panproto-protocols

51 built-in schema language definitions, each describing the rules for one data or API format.

What it does

A protocol in panproto is a definition of what a valid schema looks like for a particular format, plus a parser that reads that format's native files into panproto's internal schema representation. For example, the ATProto protocol knows that Lexicon files are JSON objects with defs blocks, what kinds of fields are allowed, and how those fields relate to each other. The OpenAPI protocol knows about paths, operations, and components.

Each protocol is assembled from smaller building-block theories by combining them with colimit (a way of merging theories that shares their common parts rather than duplicating them). For instance, a protocol that uses ordered collections includes ThOrder; one that needs validation rules includes ThConstraint. Building blocks are reused across protocols so the definitions stay compact and consistent.

Programming language parsing (TypeScript, Python, Rust, and so on) is handled separately by panproto-parse using tree-sitter grammars. The protocols here are semantic formats: annotation schemas, API specifications, database schemas, domain-specific formats, and serialization IDLs.

Quick example

use panproto_protocols::atproto;

// Get the ATProto protocol definition.
let protocol = atproto::protocol();

// Parse a Lexicon file into a panproto Schema.
let schema = atproto::parse_lexicon(&lexicon_json_bytes)?;

API overview

Building-block theories

Each protocol's schema and instance theories are composed from five reusable pieces:

License

MIT