lima
LIMA Is Metadata Annotation — a small, predictable frontmatter format. A deliberate, focused subset of YAML: the part frontmatter actually needs, with well-defined types, no surprises, and zero runtime dependencies.
[]
= "0.1"
use ;
let result = parse_references.unwrap;
// LimaValue::Mapping — "title" a String, "published" an Instant
// (2024-03-01T00:00:00Z), "draft" a Bool, "tags" an Array of Strings.
parse_core/parse_references accept the raw content between the
frontmatter delimiters (---) — stripping them is the caller's job.
Two entry points are exported:
parse_core(input, strict)— the base format only (Lima Core 1.0).parse_references(input, options)— adds($key)document references and(%key)external partials (Lima References 1.0);options.partialssupplies named values,options.strictenables strict mode.
Both return Result<LimaValue, LimaError> — LimaError implements
std::error::Error and carries a stable .code field
(LimaDiagnosticCode) for programmatic error handling, alongside a
human-readable message.
Full syntax reference, the References extension, resource limits, and strict mode: docs/guide.md (examples there are TypeScript; the syntax and semantics it describes are implementation-agnostic). Rust API docs: docs.rs/lima.
Why Lima exists, the case against YAML, and security rationale: repository README.
The Lima Core 1.0
and Lima References 1.0
specifications are the normative source of truth — this crate implements
them exactly, verified against the same
250-case conformance corpus
(count pinned by tests/corpus.rs) shared with the TypeScript
implementation.
Migrating existing YAML frontmatter: docs/migrating-from-yaml.md. Static site generator integration status: docs/integrations.md.
License
ISC — see
LICENSE (also
included in this package as LICENSE).