ROON isn’t a cataloging tool.
Based on the discussions I’ve had with a couple of the Roon folks (technical, sales and management) at various events, this isn’t going to change. There are myriad reasons for this … technical, practical, financial and intentional.
TIDAL (or any other digital streaming service) isn’t a cataloging tool either; there’s no reason at all for it to have any data whatsoever on music that isn’t available on the service - which is, by definition, digital streaming media. In other words, there’s no vinyl catalog/data because TIDAL doesn’t have vinyl material/versions.
The only way you’re going to get vinyl data into Roon any time in the foreseeable future, maybe ever, is to create local files that represent the music/data you want it to expose, each with the necessary meta-data embedded within it to drive Roon’s discovery and relationship features.
It’d be relatively trivial to build a tool to take an exported catalog from Discogs and for it create empty (or 1 second of silence) audio files that contain the necessary metadata to act as a proxy for your vinyl. Roon could then index those and they’d show up in your library as if they were actual music files (but not be playable -the same is true, of course, if you could simply enter data for vinyl albums directly).
I expect that’s as close as you’re going to get.
If someone wants to dig up all the necessary metadata fields, with definitions and the required ERD/map to make all of Roon’s features work for this I’d be willing to write such a tool.
That said …
While Discogs has an expansive metadata library, it isn’t curated nearly well enough (or at all, it seems) to be anything other than an intensely manual process. I looked at using it to catalog my 12,000+ vinyl album collection - starting with my local “Rolling Stone All-Time Top 100” collection…
And then, on Album #20, when faced with something like 6,000+ choices for “Thriller” (after over 3,500 choices for “London Calling” etc. etc.), realized I was either going to have to add all the metadata myself to make sure it was correct or manually sift through thousands of not-quite-right entries trying to find one that was “close enough” or was actually correct. Way too much work … and far too time consuming … unless you don’t care about accuracy and completeness, in which case I’d personally not bother in the first place.