ascoders / syntax-parser

Light and fast 🚀parser! With zero dependents. - Sql Parser Demo added!

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

syntax-parser

syntax-parser is a parser using pure javascript, so it can both run in browser and nodejs.

CircleCI Status NPM Version Code Coverage

syntax-parser supports:

  • lexer.
  • parser.

Lexer

createLexer can help you create a lexer.

Example

import { createLexer } from 'syntax-parser';

const myLexer = createLexer([
  {
    type: 'whitespace',
    regexes: [/^(\s+)/],
    ignore: true
  },
  {
    type: 'word',
    regexes: [/^([a-zA-Z0-9]+)/]
  },
  {
    type: 'operator',
    regexes: [/^(\+)/]
  }
]);

myLexer('a + b');
// [
//   { "type": "word", "value": "a", "position": [0, 1] },
//   { "type": "operator", "value": "+", "position": [2, 3] },
//   { "type": "word", "value": "b", "position": [4, 5] }
// ]

type

Token type name, you can use any value here, and you will use it in the parser stage.

regexes

Regexes that use to be matched for each Token type.

ignore

The matching Token will not be added to the Token result queue.

In general, whitespace can be ignored in syntax parsing.

Parser

createParser can help you create a parser. Parser requires a lexer.

import { createParser, chain, matchTokenType, many } from 'syntax-parser';

const root = () => chain(addExpr)(ast => ast[0]);

const addExpr = () =>
  chain(matchTokenType('word'), many(addPlus))(ast => ({
    left: ast[0].value,
    operator: ast[1] && ast[1][0].operator,
    right: ast[1] && ast[1][0].term
  }));

const addPlus = () =>
  chain('+', root)(ast => ({
    operator: ast[0].value,
    term: ast[1]
  }));

const myParser = createParser(
  root, // Root grammar.
  myLexer // Created in lexer example.
);

myParser('a + b');
// ast:
// [{
//   "left": "a",
//   "operator": "+",
//   "right": {
//     "left": "b",
//     "operator": null,
//     "right": null
//   }
// }]

chain

Basic grammatical element, support four parameters:

string

String means match token:

chain('select', 'table'); // Match 'select table'

array

Array means 'or':

chain('select', ['table', 'chart']); // Match both 'select table' and 'select chart'

matchTokenType

matchTokenType allow you match Token type defined in lexer.

chain('select', matchTokenType('word')); // Match 'select [any word!]'

function

It's easy to call another chain function:

const a = () => chain('select', b);
const b = () => chain('table');

many/optional

Just as literal meaning:

const a = () => chain('select', optional('table')); // Match both 'select' and 'select table'
const b = () => chain('select', many(',', matchTokenType('word'))); // Match both 'select' and 'select a' and 'select a, b' .. and so on.

optional many can also use chain as parameter. many(chain(..))

The last callback allow partial redefin of local ast:

chain('select', 'table')(
  ast => ast[0] // return 'select'
);

Tests

npm test

Monaco Editor Sql Editor

If you want to see this demo, run this command:

npm run docs

Then select demo Monaco Editor.

About

Light and fast 🚀parser! With zero dependents. - Sql Parser Demo added!

License:MIT License


Languages

Language:TypeScript 99.8%Language:JavaScript 0.2%