rug

An experimental, prose-first markup language for thoughtlessly writing structured documents — inspired by pug-js and literate programming.

You write plain text. When you need structure, a line beginning with a period drops you into a component, whose body is parsed as a small, predictable, YAML-like format. That's the whole idea — a leading period is rare enough in ordinary prose to be a clean signal.

This is just prose, with newlines and tabs preserved.

.button label: Sign up, href: /join

# a comment

And here the text resumes normally.

Usage

import parseRug from 'rug-lang'

parseRug('.card\n  title: Hello\n  count: 3')
// [
//   {
//     type: 'component',
//     name: 'card',
//     properties: { title: 'Hello', count: 3 },
//     start: 0, end: 28,
//     nameRange: [0, 5], bodyRange: [6, 28]
//   }
// ]

parseRug(input) returns a flat array of blocks — prose, component, or comment — each carrying source offsets for syntax highlighting and tooling.

• docs/syntax.md — the friendly language guide
• docs/spec.md — the formal specification (grammar, data model, processing model, API contract) for embedding rug elsewhere
• src/index.d.ts — TypeScript definitions

Develop

npm test          # run the test suite
npm run watch     # re-run scratch.js on change

Source is split into focused modules:

• src/lines.js — split input into offset-preserving lines and classify them
• src/scalar.js — scalar coercion
• src/flow.js — the inline/flow value parser
• src/structure.js — indented block bodies
• src/index.js — ties it together

Not ready for prime-time.

MIT