nano-parser

Parser combinators library for parsing strings and arrays of strings (developed for parsing ecmascript tagged template strings). Read in russian

Installing

npm install --save 'nano-parser';

Usage

    const {
        any,
        defer,
        end,
        find,
        next,
        optional,
        repeat,
        required,
        sequence
    } = require('nano-parser');

    const parser = find(/[a-z]+/);
    parser.parse('abcd'); // abcd
    parser.parse('ab1cd'); // ab
    parser.parse('1abcd'); // undefined

Examples of using:

• es6x - implementation of jsx features in pure javascript

Overview

With the help of nano-parser, you can create complex parsers, combining simple. It's possible to transform results using the method then, as well as to use the cache to improve performance.

Elementary parsers

find, next and end are most simple parsers. find parser designed to search for strings or regular expressions. next and end used when parsing an array of strings (useful for es6 template strings).

Combinators

any, conditional, sequence and repeat designed to combine other parsers with each other.

Subsidiary parsers

optional, required and defer are subsidiary. optional makes parser optional, required makes parser strictly necessary and defer provides the possibility to call the parser from itself.

Class Parser

All parsers return a instance of class Parser. Therefore, they have the same interface. The exec method is internal and is used to call from other parsers. Instead, use the parse method. The parse method takes the first argument string or array of strings, optional second parameter values are values that will be available as the second parameter of the method then. To cache the results use useCache method.

Documentation

find

find is used to search for a regular expression or a direct match with a string. Examples:

    const parser = find('a');
    expect(parser.parse('abc')).toBe('a');
    expect(parser.parse('bac')).toBe(undefined);

    const parser = find(/^[a-z]/);
    expect(parser.parse('abc')).toBe('abc');
    expect(parser.parse('bac')).toBe('bac');
    expect(parser.parse('1abc')).toBe(undefined);

    const parser = find(/^[a-z]/).not(find('x'));
    expect(parser.parse('abc')).toBe('abc');
    expect(parser.parse('bacx')).toBe(undefined);

The symbol "^" is recommended to optimize the regular expression.

any

any makes it possible to find one of the matches.

    const parser = any(
        find('a'),
        find('c')
    );
    expect(parser.parse('a')).toBe('a');
    expect(parser.parse('c')).toBe('c');
    expect(parser.parse('b')).toBe(undefined);

    const parser = any(
        find(/^[a-c]/),
        find(/^[e-g]/)
    ).not(find('abc'));
    expect(parser.parse('ab')).toBe('ab');
    expect(parser.parse('fg')).toBe('fg');
    expect(parser.parse('abc')).toBe(undefined);

sequence

With sequence you can build sequences of parsers.

    const parser = sequence(
        find('ab'),
        find(/^[c-e]+/),
        find(/^[0-9]/)
    );
    expect(parser.parse('abcc1')).toBe(['ab', 'cc', '1']);
    expect(parser.parse('abcc')).toBe(undefined);

    const parser = sequence(
        find('ab'),
        find(/^[c-e]+/),
        find(/^[0-9]/)
    ).then(results => results.join(''));
    expect(parser.parse('abcc1')).toBe('abcc1');
    expect(parser.parse('abcc')).toBe(undefined);

In the previous example you can see in action a useful method then, which is in any parser. And also, any parser has methods not and useCache. then and not return a new parser, so you can easily form a new entity, without affecting the existing:

    const parser1 = find(/^[a-c]/),
        parser2 = parser1.not(find('b')),
        parser3 = parser1.then(result => result + '1'),
        parserCached = parser1.useCache();
    expect(parser1).not.toBe(parser2);
    expect(parser1).not.toBe(parser3);
    expect(parser1).toBe(parserCached);

repeat

repeat is used to find the repeated sequences.

    const
        name = find(/^[a-z0-9]/i),
        assignment = sequence(name, find('='), name).then(results => [results[0], results[2]]),
        whiteSpace = find(/^\s+/),
        parser = repeat(assignment, whiteSpace);

    expect(parser.parse('a=a b=b c=d')).toEqual([['a', 'a'], ['b', 'b'], ['c', 'd']]);
    expect(parser.parse('a=a b=b x c=d')).toEqual([['a', 'a'], ['b', 'b']]);
    expect(parser.parse('x a=a b=b c=d')).toBe(undefined);

conditional

conditional is useful when you want to look forward.

    const condition = find(/[a-z]+\d+/),
        pattern1 = find(/[a-z]+/),
        pattern2 = pattern1.then(() => 'x'),
        parser = conditional(
            condition,
            pattern1,
            pattern2
        );

    expect(parser.parse('abc')).toBe('x');
    expect(parser.parse('abc1')).toBe('abc');

optional

optional makes parsers optional.

    const whiteSpace = find(/^\s+/),
        optionalWhiteSpace = optional(whiteSpace),
        parser1 = sequence(find('a'), whiteSpace, find('b')),
        parser2 = sequence(find('a'), optionalWhiteSpace, find('b'));

    expect(parser1.parse('a b')).toEqual(['a', ' ', 'b']);
    expect(parser1.parse('ab')).toBe(undefined);
    expect(parser2.parse('a b')).toEqual(['a', ' ', 'b']);
    expect(parser2.parse('ab')).toEqual(['a', undefined, 'b']);

required

required will generate an error that the expected characters are not found:

    const whiteSpace = find(/^\s+/),
        requeiredWhiteSpace = required(whiteSpace),
        parser1 = sequence(find('a'), whiteSpace, find('b')),
        parser2 = sequence(find('a'), requiredWhiteSpace, find('b'));

    expect(parser1.parse('a b')).toEqual(['a', ' ', 'b']);
    expect(parser1.parse('ab')).toBe(undefined);
    expect(parser2.parse('a b')).toEqual(['a', ' ', 'b']);
    expect(parser2.parse('ab')).toThrow(); // error of parsing with position

next and end

next and end needed when parsing arrays of strings. next correspond to the spacing between strings, end - end of the array. These methods are useful when parsing es6 tagged template strings.

    let lang;
    const i18Table = {
            'привет': 'hello'
        },
        anything = find(/^.*/),
        parser = repeat(any(
            anything,
            next().then((index, values) => lang === 'en' ? i18Table[values[index]] : values[index])
        )).then(results => results.join('')),
        i18n = function(strings, ...values) {
            return parser.parse(strings, values);
        };

    lang = 'en';
    expect(i18n `<div class="container">${'привет'}</div>`).toBe('<div class="container">hello</div>');
    lang = 'ru';
    expect(i18n `<div class="container">${'привет'}</div>`).toBe('<div class="container">привет</div>');

defer

defer useful for creating recursive parsers.

    const xml = sequence(
        find('<'),
        find(/^[a-z]+/),
        find('>'),
        repeat(any(
            find(/^[^<]/),
            defer(() => xml) // we can't use xml directly here, because it is undefined now
        ))
        find('<'),
        find(/^[a-z]+/),
        find('>'),
    ).then(results => ({
        tag: results[1],
        children: results[3]
    }));

    expect(xml.parse('<div><p>text</p></div>')).toEqual({
        tag: 'div',
        children: [{
            tag: 'p',
            children: ['text']
        }]
    });