cbor-ld

CBOR-LD 1.0 processor for Rust, built on top of cbor2.

cbor-ld converts JSON-LD documents to and from the current W3C CBOR-LD 1.0 wire shape:

51997([registryEntryId, payload])

Registry entry ID 0 stores an uncompressed payload. Other registry IDs use the default CBOR-LD semantic compression model: JSON-LD context processing, dynamic term IDs, registry type tables, and value codecs for URLs, multibase values, xsd:date, and xsd:dateTime.

The crate intentionally targets modern CBOR-LD 1.0. Older legacy singleton and range tags from pre-1.0 JavaScript releases are not implemented.

Why cbor-ld

Licensed under the MIT License.

Quick Start

[dependencies]
cbor-ld = "0.1"

When developing this repository alongside cbor2, use the local path dependency:

[dependencies]
cbor-ld = { path = "../cbor-ld" }
cbor2 = { path = "../cbor2", version = "1" }

Encode and decode an uncompressed CBOR-LD document:

use cbor2::Value;
use cbor_ld::{decode, encode, DecodeOptions, EncodeOptions};

let document = Value::Map(vec![(
    Value::Text("hello".into()),
    Value::Text("world".into()),
)]);

let bytes = encode(&document, EncodeOptions::uncompressed())?;
let decoded = decode(&bytes, DecodeOptions::default())?;

assert_eq!(decoded, document);
# Ok::<(), cbor_ld::Error>(())

Use compression with inline contexts:

use cbor2::Value;
use cbor_ld::{decode, encode, DecodeOptions, EncodeOptions, TypeTable};

let document = Value::Map(vec![
    (
        Value::Text("@context".into()),
        Value::Map(vec![(
            Value::Text("name".into()),
            Value::Text("https://schema.org/name".into()),
        )]),
    ),
    (Value::Text("name".into()), Value::Text("Alice".into())),
]);

let table = TypeTable::new();
let bytes = encode(&document, EncodeOptions::compressed(1, &table))?;
let decoded = decode(&bytes, DecodeOptions::default())?;

assert_eq!(
    cbor2::to_canonical_vec(&decoded)?,
    cbor2::to_canonical_vec(&document)?
);
# Ok::<(), Box<dyn std::error::Error>>(())

Use encode_with_loader and decode_with_loader when a document contains remote context URLs. The loader receives the URL and returns the loaded JSON-LD document as a cbor2::Value with an @context entry.

Type Tables

TypeTable maps original values to compressed integer IDs. Every table is normalized before use so the core context, url, and none subtables are present.

use cbor_ld::TypeTable;

let mut table = TypeTable::new();
table.insert("context", "https://example.com/context/v1", 0x8000);
table.insert("https://example.com/type#status", "active", 0x8001);

TypeTable::with_common_tables() includes the common string and cryptosuite tables mirrored from the JavaScript reference implementation.

cbor2 Integration

The implementation uses cbor2 directly rather than building a parallel CBOR layer:

• Value::Tag(0xcb1d, ...) represents the CBOR-LD envelope.
• Value::Map stores compressed term maps with integer keys.
• Value::Bytes stores registry-table byte encodings and binary URL/multibase codec outputs.
• to_canonical_vec is the default encoder for deterministic protocol bytes.
• validate checks that input is exactly one well-formed CBOR item before decode.

This keeps CBOR-LD-specific logic focused on JSON-LD semantics while reusing cbor2 for RFC 8949 encoding, tags, validation, canonical ordering, and the dynamic value model.

Current Scope

Implemented:

• CBOR-LD 1.0 tag 0xcb1d.
• Registry entry ID 0 uncompressed payloads.
• Default semantic compression for nonzero registry entry IDs.
• Embedded, remote, imported, type-scoped, and property-scoped JSON-LD contexts.
• URL, UUID URN, data URL, DID key/nym, multibase, xsd:date, and xsd:dateTime value codecs.

Not implemented:

• Pre-1.0 legacy singleton tags 0x0500..0x0501.
• Pre-1.0 legacy range tags 0x0600..0x06ff.

Verification

Useful checks while developing:

cargo fmt --all --check
cargo test
cargo clippy --all-targets -- -D warnings
git diff --check