i18n.js

import './vendor/polyfills/Array/prototype/indexOf'
import './vendor/polyfills/Object/keys'

/**
 * i18n support initialisation function
 *
 * @constructor
 * @param    {object}  options               - Options object.
 * @param    {string}  options.locale        - The locale to use, currently this is only used for pluralisation rules. If undefined, tries to use html[lang].
 * @param    {array}   options.separators    - An array of two items matching the start and end strings to search for when performing interpolation. Defaults to ['%{', '}']
 * @param    {object}  options.translations  - Key-value pairs of the translation strings to use.
 */
function i18n (options) {
  options = options || {}

  // Get user-defined locale setting,
  // Otherwise, try and detect it from the document,
  // Otherwise, set to English
  this.locale = options.locale || document.documentElement.lang || 'en'

  // Get user-defined interpolation separators,
  // Otherwise, use the same Ruby-like syntax as GOV.UK (e.g. %{keyName})
  // e.g. https://github.com/alphagov/govuk_publishing_components/blob/22944a9efaea3ae495a518717415694c71cfdb35/config/locales/en.yml
  this.separators = options.separators && options.separators.length === 2
    ? options.separators
    : ['%{', '}']

  // Flatten the list of translations
  this.translations = options.translations ? this.flatten(options.translations) : {}
}

/**
 * The most used function - looks up translation strings and sorts what
 * operations to perform on them.
 *
 * ALL properties passed to `options` will be exposed as options for string
 * interpolation, but only `count` and `fallback` initiate special functionality.
 *
 * TODO: Do we want to put string interpolation options in their own object (e.g. `options.data`) to avoid potential conflict in future?
 * TODO: Do we need some context/gendered forms support? Not sure if anything currently in Frontend necessitates this.
 *
 * @param    {string}  lookupKey         - The lookup key of the string to obtain
 * @param    {object}  options           - Options object.
 * @param    {number}  options.count     - Number used to determine which pluralisation to use.
 * @param    {string}  options.fallback  - Fallback string to use if a string associated with the key could not be found.
 * @param    {object}  options.fallback  - Object of key-value pairs, used if pluralisation support is needed in fallback text.
 * @returns  {string}                    - The formatted, localised string.
 */
i18n.prototype.t = function (lookupKey, options) {
  options = options || {}

  // Search the translations object for a matching key
  // Otherwise, use the fallback string
  // Otherwise, use the lookupKey itself as the ultimate fallback
  var outputString = this.translations[lookupKey] || options.fallback || lookupKey

  // If the `count` option is set, determine which plural suffix is needed
  // In order of priority:
  // 1. Pluralised lookup key provided in translations
  // 2. Plural object fallback text
  // 3. String fallback text
  if (typeof options.count !== 'undefined') {
    var pluralSuffix = this.getPluralSuffix(options.count)
    var pluralLookupKey = lookupKey + '_' + pluralSuffix
    if (this.translations[pluralLookupKey]) {
      outputString = this.translations[pluralLookupKey]
    } else if (typeof options.fallback === 'object') {
      outputString = options.fallback[pluralSuffix]
    } else if (typeof options.fallback === 'string') {
      outputString = options.fallback
    }
  }

  // Replace any placeholders in the string with the actual data
  outputString = this.replacePlaceholders(outputString, options)

  return outputString
}

/**
 * Ingests phrases containing placeholders and spits out the formatted phrase.
 *
 * @param    {string}  phrase          - The phrase containing the placeholder.
 * @param    {object}  placeholderMap  - Key-value pairs of placeholders and the values that replace them.
 */
i18n.prototype.replacePlaceholders = function (phrase, placeholderMap) {
  for (var key in placeholderMap) {
    phrase = phrase.replace(new RegExp(this.separators[0] + key + this.separators[1], 'g'), placeholderMap[key])
  }
  return phrase
}

/**
 * Recursively flattens nested objects, whilst maintaining the parent key name on the flattened list.
 *
 * For example, this structure:
 * {
 *   "hello": "Hello %{name}",
 *   "member_area": {
 *     "log_in": "Log in",
 *     "register": "Register account"
 *   }
 * }
 *
 * Will be flattened to this:
 * {
 *   "hello": "Hello %{name}",
 *   "member_area.log_in": "Log in",
 *   "member_area.register": "Register account"
 * }
 *
 * @param    {object} translationsObject  - A (potentially nested) object of translation key-value pairs.
 * @returns  {object}                     - A single level "flattened" object of translation key-value pairs.
 */
i18n.prototype.flatten = function (translationsObject) {
  var flattenedObject = {}
  var flattenLoop = function (translations, prefix) {
    var lookupKeys = Object.keys(translations)
    for (var i = 0; i < lookupKeys.length; i++) {
      var key = lookupKeys[i]
      var phrase = translations[key]
      var prefixedKey = prefix ? prefix + '.' + key : key
      if (typeof phrase === 'object') {
        flattenLoop(phrase, prefixedKey)
      } else {
        flattenedObject[prefixedKey] = phrase
      }
    }
  }
  flattenLoop(translationsObject)
  return flattenedObject
}

/**
 * Get the appropriate suffix for the plural form.
 *
 * The locale may include a regional indicator (such as en-GB), but we don't
 * usually care about this part, as pluralisation rules are usually the same
 * regardless of region. There are exceptions, however, (e.g. Portuguese) so
 * this searches by both the full and shortened locale codes, just to be sure.
 *
 * @param    {number}  count       - Number used to determine which pluralisation to use.
 * @returns  {string}              - The suffix associated with the correct pluralisation for this locale.
 */
i18n.prototype.getPluralSuffix = function (count) {
  var locale = this.locale
  var localeShort = (locale.length > 2) ? locale.substring(0, 2) : locale
  var keySuffix = 'other'

  for (var pluralRule in this.pluralRulesMap) {
    if (Object.prototype.hasOwnProperty.call(this.pluralRulesMap, pluralRule)) {
      var languages = this.pluralRulesMap[pluralRule]
      if (languages.indexOf(locale) > -1 || languages.indexOf(localeShort) > -1) {
        keySuffix = this.pluralRules[pluralRule](count)
        break
      }
    }
  }
  return keySuffix
}

/**
 * Different pluralisation rule sets
 *
 * Returns the appropriate suffix for the plural form associated with `n`.
 * Possible suffixes: 'zero', 'one', 'two', 'few', 'many', 'other' (the actual
 * meaning of each differs per locale). 'other' should always exist, even in
 * languages without plurals, such as Chinese.
 * https://unicode-org.github.io/cldr-staging/charts/latest/supplemental/language_plural_rules.html
 *
 * @param    {number}  n  - The `count` number being passed through. This must be a positive integer. Negative numbers and decimals aren't accounted for.
 * @returns  {string}     - The string that needs to be suffixed to the key (without separator).
 */
i18n.prototype.pluralRules = {
  arabic: function (n) {
    if (n === 0) { return 'zero' }
    if (n === 1) { return 'one' }
    if (n === 2) { return 'two' }
    if (n % 100 >= 3 && n % 100 <= 10) { return 'few' }
    if (n % 100 >= 11 && n % 100 <= 99) { return 'many' }
    return 'other'
  },
  belarusian: function (n) {
    var lastTwo = n % 100
    var last = lastTwo % 10
    if (last === 1 && lastTwo !== 11) { return 'one' }
    if (last >= 2 && last <= 4 && !(lastTwo >= 12 && lastTwo <= 14)) { return 'few' }
    if (last === 0 || (last >= 5 && last <= 9) || (lastTwo >= 11 && lastTwo <= 14)) { return 'many' }
    return 'other'
  },
  chinese: function () {
    return 'other'
  },
  cornish: function (n) {
    if (n === 1) { return 'one' }
    if (n === 2) { return 'two' }
    return 'other'
  },
  czech: function (n) {
    if (n === 1) { return 'one' }
    if (n >= 2 && n <= 4) { return 'few' }
    return 'other'
  },
  french: function (n) {
    return n === 0 || n === 1 ? 'one' : 'other'
  },
  german: function (n) {
    return n === 1 ? 'one' : 'other'
  },
  hebrew: function (n) {
    if (n === 1) { return 'one' }
    if (n === 2) { return 'two' }
    if (n % 10 === 0 && !(n >= 0 && n <= 10)) { return 'many' }
    return 'other'
  },
  icelandic: function (n) {
    return n % 10 === 1 && n % 100 !== 11 ? 'one' : 'other'
  },
  irish: function (n) {
    if (n === 1) { return 'one' }
    if (n === 2) { return 'two' }
    if (n >= 3 && n <= 6) { return 'few' }
    if (n >= 7 && n <= 10) { return 'many' }
    return 'other'
  },
  latvian: function (n) {
    if (n % 10 === 0 || (n % 100 >= 11 && n % 100 <= 19)) { return 'zero' }
    if (n % 10 === 1 && n % 100 !== 11) { return 'one' }
    return 'other'
  },
  lithuanian: function (n) {
    var lastTwo = n % 100
    var last = lastTwo % 10
    if (last === 1 && !(lastTwo >= 11 && lastTwo <= 19)) { return 'one' }
    if (last >= 2 && last <= 9 && !(lastTwo >= 11 && lastTwo <= 19)) { return 'few' }
    return 'other'
  },
  polish: function (n) {
    if (n === 1) { return 'one' }
    var lastTwo = n % 100
    var last = lastTwo % 10
    if (last >= 2 && last <= 4 && !(lastTwo >= 12 && lastTwo <= 14)) { return 'few' }
    if (last === 0 || last === 1 || (last >= 5 && last <= 9) || (lastTwo >= 12 && lastTwo <= 14)) { return 'many' }
    return 'other'
  },
  romanian: function (n) {
    if (n === 1) { return 'one' }
    if (n === 0 || (n % 100 >= 2 && n % 100 <= 19)) { return 'few' }
    return 'other'
  },
  russian: function (n) {
    var lastTwo = n % 100
    var last = lastTwo % 10
    if (last === 1 && lastTwo !== 11) { return 'one' }
    if (last >= 2 && last <= 4 && !(lastTwo >= 12 && lastTwo <= 14)) { return 'few' }
    if (last === 0 || (last >= 5 && last <= 9) || (lastTwo >= 11 && lastTwo <= 14)) { return 'many' }
    return 'other'
  },
  scottish: function (n) {
    if (n === 1 || n === 11) { return 'one' }
    if (n === 2 || n === 12) { return 'two' }
    if ((n >= 3 && n <= 10) || (n >= 13 && n <= 19)) { return 'few' }
    return 'other'
  },
  spanish: function (n) {
    if (n === 1) { return 'one' }
    if (n % 1000000 === 0 && n !== 0) { return 'many' }
    return 'other'
  },
  welsh: function (n) {
    if (n === 0) { return 'zero' }
    if (n === 1) { return 'one' }
    if (n === 2) { return 'two' }
    if (n === 3) { return 'few' }
    if (n === 6) { return 'many' }
    return 'other'
  }
}

/**
 * Map of plural rules to languages where those rules apply.
 *
 * Note: These groups are named for the most dominant or recognisable language
 * that uses each system. The groupings do not imply that the languages are
 * related to one another. Many languages have evolved the same systems
 * independently of one another.
 *
 * Spreadsheet of which languages require which plural forms:
 * https://docs.google.com/spreadsheets/d/1sFMhSX-ma7-_myYoRGh6DboHfN0xuN-1VRJ0ClVvNa8/edit
 *
 * Currently supported:
 * Arabic: Arabic (ar)
 * Belarusian: Belarusian (be)
 * Chinese: Burmese (my), Chinese (zh), Indonesian (id), Japanese (ja), Javanese (jv), Korean (ko), Malay (ms), Thai (th), Vietnamese (vi)
 * Cornish: Cornish (kw)
 * Czech: Czech (cs)
 * French: Armenian (hy), French (fr), Hindi (hi), Persian Farsi (fa), Punjabi (pa), Zulu (zu)
 * German: Afrikaans (af), Albanian (sq), Azerbaijani (az), Basque (eu), Bulgarian (bg), Catalan (ca), Danish (da), Dutch (nl), English (en), Estonian (et), Finnish (fi), Georgian (ka), German (de), Greek (el), Hungarian (hu), Luxembourgish (lb), Norwegian (no), Somali (so), Swahili (sw), Swedish (sv), Tamil (ta), Telugu (te), Turkish (tr), Urdu (ur)
 * Hebrew: Hebrew (he)
 * Icelandic: Icelandic (is), Macedonian (mk)
 * Irish: Irish Gaelic (ga)
 * Latvian: Latvian (lv)
 * Lithuanian: Lithuanian (lt)
 * Polish: Bosnian (bs), Croatian (hr), Polish (pl)
 * Romanian: Romanian (ro)
 * Russian: Russian (ru), Ukrainian (uk)
 * Scottish: Scottish Gaelic (gd)
 * Spanish: European Portuguese (pt-PT), Italian (it), Spanish (es)
 * Welsh: Welsh (cy)
 */
i18n.prototype.pluralRulesMap = {
  arabic: ['ar'],
  belarusian: ['be'],
  chinese: ['my', 'zh', 'id', 'ja', 'jv', 'ko', 'ms', 'th', 'vi'],
  cornish: ['kw'],
  czech: ['cs'],
  french: ['hy', 'fr', 'hi', 'fa', 'pa', 'zu'],
  german: ['af', 'sq', 'az', 'eu', 'bg', 'ca', 'da', 'nl', 'en', 'et', 'fi', 'ka', 'de', 'el', 'hu', 'lb', 'no', 'so', 'sw', 'sv', 'ta', 'te', 'tr', 'ur'],
  hebrew: ['he'],
  icelandic: ['is', 'mk'],
  irish: ['ga'],
  latvian: ['lv'],
  lithuanian: ['lt'],
  polish: ['bs', 'hr', 'pl'],
  romanian: ['ro'],
  russian: ['ru', 'uk'],
  scottish: ['gd'],
  spanish: ['pt-PT', 'it', 'es'],
  welsh: ['cy']
}

export default i18n