Hi folks, I started a project to map out roms on the many dat sources (no-intro), to external sites such as tgdb, igdb, giantbomb.
I made it all available here: https://github.com/viniciusccarvalho/retro-database
First pass is a way to convert most of libretro dats into csv files that anyone could import.
Second is an attempt to match those games to corresponding entries on remote sites. The first take is not too bad, besides arcade who has a lot of false matches, most systems would be over 75% of exact/similar match.
I feel with a little work, this could be a good shared database to people to implement their own scrappers that would not require remote calls or be throttled by request limits.
If anyone is interested in collaborating in curating this, I will eventually publish the code that automatically parses all files and remote sites and should generate new versions of the db on every new dat changes.