tree-sitter-peg

A Tree-sitter grammar for Parsing Expression Grammar (PEG) files.

This grammar supports the PEG format used by pigeon and similar PEG parsers, including an optional Go preamble section.

Features

• Full PEG expression parsing: sequences, ordered choices (/), quantifiers (*, +, ?), predicates (&, !), grouping ((...))
• String literals with escape sequences (single-quoted '...' and double-quoted "...")
• Character classes ([a-z])
• Any-character (.)
• Line comments (# ...)
• Optional Go preamble (package, type ... Peg { ... })

Node Types

Node	Description
source_file	Root node
go_section	Optional Go preamble
peg_section	Contains all PEG rules
peg_rule	A single rule definition
rule_name	The name being defined (field: rule)
operator_assign	The <- operator
rule_body	The rule's expression
choice_expr	Alternatives separated by /
sequence_expr	Sequence of expressions
prefix_expr	Predicate + expression
suffix_expr	Expression + quantifier
primary_expr	Atom or group
group	Parenthesized expression
identifier	Rule reference
literal	String literal
single_quoted	'...' literal
double_quoted	"..." literal
char_class	[...] character class
any_char	. wildcard
and_predicate	& lookahead
not_predicate	! negative lookahead
quantifier	*, +, or ?
comment	# ... line comment

Installation

Prerequisites

• Node.js (for grammar development)
• A C compiler (for building the shared library)
• tree-sitter CLI

Building from source

git clone https://github.com/shen390s/tree-sitter-peg.git
cd tree-sitter-peg/tree-sitter-peg
npm install
npx tree-sitter generate

For Neovim (nvim-treesitter)

Add to your nvim-treesitter configuration:

local parser_config = require("nvim-treesitter.parsers").get_parser_configs()
parser_config.peg = {
  install_info = {
    url = "https://github.com/shen390s/tree-sitter-peg",
    files = { "tree-sitter-peg/src/parser.c" },
    branch = "master",
  },
  filetype = "peg",
}

For Emacs 29+ (treesit)

Using the companion peg-ts-mode package:

;; Install the grammar
(add-to-list 'treesit-language-source-alist
             '(peg "https://github.com/shen390s/tree-sitter-peg"
                   nil nil nil "tree-sitter-peg"))
(treesit-install-language-grammar 'peg)

Or interactively after loading peg-ts-mode:

M-x peg-ts-mode-install-grammar

For Helix

Add to languages.toml:

[[language]]
name = "peg"
scope = "source.peg"
file-types = ["peg"]
roots = []

[language.auto-pairs]
'(' = ')'
'[' = ']'
'{' = '}'
'"' = '"'
"'" = "'"

[[grammar]]
name = "peg"
source = { git = "https://github.com/shen390s/tree-sitter-peg", rev = "master", subpath = "tree-sitter-peg" }

Development

Generating the parser

npx tree-sitter generate

Running tests

npx tree-sitter test

Parsing an example file

npx tree-sitter parse examples/gnparser.peg

Example

package parser

type Engine Peg {
  baseEngine
}

# Match a greeting
Greeting <- Hello _ Name '!'

Hello <- 'hello' / 'hi' / 'hey'

Name <- [A-Z] [a-z]+

_ <- ' '+

License

MIT