parsing

Parsing

A JSON-esque grammar-based Packrat PEG parser.

Getting started

$ npm install parsing
$ node

/**
 * Based on this EBNF grammar
 * - from http://stackoverflow.com/questions/6805172/how-do-you-abstract-some-expression-to-bnf#answer-6805185
 *
 * AEXP => AS+
 * AS   => id ':=' EX1 ';'
 * EX1  => EX2 (('+' | '-') EX2)*
 * EX2  => EX3 (('*' | '/') EX3)*
 * EX3  => EX4 ('^' EX3)*
 * EX4  => ('+'|'-')? EX5
 * EX5  => id | number | '(' EX1 ')'
 */
var grammarSpec = {
    ignore: 'whitespace',
    rules: {
        'assign': /:=/,
        'character': /[;*\/^+-]/,
        'id': /[\w$][\w\d$]*/,
        'number': /\d(?:\.\d+)?/,
        'whitespace': /\s+/,
        'AEXP': {
            components: {name: 'assignment', oneOrMoreOf: 'AS'}
        },
        'AS': {
            components: [{name: 'target', what: 'id'}, 'assign', {name: 'expression', what: 'EX1'}, {'character': ';'}]
        },
        'EX1': {
            captureAs: 'EX',
            components: [{name: 'left', what: 'EX2'}, {name: 'right', zeroOrMoreOf: [{name: 'operator', oneOf: [{'character': '+'}, {'character': '-'}]}, {name: 'operand', what: 'EX2'}]}],
            ifNoMatch: {component: 'right', capture: 'left'}
        },
        'EX2': {
            captureAs: 'EX',
            components: [{name: 'left', what: 'EX3'}, {name: 'right', zeroOrMoreOf: [{name: 'operator', oneOf: [{'character': '*'}, {'character': '/'}]}, {name: 'operand', what: 'EX3'}]}],
            ifNoMatch: {component: 'right', capture: 'left'}
        },
        'EX3': {
            captureAs: 'EX',
            components: [{name: 'left', what: 'EX4'}, {name: 'right', zeroOrMoreOf: [{name: 'operator', what: {'character': '^'}}, {name: 'operand', rule: 'EX3'}]}],
            ifNoMatch: {component: 'right', capture: 'left'}
        },
        'EX4': {
            captureAs: 'EX',
            components: [{name: 'operator', optionally: {oneOf: [{'character': '+'}, {'character': '-'}]}}, {name: 'operand', what: 'EX5'}],
            ifNoMatch: {component: 'operator', capture: 'operand'}
        },
        'EX5': {
            components: [{oneOf: ['id', 'number', [{'character': '('}, 'EX1', {'character': ')'}]]}]
        }
    },
    start: 'AEXP'
};

var parser = require('parsing').create(grammarSpec);

console.log(parser.parse('waldo:=1;'));
/**
 * Gives:
 * {
 *     name: 'AEXP',
 *     assignment: [{
 *         name: 'AS',
 *         target: 'waldo',
 *         expression: '1'
 *     }]
 * }
 */

Extending a grammar / Defining custom rules

Sometimes it can be handy to extend an existing grammar by defining custom rules and overriding some existing ones. For example, given this simple grammar we could add a new statement type to it:

var grammarSpec = {
    ignore: 'whitespace',
    rules: {
        'go_statement': {
            components: [{what: /go/, allowMerge: false}]
        },
        'end_statement': {
            components: [{what: /end/, allowMerge: false}]
        },
        'whitespace': /\s+/,
        'single_statement': {
            components: {oneOf: ['go_statement', 'end_statement']}
        },
        'statement': {
            components: ['single_statement', /;/]
        },
        'program': {
            components: {name: 'statements', zeroOrMoreOf: 'statement'}
        }
    },
    start: 'program'
};

// Note the third `options` argument here, which we use to pass the custom rule specs,
// adding support for the new `do_something_custom;` statement to the above grammar
var parser = require('parsing').create(grammarSpec, null, {
    rules: {
        'do_something_statement': {
            components: [{what: /do_something_custom/, allowMerge: false}]
        },
        'single_statement': {
            // Override `single_statement`, but refer back to the original
            components: {oneOf: ['do_something_statement', 'single_statement']}
        }
    }
});

console.log(parser.parse('go; do_something_custom; end;'));
/**
 * Gives:
 * {
 *     name: 'program',
 *     statements: [
 *         { name: 'go_statement' },
 *         { name: 'do_something_statement' },
 *         { name: 'end_statement' }
 *     ]
 * }
 */

Keeping up to date

• Follow me on Twitter for updates: https://twitter.com/@asmblah