YC-Lammy/pegy

Derive based parser generator for rust

Pegy - derive based Parser in Rust

Features

• derive based parser generation
• async api for parsing
• parse input from str, async readers or custom input types.
• AST generation on the run
• meaningful error messages for errors in grammar definition
• supports no_std

MSRV

The current MSRV supported is 1.75

Example

use pegy::Parse;

#[derive(Debug, Default, Parse)]
#[grammar($item0:['a'-'z''A'-'Z''0'-'9'])]
pub struct AlphaDigit(char);

#[derive(Debug, Default, Parse)]
#[grammar($item0:['0'-'9'])]
pub struct Digit(char);

#[derive(Debug, Default, Parse)]
#[grammar(!Digit $item0:AlphaDigit+)]
pub struct Ident(Vec<AlphaDigit>);

pub fn main(){
    let re: pegy::Result<Ident> = pegy::parse_blocking::<Ident, _>("myIdent");
    assert!(re.is_ok());
}

Comparison with similar traits

crate	action code	integration	input type	streaming input
pegy	in grammar	proc macro(derive)	&str, AsyncRead, custom	Yes
peg	in grammar	proc macro(block)	&str, &[T], custom	No
pest	external	proc macro(file)	&str	No

Expression Reference

Term

• "some string" - string literal: matches a str slice. returns &'static str.
• 'c' - character literal: matches a character. returns char.
• Ident - rule: matches a Parse rule. It must be a valid type and imlplements pegy::Parse. returns Ident type.
• ['a'-'z''A'-'Z''$'] - character class: matches a range of characters. returns char.

Quantifier

• ? - optional: matches zero or one term. returns Option<T>
• * - repeat: matches zero or more terms. returns Vec<T>
• + - repeat atleast: matches one or more terms. returns Vec<T>.
• {min, max} - repeat range: matches at least min and at most max number of terms. returns Vec<T>

Special

• $ident:term - field binding: bind the result of the term to the field ident of result. returns ().
• ( alternatives ) - group: matches the terms and returns a Span.
• terms | terms | terms - alternatives: trys to match the first terms, if failed, matches the second one and so on until a match is found. returns a Span.
• !term - negative lookahead: matches the term without consuming any characters.
• _ term - quiet: matches the term and returns ().