@carlosvpi/parser

A <6kB Parsing utility

Install

npm install @carlosvpi/parser

yarn add @carlosvpi/parser

Parse a string

Import the parser

const getParser = require('@carlosvpi/parser')

import getParser from '@carlosvpi/parser'

Use the parser

const grammar = `
  S = 's' A | '0';
  A = 'a' S;
`
const parser = getParser(grammar)

parser is a hash. Each rule head of the grammar is a key in the hash. In this case,

typeof parser.S === 'function'
typeof parser.A === 'function'

If the string head is the head of a rule, and input is some string, then

parser[head](input)

parses the input using the given grammar, using the head as an axiom. The output of parser[head](input) is a parsed tree.

The grammar

The grammar that the getParser function (from @carlosvpi/parser) takes is a string that follows the same syntax described in the Wikipedia page for EBNF.

The tree

The result of parser[head](input) is a tree where nodes have this shape:

[node, endIndex, errors]

endIndex is the las index of the input that has been parsed. If it does not equal the length of the input, it means that the rest of the input does not follow the grammar.

errors is a list of strings, each indicating a possible error during the parsing. Erros have this shape: "✘ 2:12 | Expected 'non-terminal W', got 'rest of input"

The node is a pair [root, ...children], where root is a string with these possible values:

• '$LITERAL': has only one child, the parsed literal as a string,
• '$MATCH': its children are the result of applying a regex to the input using js regex match,
• '$CONCAT': its children are other nodes as an array,
• head: this node corresponds to the parsing the input using a rule head

Example

This code

parser(`
  S = 's' A | '0';
  A = 'a' S;
`).S('sasa0')[0]

produces this tree:

[
   "S",
   [
      "$CONCAT",
      [
         "$LITERAL",
         "s"
      ],
      [
         "A",
         [
            "$CONCAT",
            [
               "$LITERAL",
               "a"
            ],
            [
               "S",
               [
                  "$CONCAT",
                  [
                     "$LITERAL",
                     "s"
                  ],
                  [
                     "A",
                     [
                        "$CONCAT",
                        [
                           "$LITERAL",
                           "a"
                        ],
                        [
                           "S",
                           [
                              "$CONCAT",
                              [
                                 "$LITERAL",
                                 "0"
                              ]
                           ]
                        ]
                     ]
                  ]
               ]
            ]
         ]
      ]
   ]
]