beginner-friendly-lang/tree-sitter/grammar.js

/**
 * @file Parser for the Crepuscular(ray) functional programming language
 * @author Alexander Nutz <alexander.nutz@vxcc.dev>
 * @license MIT
 */

/// <reference types="tree-sitter-cli/dsl" />
// @ts-check

module.exports = grammar({
  name: "crepuscular",

  reserved: {
    toplevel_kw: $ =>
    ['type', 'with', 'extensible', 'extend', 'union', 'def'],
  },

  extras: ($) => [
    /\s/, // whitespace
    $.comment,
  ],

  word: $ => $._identifier_tok,

  rules: {
    source_file: $ => repeat($._definition),

    _identifier_tok: $ => token(/[a-zA-Z_]+[a-zA-Z0-9_]*/),
    identifier: $ => reserved('toplevel_kw', $._identifier_tok),
    path: $ => prec.left(seq($.identifier, repeat(seq('.', $.identifier)))),

    comment: $ =>
      token(seq("#", /.*/)),

    _definition: $ => choice(
      $.full_partial_type_definition,
      $.type_definition,
      $.extensible_union,
      $.extend_decl,
      $.def,
    ),

    extensible_union: $ => seq(
      'extensible', 'union', $.path),

    extend_decl: $ => seq(
      'extend', $.path, 'with', $.tag, $._type),

    full_partial_type_definition: $ => seq(
      "type",
      "?", $.path,
      "=",
      $._type
    ),

    type_definition: $ => seq(
      "type",
      repeat($.identifier),
      $.path,
      "=",
      $._type
    ),

    _type_atom: $ => choice(
      $.just_type,
      $.partial_type,
      seq('(', $._type, ')'),
      $.record_type,
    ),

    _type_non_fn: $ => choice(
      $._type_atom,
      $.tagged_type,
      $.union_type,
      $.partial_union_type,
      $.parametrized_type,
      $.with_type,
      $.recursive_type,
    ),

    _type: $ => choice(
      $._type_non_fn,
      $.fn_type,
    ),

    union_type: $ => prec.left(1,
      seq($._type, '|', $._type)),

    partial_union_type: $ => prec.left(1,
      seq($._type, '|', '...', $.partial_type)),

    tag: $ => new RustRegex("'(?:[a-zA-Z_][a-zA-Z0-9_]*(?:[.][a-zA-Z_0-9]+)*)"),
    tagged_type: $ => prec.right(3,
      seq($.tag, optional(
        choice(
          $._type_atom,
          $.parametrized_type)))),

    multi_type_parameters: $ => seq('[', $._type, repeat(seq(',', $._type)), ']'),
    parametrized_type: $ => prec.left(4, seq(
      choice(
        $.multi_type_parameters,
        $._type_atom
      ),
      repeat1($.path)
    )),

    with_type: $ => seq('with', $.identifier, repeat(seq(',', $.identifier)), ':', $._type),

    recursive_type: $ => seq('&', $.identifier, $._type),

    partial_type: $ => seq('?', $.identifier),

    fn_type: $ => prec.left(-10,
      seq($._type, '->', $._type)),

    just_type: $ => prec(-1, $.path),

    record_type_field: $ => seq($.identifier, ':', $._type),
    record_type: $ => seq(
      '{',
      repeat(seq($.record_type_field, ',')),
      optional(choice(
        $.record_type_field,
        seq('...', $.partial_type),
      )),
      '}'),

    escape_sequence: $ =>
      token.immediate(
        seq('\\', /[tbrnf0"'\\]/)),

    char_middle: $ => /./,
    string_middle: $ => /[^\\"]+/,

    char_literal: $ =>
      seq('\'', choice($.escape_sequence, $.char_middle), '\''),

    string_literal: $ =>
      seq('"', repeat(choice($.escape_sequence, $.string_middle)), '"'),

    num_literal: $ =>
      seq(
        choice(
          /[0-9]+/,
          /\-[0-9]+/
        ),
        optional(token.immediate(/[.][0-9]+/))
      ),

    list_expression: $ =>
      seq(
        '[',
        repeat(seq($.expression, ',')),
        optional($.expression),
        ']'),

    field_access: $ => prec.left(
      seq($._atom, ':', $.identifier)),

    function_call: $ => prec.left(1,
      seq($._atom, '(',
        repeat(seq($.expression, ',')), optional($.expression),
        ')')),

    ident_expr: $ => $.identifier,

    record_expr_field: $ =>
      seq($.identifier, ':', $.expression),

    record_expr: $ => seq(
      '{',
      repeat(seq($.record_expr_field, ',')),
      optional($.record_expr_field),
      '}'),

    _atom: $ => choice(
      prec(0, seq('(', $.expression, ')')),
      $.ident_expr,
      $.char_literal,
      $.string_literal,
      $.num_literal,
      $.list_expression,
      $.field_access,
      $.function_call,
      $.record_expr,
      // TODO
    ),

    let_binding: $ => seq(
      'let',
      $.identifier,
      '=',
      $.expression,
      optional('in'),
      $.expression,
    ),

    await_binding: $ => seq(
      'await',
      $.identifier,
      '=',
      $.expression,
      optional('in'),
      $.expression,
    ),

    type_downcast: $ => seq(
      $._atom,
      '::',
      $._type,
    ),

    lambda: $ => prec.right(4, seq(
      $.identifier,
      optional(seq(':', $._type_non_fn)),
      '->',
      $.expression
    )),

    expression: $ => choice(
      $._atom,
      $.let_binding,
      $.await_binding,
      $.type_downcast,
      $.lambda,
      // TODO
    ),

    def: $ => seq(
      'def', $.path,
      choice(
        seq(':', $._type),
        seq(
          optional(seq(':', $._type)),
          seq('=', $.expression),
        )
      ),
    ),
  }
});