regex.js

/**
 * @param {string} value
 * @returns {RegExp}
 * */
function escape(value) {
  return new RegExp(value.replace(/[-/\\^$*+?.()|[\]{}]/g, '\\$&'), 'm');
}

/**
 * @param {RegExp | string } re
 * @returns {string}
 */
function source(re) {
  if (!re) return null;
  if (typeof re === "string") return re;

  return re.source;
}

/**
 * @param {RegExp | string } re
 * @returns {string}
 */
function lookahead(re) {
  return concat('(?=', re, ')');
}

/**
 * @param {RegExp | string } re
 * @returns {string}
 */
function anyNumberOfTimes(re) {
  return concat('(?:', re, ')*');
}

/**
 * @param {RegExp | string } re
 * @returns {string}
 */
function optional(re) {
  return concat('(?:', re, ')?');
}

/**
 * @param {...(RegExp | string) } args
 * @returns {string}
 */
function concat(...args) {
  const joined = args.map((x) => source(x)).join("");
  return joined;
}

/**
 * @param { Array<string | RegExp | Object> } args
 * @returns {object}
 */
function stripOptionsFromArgs(args) {
  const opts = args[args.length - 1];

  if (typeof opts === 'object' && opts.constructor === Object) {
    args.splice(args.length - 1, 1);
    return opts;
  } else {
    return {};
  }
}

/** @typedef { {capture?: boolean} } RegexEitherOptions */

/**
 * Any of the passed expresssions may match
 *
 * Creates a huge this | this | that | that match
 * @param {(RegExp | string)[] | [...(RegExp | string)[], RegexEitherOptions]} args
 * @returns {string}
 */
function either(...args) {
  /** @type { object & {capture?: boolean} }  */
  const opts = stripOptionsFromArgs(args);
  const joined = '('
    + (opts.capture ? "" : "?:")
    + args.map((x) => source(x)).join("|") + ")";
  return joined;
}

/**
 * @param {RegExp | string} re
 * @returns {number}
 */
function countMatchGroups(re) {
  return (new RegExp(re.toString() + '|')).exec('').length - 1;
}

/**
 * Does lexeme start with a regular expression match at the beginning
 * @param {RegExp} re
 * @param {string} lexeme
 */
function startsWith(re, lexeme) {
  const match = re && re.exec(lexeme);
  return match && match.index === 0;
}

// BACKREF_RE matches an open parenthesis or backreference. To avoid
// an incorrect parse, it additionally matches the following:
// - [...] elements, where the meaning of parentheses and escapes change
// - other escape sequences, so we do not misparse escape sequences as
//   interesting elements
// - non-matching or lookahead parentheses, which do not capture. These
//   follow the '(' with a '?'.
const BACKREF_RE = /\[(?:[^\\\]]|\\.)*\]|\(\??|\\([1-9][0-9]*)|\\./;

// **INTERNAL** Not intended for outside usage
// join logically computes regexps.join(separator), but fixes the
// backreferences so they continue to match.
// it also places each individual regular expression into it's own
// match group, keeping track of the sequencing of those match groups
// is currently an exercise for the caller. :-)
/**
 * @param {(string | RegExp)[]} regexps
 * @param {{joinWith: string}} opts
 * @returns {string}
 */
function _rewriteBackreferences(regexps, { joinWith }) {
  let numCaptures = 0;

  return regexps.map((regex) => {
    numCaptures += 1;
    const offset = numCaptures;
    let re = source(regex);
    let out = '';

    while (re.length > 0) {
      const match = BACKREF_RE.exec(re);
      if (!match) {
        out += re;
        break;
      }
      out += re.substring(0, match.index);
      re = re.substring(match.index + match[0].length);
      if (match[0][0] === '\\' && match[1]) {
        // Adjust the backreference.
        out += '\\' + String(Number(match[1]) + offset);
      } else {
        out += match[0];
        if (match[0] === '(') {
          numCaptures++;
        }
      }
    }
    return out;
  }).map(re => `(${re})`).join(joinWith);
}

export { _rewriteBackreferences, anyNumberOfTimes, concat, countMatchGroups, either, escape, lookahead, optional, source, startsWith };