// Domain helpers shared by the import script (plain node), the Nitro server routes
// and the browser form. Plain ESM on purpose so `node scripts/*.mjs` can import it
// directly — do not convert this file to TypeScript.

/**
 * Second-level suffixes that are NOT registrable on their own. Kept deliberately small:
 * a full Public Suffix List is overkill for German hospital domains, but these show up.
 */
const SECOND_LEVEL_SUFFIXES = new Set([
  'co', 'com', 'org', 'net', 'gov', 'edu', 'ac', 'or', 'ne', 'go', 'mil', 'nom',
])

/**
 * Freemail / consumer mailbox providers. A hospital account must never be registered
 * with one of these — the address would prove nothing about the institution.
 */
export const FREEMAIL_DOMAINS = new Set([
  'gmail.com', 'googlemail.com',
  'gmx.de', 'gmx.net', 'gmx.at', 'gmx.ch', 'gmx.com', 'gmx.info', 'gmx.biz',
  'web.de', 'freenet.de', 'arcor.de', 'online.de', 'email.de', 'mail.de',
  't-online.de', 'telekom.de', 'magenta.de',
  'outlook.com', 'outlook.de', 'outlook.at', 'hotmail.com', 'hotmail.de', 'live.com',
  'live.de', 'live.at', 'msn.com', 'windowslive.com',
  'yahoo.com', 'yahoo.de', 'ymail.com', 'rocketmail.com',
  'aol.com', 'aol.de',
  'icloud.com', 'me.com', 'mac.com',
  'posteo.de', 'posteo.net', 'mailbox.org', 'protonmail.com', 'protonmail.ch', 'proton.me',
  'tutanota.com', 'tutanota.de', 'tuta.io', 'tuta.com',
  'zoho.com', 'yandex.com', 'yandex.ru', 'mail.ru', 'gmx.us',
  'firemail.de', 'unitybox.de', 'unity-mail.de', 'kabelmail.de', 'vodafone.de',
  'ewetel.net', 'nexgo.de', 'inbox.com', 'fastmail.com', 'hushmail.com',
  'trashmail.com', 'trashmail.de', 'mailinator.com', 'guerrillamail.com',
  '10minutemail.com', 'wegwerfmail.de', 'temp-mail.org', 'yopmail.com',
])

/** Lowercase + strip whitespace/trailing dots. */
const clean = (value) => String(value ?? '').trim().toLowerCase().replace(/\.+$/, '')

/**
 * Host part of a URL (or of a bare "www.example.de/pfad" string), without www. prefix.
 * Returns '' when nothing usable is found.
 */
export function hostFromUrl(url) {
  let raw = clean(url)
  if (!raw) return ''
  raw = raw.replace(/^[a-z][a-z0-9+.-]*:\/\//, '')  // strip scheme
  raw = raw.replace(/^[^@/]*@/, '')                 // strip user:pass@
  raw = raw.split(/[/?#]/)[0] || ''                 // strip path/query/hash
  raw = raw.split(':')[0] || ''                     // strip port
  raw = raw.replace(/^www\d*\./, '')                // strip www. / www2.
  if (!raw || !raw.includes('.')) return ''
  if (!/^[a-z0-9.-]+$/.test(raw)) return ''
  return raw.replace(/\.+$/, '')
}

/** Domain part of an e-mail address (everything after the last @), without www. */
export function hostFromEmail(email) {
  const raw = clean(email)
  const at = raw.lastIndexOf('@')
  if (at < 0) return ''
  return hostFromUrl(raw.slice(at + 1))
}

/**
 * Registrable domain ("eTLD+1") of a host: klinik.helios-gesundheit.de → helios-gesundheit.de.
 * Collapsing subdomains is intentional — a hospital group shares ONE registrable domain
 * across many Standorte, and that is exactly the key we match registrations against.
 */
export function registrableDomain(host) {
  const h = hostFromUrl(host)
  if (!h) return ''
  const parts = h.split('.').filter(Boolean)
  if (parts.length < 2) return ''
  const last = parts[parts.length - 1]
  const secondLast = parts[parts.length - 2]
  // e.g. example.co.uk / example.com.de → keep three labels
  if (parts.length >= 3 && last.length === 2 && SECOND_LEVEL_SUFFIXES.has(secondLast)) {
    return parts.slice(-3).join('.')
  }
  return parts.slice(-2).join('.')
}

/** Registrable domain derived from a URL string. */
export const domainFromUrl = (url) => registrableDomain(hostFromUrl(url))

/** Registrable domain derived from an e-mail address. */
export const domainFromEmail = (email) => registrableDomain(hostFromEmail(email))

/** True when the address belongs to a consumer mailbox provider. */
export function isFreemailDomain(domain) {
  const d = registrableDomain(domain)
  if (!d) return false
  return FREEMAIL_DOMAINS.has(d) || FREEMAIL_DOMAINS.has(clean(domain))
}

/** Syntactic e-mail check — deliberately permissive, the confirmation mail is the real proof. */
export function isValidEmail(email) {
  const raw = clean(email)
  if (!raw || raw.length > 254) return false
  return /^[^\s@,;:<>()[\]\\"]+@[a-z0-9]([a-z0-9-]*[a-z0-9])?(\.[a-z0-9]([a-z0-9-]*[a-z0-9])?)+$/.test(raw)
}

/**
 * German umlaut/ß transliteration plus a diacritic-free variant, so that both
 * "München" and "Muenchen"/"Munchen" find the same Standort in the FTS index.
 */
export function searchVariants(text) {
  const base = String(text ?? '').toLowerCase()
  if (!base) return []
  const expanded = base
    .replace(/ä/g, 'ae').replace(/ö/g, 'oe').replace(/ü/g, 'ue')
    .replace(/ß/g, 'ss').replace(/é|è|ê/g, 'e').replace(/á|à|â/g, 'a')
  const stripped = base
    .replace(/ä/g, 'a').replace(/ö/g, 'o').replace(/ü/g, 'u').replace(/ß/g, 's')
    .normalize('NFD').replace(/[̀-ͯ]/g, '')
  const out = new Set([base])
  if (expanded !== base) out.add(expanded)
  if (stripped !== base) out.add(stripped)
  return [...out]
}

export default {
  FREEMAIL_DOMAINS,
  hostFromUrl,
  hostFromEmail,
  registrableDomain,
  domainFromUrl,
  domainFromEmail,
  isFreemailDomain,
  isValidEmail,
  searchVariants,
}