layers.pub now has an official data PDS at repo.layers.pub as well as a python client and CLI that you can read about at layers.pub/lairs/!
Post
The layers repo currently contains all license-compatible UniMorph lexicons (unimorph.github.io), Universal Dependencies treebanks (universaldependencies.org), and PropBank annotations (propbank.github.io) pointing into those treebanks, alongside additional lexical resources.
Additionally, repo.megaattitude.io and repo.decomp.io host the MegaAttitude and Universal Decompositional Semantics datasets, the latter of which has pointers into the UD treebanks living on repo.layers.pub.
On the writing side: you can author records as typed models and lairs applies them to your PDS as a minimal diff. It also has codecs for converting between Layers and formats like CoNLL-U and brat, and exporters to HuggingFace datasets, PyTorch, and WebDataset.
0 reposts0 likes
lairs is built on didactic (panproto.dev/didactic/), which provides pydantic-like typed models for the pub.layers.* lexicons.
didactic is itself built on panproto.dev, which handles migration within and between formats (e.g. between layers data model versions or from CoNLL-U to layers).
If you find this stuff interesting, also check out the broader @atproto.science ecosystem!
We made a thing! Check out our new #ATScience ecosystem map ✨ All projects, organized in one place: atproto.science/ecosystem/