cinny/src/app/utils/translation/langUtils.ts

// Pure language-code helpers for the on-device translation feature. No DOM / no
// engine — unit-testable in isolation (langUtils.test.ts).

/**
 * Normalize a BCP-47 tag to the base language subtag, lowercased:
 *   "en-US" -> "en", "ZH" -> "zh", "pt-BR" -> "pt".
 * Chinese is a common exception where script matters (zh-Hant vs zh-Hans); we
 * keep the script subtag for `zh` so a Traditional/Simplified distinction isn't
 * flattened. Everything else collapses to the primary subtag.
 */
export function normalizeLang(code: string | undefined | null): string {
  if (!code) return '';
  const lower = code.trim().toLowerCase();
  if (!lower) return '';
  const parts = lower.split(/[-_]/);
  const base = parts[0];
  if (base === 'zh' && (parts[1] === 'hant' || parts[1] === 'hans')) {
    return `zh-${parts[1]}`;
  }
  return base;
}

/** True when two language tags refer to the same base language. */
export function sameLanguage(a: string | undefined | null, b: string | undefined | null): boolean {
  const na = normalizeLang(a);
  const nb = normalizeLang(b);
  return na !== '' && na === nb;
}

/**
 * Curated list of target languages offered in the "Translate messages into"
 * setting — the common languages the on-device engine can produce. Codes are the
 * normalized base tags passed to the engine.
 */
export const TRANSLATE_TARGET_LANGUAGES: ReadonlyArray<{ code: string; name: string }> = [
  { code: 'en', name: 'English' },
  { code: 'es', name: 'Spanish' },
  { code: 'fr', name: 'French' },
  { code: 'de', name: 'German' },
  { code: 'it', name: 'Italian' },
  { code: 'pt', name: 'Portuguese' },
  { code: 'nl', name: 'Dutch' },
  { code: 'pl', name: 'Polish' },
  { code: 'ru', name: 'Russian' },
  { code: 'uk', name: 'Ukrainian' },
  { code: 'tr', name: 'Turkish' },
  { code: 'ar', name: 'Arabic' },
  { code: 'he', name: 'Hebrew' },
  { code: 'hi', name: 'Hindi' },
  { code: 'bn', name: 'Bengali' },
  { code: 'ja', name: 'Japanese' },
  { code: 'ko', name: 'Korean' },
  { code: 'zh', name: 'Chinese' },
  { code: 'vi', name: 'Vietnamese' },
  { code: 'th', name: 'Thai' },
  { code: 'id', name: 'Indonesian' },
  { code: 'sv', name: 'Swedish' },
  { code: 'cs', name: 'Czech' },
  { code: 'ro', name: 'Romanian' },
  { code: 'el', name: 'Greek' },
  { code: 'fa', name: 'Persian' },
];

const CURATED_NAME = new Map(TRANSLATE_TARGET_LANGUAGES.map((l) => [l.code, l.name]));

/**
 * Human-readable language name for a code, e.g. "de" -> "German". Prefers
 * `Intl.DisplayNames` (localized) and falls back to the curated map, then the
 * raw code. `uiLocale` picks the display locale for the name.
 */
export function languageName(code: string, uiLocale = 'en'): string {
  const norm = normalizeLang(code);
  if (!norm) return code;
  try {
    const dn = new Intl.DisplayNames([uiLocale], { type: 'language' });
    const name = dn.of(norm);
    if (name && name.toLowerCase() !== norm) return name;
  } catch {
    /* Intl.DisplayNames unsupported or bad code — fall through */
  }
  return CURATED_NAME.get(norm) ?? norm;
}

/** Is `code` in the curated target-language list? (validates a persisted setting) */
export function isSupportedTargetLang(code: string): boolean {
  return CURATED_NAME.has(normalizeLang(code));
}

/**
 * [Gitea #349] Why translation is unavailable, by platform. The desktop app
 * on Linux/macOS (WebKit) lacks the on-device APIs too, so it must not send
 * people to "the Lotus desktop app" they're already using.
 */
export const translationUnavailableText = (desktopApp: boolean): string =>
  desktopApp
    ? 'On-device translation isn’t available in this version of the desktop app. It works in Chrome or Edge 138+ on a computer.'
    : 'On-device translation isn’t available in this browser. Use Chrome or Edge 138+ on a computer.';

/** Pure: why a manual Translate did nothing (#287). */
export const translateSkipMessage = (from: string | undefined, targetLang: string): string =>
  from
    ? `This message is already in ${languageName(targetLang) || targetLang}.`
    : "Couldn't tell what language this message is in, so it wasn't translated.";