/**
 * Catalog checker. Runs before every build, and on its own via `pnpm check:i18n`.
 *
 * Four questions, all of which have bitten a multilingual site before:
 *
 *   missing    a key English has and this locale does not (falls back, but silently)
 *   unknown    a key a locale has and English does not (a typo, or a stale key)
 *   undefined  a `t('...')` in the source that no catalog defines (would render a
 *              humanised key on a real page)
 *   client     an island reaching for a key outside CLIENT_NAMESPACES, which would
 *              not be in the page's inline payload and would render as a humanised key
 *
 * Missing keys are a warning: English is a working fallback and a half-translated
 * language is better than no language. Everything else is an error, because every one
 * of them puts a wrong string in front of a visitor.
 *
 * Usage: node scripts/check-i18n.mjs [--strict]
 *   --strict also fails on missing keys, for a release build.
 */
import { readdir, readFile } from 'node:fs/promises';
import path from 'node:path';
import { fileURLToPath } from 'node:url';

const root = path.resolve(path.dirname(fileURLToPath(import.meta.url)), '..');
const i18nDir = path.join(root, 'src/i18n');
const srcDir = path.join(root, 'src');

const strict = process.argv.includes('--strict');

/* ---------- the locale table, read from config.ts ---------- */

/*
 * Parsed out of the source rather than imported, because config.ts is TypeScript and
 * this script runs on plain node before anything is compiled. The two shapes it reads
 * are the two the file is allowed to have; a change to either fails loudly here rather
 * than quietly checking nothing.
 */
const configSource = await readFile(path.join(i18nDir, 'config.ts'), 'utf8');

const localeCodes = [...configSource.matchAll(/\{\s*code:\s*'([a-z-]+)'/g)].map((m) => m[1]);
const defaultLocale = /DEFAULT_LOCALE\s*=\s*'([a-z-]+)'/.exec(configSource)?.[1];
const clientBlock = /CLIENT_NAMESPACES\s*=\s*\[([^\]]*)\]/s.exec(configSource)?.[1] ?? '';
const clientNamespaces = new Set([...clientBlock.matchAll(/'([\w-]+)'/g)].map((m) => m[1]));

if (localeCodes.length === 0 || !defaultLocale || clientNamespaces.size === 0) {
  console.error('check-i18n: could not read LOCALES, DEFAULT_LOCALE or CLIENT_NAMESPACES from src/i18n/config.ts');
  process.exit(2);
}

/*
 * `locales.mjs` is the same list in plain JavaScript, for `astro.config.mjs` and this
 * script, both of which run before TypeScript exists. Two lists is one more than
 * ideal, so they are compared here rather than trusted.
 */
const { DEFAULT_LOCALE: mjsDefault, LOCALE_CODES: mjsCodes } = await import(
  path.join(i18nDir, 'locales.mjs')
);

if (mjsDefault !== defaultLocale || mjsCodes.join(',') !== localeCodes.join(',')) {
  console.error('check-i18n: src/i18n/config.ts and src/i18n/locales.mjs disagree.');
  console.error(`  config.ts:   default ${defaultLocale}, locales [${localeCodes.join(', ')}]`);
  console.error(`  locales.mjs: default ${mjsDefault}, locales [${mjsCodes.join(', ')}]`);
  console.error('Both need the new language. See README, "Adding a language".');
  process.exit(2);
}

/* ---------- catalogs ---------- */

const catalogs = new Map();
for (const code of localeCodes) {
  const file = path.join(i18nDir, `${code}.json`);
  try {
    catalogs.set(code, JSON.parse(await readFile(file, 'utf8')));
  } catch (error) {
    console.error(`check-i18n: cannot read src/i18n/${code}.json (${error.message})`);
    console.error(`Every locale in LOCALES needs a catalog. Copy en.json to ${code}.json and translate it.`);
    process.exit(2);
  }
}

const base = catalogs.get(defaultLocale);

/** CLDR plural categories, which are a key's suffix rather than a key of their own. */
const PLURAL_SUFFIX = /\.(zero|one|two|few|many|other)$/;

/** `card.reviews.one` and `card.reviews` are the same key as far as coverage goes. */
const stem = (key) => key.replace(PLURAL_SUFFIX, '');

/**
 * Every stem a catalog can resolve.
 *
 * Stems rather than literal keys, because a locale is allowed more plural categories
 * than English has: Spanish spells `time.ago.month` as `.one` and `.other` where
 * English needs one form, and neither is missing anything.
 */
const stems = (catalog) => new Set(Object.keys(catalog).map(stem));

const baseStems = stems(base);

const problems = [];
const warnings = [];

for (const code of localeCodes) {
  if (code === defaultLocale) continue;
  const catalog = catalogs.get(code);
  const own = stems(catalog);

  const missing = [...baseStems].filter((key) => !own.has(key)).sort();
  const unknown = [...own].filter((key) => !baseStems.has(key)).sort();

  if (missing.length > 0) {
    (strict ? problems : warnings).push({
      code,
      kind: 'missing',
      keys: missing,
      note: `${missing.length} key(s) fall back to ${defaultLocale}`,
    });
  }
  if (unknown.length > 0) {
    problems.push({
      code,
      kind: 'unknown',
      keys: unknown,
      note: `${unknown.length} key(s) are not in ${defaultLocale}.json, so nothing reads them`,
    });
  }
}

/* ---------- the warning copy, which lives in two packages ---------- */

/*
 * `shared/src/warnings.ts` decides which banner a mint gets and writes the English
 * sentence for it, because the API's fixture tests assert on that sentence. The
 * catalogs carry the same sentences so they can be translated. Two copies of safety
 * copy is exactly the sort of thing that drifts, and a mint page and a mint card
 * disagreeing about whether withdrawals work is the worst possible thing to get wrong,
 * so the two are diffed here on every build.
 *
 * Read from shared's build output rather than its source: it is a prerequisite of the
 * web build, so it is always there, and importing it beats parsing TypeScript.
 */
const SHARED_TABLES = [
  { module: 'WARNING_COPY_EN', prefix: 'mint.warnings.' },
  { module: 'CHIP_COPY_EN', prefix: 'mint.chip.' },
];

try {
  const shared = await import(path.join(root, '../shared/dist/warnings.js'));

  for (const { module, prefix } of SHARED_TABLES) {
    const table = shared[module];
    if (!table) {
      problems.push({
        code: defaultLocale,
        kind: 'shared',
        keys: [module],
        note: `shared/dist/warnings.js no longer exports ${module}`,
      });
      continue;
    }

    const drift = [];
    for (const [key, english] of Object.entries(table)) {
      const catalogKey = prefix + key;
      const inCatalog = base[catalogKey];
      if (inCatalog === undefined) drift.push(`${catalogKey}  (in ${module}, not in ${defaultLocale}.json)`);
      else if (inCatalog !== english) {
        drift.push(`${catalogKey}\n      shared:  ${english}\n      catalog: ${inCatalog}`);
      }
    }
    // And the other direction: a catalog key under this prefix that shared never writes
    // is a key no banner will ever ask for.
    for (const key of Object.keys(base)) {
      if (!key.startsWith(prefix)) continue;
      if (table[key.slice(prefix.length)] === undefined) {
        drift.push(`${key}  (in ${defaultLocale}.json, not in ${module})`);
      }
    }

    if (drift.length > 0) {
      problems.push({
        code: defaultLocale,
        kind: 'shared',
        keys: drift,
        note: `${drift.length} difference(s) between shared's ${module} and ${defaultLocale}.json`,
      });
    }
  }
} catch (error) {
  problems.push({
    code: defaultLocale,
    kind: 'shared',
    keys: [String(error.message)],
    note: 'could not read shared/dist/warnings.js. Run `pnpm --filter ./shared build` first',
  });
}

/* ---------- key usage in the source ---------- */

async function walk(dir) {
  const out = [];
  for (const entry of await readdir(dir, { withFileTypes: true })) {
    const full = path.join(dir, entry.name);
    if (entry.isDirectory()) {
      if (entry.name === 'i18n' || entry.name === 'bones') continue;
      out.push(...(await walk(full)));
    } else if (/\.(astro|ts)$/.test(entry.name)) {
      out.push(full);
    }
  }
  return out;
}

/**
 * `t('some.key')` and `t('some.key', { ... })`, with either quote.
 *
 * Only literal keys are found, which is the point: a key built at runtime cannot be
 * checked, so the codebase does not build them. The two places that legitimately need
 * a computed key (`nut.${n}` and `status.${mint.status}`) use a helper with its own
 * exhaustive list, declared below.
 */
const T_CALL = /\bt\(\s*(['"])([\w.-]+)\1/g;

/**
 * Key families built from a value at runtime, listed here so the checker knows they
 * are reached. Each entry is a prefix; every catalog key under it counts as used.
 */
const DYNAMIC_PREFIXES = [
  'nut.',                    // nutName(n)
  'status.',                 // statusLabel(mint.status)
  'reviews.dialog.stars.',   // the star picker's phrase per rating
  'reviews.breakdown.',      // ratingBreakdown, one key per star count
  'mint.warnings.',          // warningStrings(), keyed by kind
  'mint.chip.',              // mintChip()
  'login.method.',           // methodChipHtml(session.method)
  'wallets.platform.',       // the platform list on each wallet card
  'wallets.item.',           // one blurb per wallet, keyed by its id
  'login.connect.',          // CONNECT_COPY, keyed by method
  'time.ago.',               // formatRelative, keyed by unit
  'time.short.',             // formatShortDuration, keyed by unit
];

const isDynamic = (key) => DYNAMIC_PREFIXES.some((prefix) => key.startsWith(prefix));

const files = await walk(srcDir);
const used = new Set();
const undefinedKeys = [];
const clientLeaks = [];

for (const file of files) {
  const source = await readFile(file, 'utf8');
  const relative = path.relative(root, file);

  /*
   * An .astro file is two programs: frontmatter and template render at build time, and
   * anything inside a <script> ships to the browser. Only the second half is limited to
   * the client namespaces, so the two are checked separately.
   */
  const scripts = [...source.matchAll(/<script(?![^>]*\bis:inline\b)[^>]*>([\s\S]*?)<\/script>/g)]
    .map((m) => m[1])
    .join('\n');
  // A .ts file under lib/ or scripts/ is island code wholesale.
  const shipsToBrowser = /\/(lib|scripts)\//.test(relative) && relative.endsWith('.ts');

  for (const match of source.matchAll(T_CALL)) {
    const key = match[2];
    used.add(key);
    if (!baseStems.has(stem(key))) {
      undefinedKeys.push({ key, file: relative });
    }
  }

  const clientSource = shipsToBrowser ? source : scripts;
  for (const match of clientSource.matchAll(T_CALL)) {
    const key = match[2];
    const namespace = key.split('.')[0];
    if (!clientNamespaces.has(namespace)) {
      clientLeaks.push({ key, file: relative, namespace });
    }
  }
}

if (undefinedKeys.length > 0) {
  problems.push({
    code: defaultLocale,
    kind: 'undefined',
    keys: undefinedKeys.map((u) => `${u.key}  (${u.file})`),
    note: `${undefinedKeys.length} key(s) used in the source are not defined in ${defaultLocale}.json`,
  });
}

if (clientLeaks.length > 0) {
  problems.push({
    code: '-',
    kind: 'client',
    keys: clientLeaks.map((c) => `${c.key}  (${c.file}, namespace "${c.namespace}")`),
    note: `${clientLeaks.length} island key(s) outside CLIENT_NAMESPACES: add the namespace to config.ts or move the string`,
  });
}

/* ---------- keys nothing reads ---------- */

const unused = [...baseStems]
  .filter((key) => !used.has(key) && !isDynamic(key))
  .filter((key) => !base[key] === false || true)
  .sort();

/* ---------- report ---------- */

const report = (list, level) => {
  for (const entry of list) {
    console.log(`\n${level} [${entry.kind}] ${entry.code}: ${entry.note}`);
    for (const key of entry.keys.slice(0, 40)) console.log(`    ${key}`);
    if (entry.keys.length > 40) console.log(`    … and ${entry.keys.length - 40} more`);
  }
};

const total = Object.keys(base).length;
console.log(
  `check-i18n: ${localeCodes.length} locales, ${total} keys in ${defaultLocale}.json, ` +
    `${used.size} literal key(s) used in src/.`,
);

if (process.env['LIST_UNUSED'] && unused.length > 0) {
  console.log(`\n  ${unused.length} key(s) with no literal or dynamic use:`);
  for (const key of unused) console.log(`    ${key}`);
}

report(warnings, 'WARN ');
report(problems, 'ERROR');

if (problems.length > 0) {
  console.log(`\ncheck-i18n failed with ${problems.length} problem group(s).`);
  process.exit(1);
}

if (warnings.length > 0) {
  console.log('\ncheck-i18n passed with warnings. Untranslated keys fall back to English.');
} else {
  console.log('Every locale is complete and every key used in the source is defined.');
}