doken

A minimalistic, general purpose tokenizer generator.

Usage

Use npm to install:

$ npm install doken

Import doken and create a tokenizer with regular expression rules:

const {createTokenizer, regexRule} = require('doken')
 
const tokenizeJSON = createTokenizer({
  rules: [
    regexRule('_whitespace', /\s+/y, {lineBreaks: true}),
    regexRule('brace', /[{}]/y),
    regexRule('bracket', /[\[\]]/y),
    regexRule('colon', /:/y),
    regexRule('comma', /,/y),
    regexRule('string', /"([^"\n\\]|\\[^\n])*"/y),
    regexRule('number', /(-|\+)?\d+(.\d+)?/y),
    regexRule('boolean', /(true|false)\b/y),
    regexRule('null', /null\b/y)
  ]
})
 
let tokens = tokenizeJSON(`{"a": "Hello World!"}`)
 
console.log([...tokens])

API

Rule object

A rule object contains the following fields:

• type <string> - The type of the token this rule generates. If type starts with an underscore _, the token will not be emitted by the tokenizer.

• lineBreaks <boolean> (Optional) - Set this property to true if this rule might match line breaks and you want to track it correctly.

• match <Function> - A function with the following signature:

  (input: string, position: number) =>
    null |
    {
      length: number,
      value: any
    }

  This function will try to get the token of given type at given position in input if applicable. Return null if input at position is not a token with the given type, otherwise return an object.

  The first length characters of input after position will be the matched token. You can optionally return a value which can contain any data that will be attached to the token. If value is not given, it will default to the first length characters of input after position.

Token object

A token will be represented by an object with the following fields:

• type <string> | null - The type of the token or null if no given rules match the input.
• length <number> - The length of the token.
• value <any> - The value generated by the rule.
• pos <number> - The zero-based position of the first character of the token.
• row <number> - The zero-based row of the first character of the token.
• col <number> - The zero-based column of the first character of the token.

doken.createTokenizer(options)

• options <object>
  • rules <Array<Rule>>
  • strategy 'first' | 'longest' (Optional) - Default: 'first'
• Returns: <Function>

Generates a tokenize function with the following signature:

(input: string) => IterableIterator<Token>

This function will attempt to tokenize given input, yielding tokens matched by given rules one by one.

Set strategy to 'longest' to match the token with the rule that matches the most characters instead of using the rule that matches first.

doken.regexRule(type, regex[, options])

• type <string>
• regex <RegExp>
• options <object>
  • lineBreaks <boolean> (Optional) - Set this property to true if this rule might match line breaks and you want to track it correctly.
  • value <Function> (Optional) - A function for calculating the token value out of the match.
  • condition <Function> (Optional) - A function for indicating whether to discard match or not.
• Returns: <Rule>

Returns a rule that attempts to match input string with the given regex.

value can be set to a function (match: RegExpExecArray) => any. The generated token will have the returned value as value.

condition can be set to a function (match: RegExpExecArray) => boolean. Return false to indicate to discard matched result and go on with the next rule.