Grammar reference

The grammar language is identical across every target — you write a .lark grammar once and parse with it from Python, Rust, TypeScript, or C.

Rules and terminals

• Rules are lowercase (expr, start) and produce tree nodes.
• Terminals are UPPERCASE (NUMBER, NAME) and produce tokens.
• %ignore drops a terminal (whitespace, comments); %import pulls in shared definitions; the ?, !, and _ modifiers change how a rule shapes the tree.

start: NUMBER (OP NUMBER)*
OP: "+" | "-" | "*" | "/"
NUMBER: /[0-9]+/
%ignore " "

The same grammar, every target

Python

import hyperlark as lark
p = lark.Lark(GRAMMAR, parser="lalr")
print(p.parse("1 + 2 * 3").pretty())

Rust

use hyperlark::Lark;
let p = Lark::new(GRAMMAR)?;
println!("{}", p.parse("1 + 2 * 3")?.pretty());

TypeScript

import { Lark } from "hyperlark";
const p = new Lark(GRAMMAR);
console.log(p.parse("1 + 2 * 3"));

C

Lark *p = NULL;
lark_from_source(GRAMMAR, strlen(GRAMMAR), LARK_OPT_NONE, &p);
LarkParseResult *r = NULL;
lark_parse(p, "1 + 2 * 3", 9, NULL, &r);

Learn more

The grammar language follows Lark’s grammar reference — hyperlark implements it natively. See the feature matrix for any target-specific notes.