browser? Eg new material is available or edits have been made, so sync it from backend and it gets merged in.
I've been working on using rxdb to sync and store in browser - it can use its own indexeddb abstraction, sqlite or it's own OPFS-based DB. It can also load any of these into memory in its memory-mapped mechanism. I've also made a mechanism to load everything into flexsearch in a sharedworker, so that you can do full text search fairly performantly.
It's a lot of complexity though. I'd be curious to hear any of your thoughts. Or even to chat if you're open to it!
