panproto-protocols

Protocol definitions and schema parsers for semantic data formats.

Coverage

The crate contains 54 format-specific schema parsers exposed through two generic dispatch functions. parse_schema_document accepts the 43 JSON-document protocols. parse_schema_source accepts the 11 text-source protocols. The latter group includes SQL DDL, GraphQL SDL, Protobuf, CDDL, CQL, Cypher, Redis schema text, ASN.1, Bond, FlatBuffers, and CoNLL-U. The raw_file module is a separate fallback for project assembly.

The protocol modules cover annotation formats, API descriptions, configuration, data-schema formats, data-science schemas, databases, domain formats, serialization IDLs, and web or document formats. Programming-language grammars are provided by panproto-parse, not this crate.

Only ATProto currently has generic bundle dispatch through parse_schema_bundle and parse_schema_bundle_project. Other protocols must be parsed one document at a time unless their module exposes a separate specialized API.

Theories

Each Protocol names a schema theory and an instance theory. theories provides 11 reusable constructors: ThGraph, ThConstraint, ThMulti, ThWType, ThMeta, ThSimpleGraph, ThHypergraph, ThInterface, ThFunctor, ThFlat, and ThGraphInstance. Individual protocols may register additional composed theories.

The generalized-algebraic-theory terminology follows Cartmell (1986). The use of colimits to combine theory presentations follows the structured-specification line begun by Burstall and Goguen (1977).

Example

use panproto_protocols::atproto;

let protocol = atproto::protocol();
let document: serde_json::Value = serde_json::from_slice(&lexicon_bytes)?;
let schema = atproto::parse_lexicon(&document)?;

Main entry points

Each format module also exposes its own protocol(), parser, emitter, and theory registration functions. See docs.rs for those module-specific signatures.

License

MIT