koranAddresses.js

/**
 * Attempts to parse the string into a SK address.
 * https://en.wikipedia.org/wiki/Addresses_in_South_Korea#Postal_Address
 * https://en.wikipedia.org/wiki/ISO_3166-2:KR
 * https://en.wikipedia.org/wiki/Administrative_divisions_of_South_Korea#Gun_.28County.29
 * @returns An object of the form {province, city, street, zip}
 */
function parseAddress(string) {
    var address = {};
    address.extra = [];

    //for crawling purposes, these are all provinces.
    var province = ["-do", "-teukbyeol-jachido", "-teukbyeolsi", "busan", "-teukbyeol-jachisi", "-gwangyeoksi"];
    //Cities and districts
    var city = ["-si", "-gun", "-gu", "seoul", "mokpo"];
    //town, township and neighborhood
    var town = ["-eup", "-myeon", "-dong"];
    //rural areas 
    var rural = ["-tong", "-ri"];
    var hamlet = ["-ban"];
    //zipcode
    var zipRegex = /(\d+-\d+)$/;
    var parts = string.split(" "); //space note

    parts.forEach(function(item, idx) {
        var item = item.toLowerCase().replace(",", "");
        if (province.find(compare, item)) {
            address.province = item.replace(/-.+/, "").capitalize();
        } else if (city.find(compare, item)) {
            address.city = item.replace(/-.+/, "").capitalize();
        } else if (town.find(compare, item)) {
            address.town = item.replace().capitalize(); //keep the -dong in the town name.
        } else if (rural.find(compare, item)) {
            address.ruralArea = item.capitalize();
        } else if (hamlet.find(compare, item)) {
            address.hamlet = item.capitalize();
        }
        else if (item.match(zipRegex)) {
            address.zip = item;
        } else if (item.match(/south|korea/)) {
            //discard
        }
        else {
            address.extra.push(item.capitalize());
        }
    });

    //find callback
    function compare(element, idx, arr) {
        var reg = new RegExp(element + "$", "i");
        return reg.test(this);
    }
    address.streetLevelAddress = address.extra;
    if (address.town) address.extra.push(address.town);
    if (address.rural) address.extra.push(address.rural);
    if (address.hamlet) address.extra.push(address.hamlet);
    address.streetLevelAddress = address.extra.join(" ");
    return address;
}

/*
* Polyfill required for <ES6
*/

//polyfill arr.find method until the current version of node supports ES6 features
//https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Array/find
if (!Array.prototype.find) {
  Array.prototype.find = function(predicate) {
      if (this === null) {
          throw new TypeError('Array.prototype.find called on null or undefined');
      }
      if (typeof predicate !== 'function') {
          throw new TypeError('predicate must be a function');
      }
      var list = Object(this);
      var length = list.length >>> 0;
      var thisArg = arguments[1];
      var value;
        for (var i = 0; i < length; i++) {
          value = list[i];
          if (predicate.call(thisArg, value, i, list)) {
              return value;
          }
      }
  return undefined;
};
}

  
/**
 * Capitalize the first character of a string. 
 * Usage: myString.capitalize() -> MyString.
 */
String.prototype.capitalize = function() {
    return this.charAt(0).toUpperCase() + this.slice(1);
};