For example, I can see:
- https://ufos.microcosm.blue/collection/?nsid=sh.tangled.repo
- https://ufos.microcosm.blue/collection/?nsid=sh.tangled.publ...
- https://ufos.microcosm.blue/collection/?nsid=sh.tangled.repo...
Here's a link to all of them: https://ufos.microcosm.blue/collection/?prefix=sh.tangled
Is it meant to be some kind of filtered sampling from the stream of ATProto events?
Think of these as names of tables (or collections) in a distributed database. Or as type definitions. Or as app-defined data formats.
Each lexicon is a schema for a model. So you’re looking at a list of such “types” — a “repo”, a “follow”, a “star” etc.
There’s a “Tangled repo” lexicon, a “Bluesky post” lexicon, a “Leaflet publication” lexicon, and so on. Lexicons are specified and evolved by developers of each concrete application. Other apps can use those type definitions to read or write that kind of data.
See https://www.pfrazee.com/blog/why-not-rdf#lexicon for a short intro.
The UFO tool samples the global event stream and keeps stats on which lexicons are showing up in it (i.e. types of JSON that are being created on the network). You can expand the “samples” tab to show a few concrete JSON blobs so you get the idea of what the data represents.