4.3 KiB

Raw Blame History

Regular Expression Tokenizer

Tokenizes strings that represent a regular expressions.

Usage

var ret = require('ret');

var tokens = ret(/foo|bar/.source);

tokens will contain the following object

{
  "type": ret.types.ROOT
  "options": [
    [ { "type": ret.types.CHAR, "value", 102 },
      { "type": ret.types.CHAR, "value", 111 },
      { "type": ret.types.CHAR, "value", 111 } ],
    [ { "type": ret.types.CHAR, "value",  98 },
      { "type": ret.types.CHAR, "value",  97 },
      { "type": ret.types.CHAR, "value", 114 } ]
  ]
}

Token Types

ret.types is a collection of the various token types exported by ret.

ROOT

Only used in the root of the regexp. This is needed due to the posibility of the root containing a pipe | character. In that case, the token will have an options key that will be an array of arrays of tokens. If not, it will contain a stack key that is an array of tokens.

{
  "type": ret.types.ROOT,
  "stack": [token1, token2...],
}

{
  "type": ret.types.ROOT,
  "options" [
    [token1, token2...],
    [othertoken1, othertoken2...]
    ...
  ],
}

GROUP

Groups contain tokens that are inside of a parenthesis. If the group begins with ? followed by another character, it's a special type of group. A ':' tells the group not to be remembered when exec is used. '=' means the previous token matches only if followed by this group, and '!' means the previous token matches only if NOT followed.

Like root, it can contain an options key instead of stack if there is a pipe.

{
  "type": ret.types.GROUP,
  "remember" true,
  "followedBy": false,
  "notFollowedBy": false,
  "stack": [token1, token2...],
}

{
  "type": ret.types.GROUP,
  "remember" true,
  "followedBy": false,
  "notFollowedBy": false,
  "options" [
    [token1, token2...],
    [othertoken1, othertoken2...]
    ...
  ],
}

POSITION

\b, \B, ^, and $ specify positions in the regexp.

{
  "type": ret.types.POSITION,
  "value": "^",
}

SET

Contains a key set specifying what tokens are allowed and a key not specifying if the set should be negated. A set can contain other sets, ranges, and characters.

{
  "type": ret.types.SET,
  "set": [token1, token2...],
  "not": false,
}

RANGE

Used in set tokens to specify a character range. from and to are character codes.

{
  "type": ret.types.RANGE,
  "from": 97,
  "to": 122,
}

REPETITION

{
  "type": ret.types.REPETITION,
  "min": 0,
  "max": Infinity,
  "value": token,
}

REFERENCE

References a group token. value is 1-9.

{
  "type": ret.types.REFERENCE,
  "value": 1,
}

CHAR

Represents a single character token. value is the character code. This might seem a bit cluttering instead of concatenating characters together. But since repetition tokens only repeat the last token and not the last clause like the pipe, it's simpler to do it this way.

{
  "type": ret.types.CHAR,
  "value": 123,
}

Errors

ret.js will throw errors if given a string with an invalid regular expression. All possible errors are

Invalid group. When a group with an immediate ? character is followed by an invalid character. It can only be followed by !, =, or :. Example: /(?_abc)/
Nothing to repeat. Thrown when a repetitional token is used as the first token in the current clause, as in right in the beginning of the regexp or group, or right after a pipe. Example: /foo|?bar/, /{1,3}foo|bar/, /foo(+bar)/
Unmatched ). A group was not opened, but was closed. Example: /hello)2u/
Unterminated group. A group was not closed. Example: /(1(23)4/
Unterminated character class. A custom character set was not closed. Example: /[abc/

Install

npm install ret

Tests

Tests are written with vows

npm test

License

MIT

4.3 KiB Raw Blame History