transliteration/lib/node/transliterate.js

Unicode to ACSII transliteration / slugify module. Works in node.js, web browser and command line.

'use strict';

exports.__esModule = true;
exports.replaceStr = undefined;

var _utils = require('./utils');

var codemap = {};
var defaultOptions = {
  unknown: '[?]',
  replace: [],
  replaceAfter: [],
  ignore: []
};
var configOptions = {};

/* istanbul ignore next */
var replaceStr = exports.replaceStr = function replaceStr(str, replace) {
  for (var _iterator = replace, _isArray = Array.isArray(_iterator), _i = 0, _iterator = _isArray ? _iterator : _iterator[Symbol.iterator]();;) {
    var _ref;

    if (_isArray) {
      if (_i >= _iterator.length) break;
      _ref = _iterator[_i++];
    } else {
      _i = _iterator.next();
      if (_i.done) break;
      _ref = _i.value;
    }

    var item = _ref;

    if (item[0] instanceof RegExp) {
      var flag = item[0].flags;
      if (!item[0].global) {
        flag += 'g';
        item[0] = new RegExp(item[0].toString().replace(/^\/|\/$/), flag);
      }
    } else if (typeof item[0] === 'string') {
      item[0] = new RegExp((0, _utils.escapeRegExp)(item[0]), 'g');
    }
    if (item[0] instanceof RegExp) {
      str = str.replace(item[0], item[1]);
    }
  }
  return str;
};

/**
 * @param {string} str The string which is being transliterated
 * @param {object} options options
 */
/* istanbul ignore next */
var transliterate = function transliterate(str, options) {
  var opt = options ? (0, _utils.mergeOptions)(defaultOptions, options) : (0, _utils.mergeOptions)(defaultOptions, configOptions);
  str = String(str);
  if (opt.ignore instanceof Array && opt.ignore.length > 0) {
    for (var i in opt.ignore) {
      var splitted = str.split(opt.ignore[i]);
      var result = [];
      for (var j in splitted) {
        var ignore = opt.ignore.slice(0);
        ignore.splice(i, 1);
        result.push(transliterate(splitted[j], (0, _utils.mergeOptions)(opt, { ignore: ignore })));
      }
      return result.join(opt.ignore[i]);
    }
  }
  str = replaceStr(str, opt.replace);
  var strArr = (0, _utils.ucs2decode)((0, _utils.fixChineseSpace)(str));
  var strArrNew = [];

  for (var _iterator2 = strArr, _isArray2 = Array.isArray(_iterator2), _i2 = 0, _iterator2 = _isArray2 ? _iterator2 : _iterator2[Symbol.iterator]();;) {
    var _ref2;

    if (_isArray2) {
      if (_i2 >= _iterator2.length) break;
      _ref2 = _iterator2[_i2++];
    } else {
      _i2 = _iterator2.next();
      if (_i2.done) break;
      _ref2 = _i2.value;
    }

    var ord = _ref2;

    // These characters are also transliteratable. Will improve it later if needed
    if (ord > 0xffff) {
      strArrNew.push(opt.unknown);
    } else {
      var offset = ord >> 8;
      if (typeof codemap[offset] === 'undefined') {
        try {
          codemap[offset] = require('../../data/x' + offset.toString(16) + '.json');
        } catch (e) {
          codemap[offset] = [];
        }
      }
      ord &= 0xff;
      var t = codemap[offset][ord];
      if (typeof t === 'undefined' || t === null) {
        strArrNew.push(opt.unknown);
      } else {
        strArrNew.push(codemap[offset][ord]);
      }
    }
  }
  // trim spaces at the begining and ending of the string
  if (strArrNew.length > 1) {
    opt.replaceAfter.push([/(^ +?)|( +?$)/g, '']);
  }
  var strNew = strArrNew.join('');

  strNew = replaceStr(strNew, opt.replaceAfter);
  return strNew;
};

transliterate.setCodemap = function (customCodemap) {
  codemap = customCodemap || codemap;
  return codemap;
};

transliterate.config = function (options) {
  if (options === undefined) {
    return configOptions;
  }
  configOptions = (0, _utils.mergeOptions)(defaultOptions, options);
  return configOptions;
};

exports.default = transliterate;