yourpart3/frontend/scripts/audit-ceb-locale.mjs

#!/usr/bin/env node
/**
 * Vergleicht en- mit ceb-Locale-Chunks (wie in src/i18n/index.js).
 * - identical: Blattstrings, die nach Merge noch gleich en sind (typisch: fehlende Bisaya-Übersetzung).
 * - germanHints: Blattstrings in ceb mit deutschen Anzeichen (Umlaute / typische Wörter).
 *
 * Usage: node frontend/scripts/audit-ceb-locale.mjs
 */
import fs from 'fs';
import path from 'path';
import { fileURLToPath } from 'url';

const __dirname = path.dirname(fileURLToPath(import.meta.url));
const localesRoot = path.join(__dirname, '../src/i18n/locales');

const EN_CHUNKS = [
  'general.json',
  'header.json',
  'navigation.json',
  'home.json',
  'chat.json',
  'register.json',
  'passwordReset.json',
  'error.json',
  'activate.json',
  'settings.json',
  'admin.json',
  'socialnetwork.json',
  'friends.json',
  'falukant.json',
  'blog.json',
  'minigames.json',
  'message.json',
  'personal.json',
  'seo.json',
];

function isPlainObject(value) {
  return value !== null && typeof value === 'object' && !Array.isArray(value);
}

function deepMerge(target, source) {
  if (!isPlainObject(source)) return target;
  const base = isPlainObject(target) ? { ...target } : {};
  for (const key of Object.keys(source)) {
    const sv = source[key];
    const tv = base[key];
    if (isPlainObject(sv) && isPlainObject(tv)) {
      base[key] = deepMerge(tv, sv);
    } else if (isPlainObject(sv)) {
      base[key] = deepMerge({}, sv);
    } else {
      base[key] = sv;
    }
  }
  return base;
}

function loadJson(rel) {
  const p = path.join(localesRoot, rel);
  const raw = fs.readFileSync(p, 'utf8');
  return JSON.parse(raw);
}

function flatten(obj, prefix = '', out = {}) {
  if (!isPlainObject(obj)) return out;
  for (const k of Object.keys(obj)) {
    const v = obj[k];
    const key = prefix ? `${prefix}.${k}` : k;
    if (isPlainObject(v)) {
      flatten(v, key, out);
    } else if (typeof v === 'string') {
      out[key] = v;
    }
  }
  return out;
}

const germanWordHints =
  /\b(der|die|das|und|nicht|wird|werden|schon|keine|kein|Ziel|Ziele|Zug|Züge|Spiel|Level|Bitte|Hier|wurde|können|müssen)\b/i;
const germanChars = /[äöüÄÖÜß]/;

function audit() {
  let enMerged = {};
  let cebMerged = {};

  for (const f of EN_CHUNKS) {
    const enPath = path.join('en', f);
    const cebPath = path.join('ceb', f);
    if (!fs.existsSync(path.join(localesRoot, cebPath))) {
      console.warn(`Warnung: fehlt ${cebPath} (übersprungen)`);
      continue;
    }
    enMerged = deepMerge(enMerged, loadJson(enPath));
    cebMerged = deepMerge(cebMerged, loadJson(enPath));
    cebMerged = deepMerge(cebMerged, loadJson(cebPath));
  }

  const enFlat = flatten(enMerged);
  const cebFlat = flatten(cebMerged);

  const identical = [];
  const germanHints = [];

  for (const key of Object.keys(cebFlat)) {
    const ev = enFlat[key];
    const cv = cebFlat[key];
    if (typeof cv !== 'string') continue;
    if (ev === cv && typeof ev === 'string') {
      identical.push(key);
    }
    if (germanChars.test(cv) || germanWordHints.test(cv)) {
      germanHints.push({ key, sample: cv.slice(0, 120) });
    }
  }

  identical.sort();
  germanHints.sort((a, b) => a.key.localeCompare(b.key));

  const report = {
    generated: new Date().toISOString(),
    summary: {
      totalLeafStringsCeb: Object.keys(cebFlat).length,
      identicalToEnCount: identical.length,
      germanHintCount: germanHints.length,
    },
    identicalToEn: identical,
    germanHintsInCeb: germanHints,
  };

  const outFile = path.join(__dirname, '../ceb-locale-audit-report.json');
  fs.writeFileSync(outFile, JSON.stringify(report, null, 2), 'utf8');
  console.log(`Bericht geschrieben: ${outFile}`);
  console.log(
    `Zusammenfassung: ${report.summary.identicalToEnCount} Blätter noch identisch mit EN, ` +
      `${report.summary.germanHintCount} Strings mit DE-Hinweis (heuristisch).`
  );
}

audit();