tractatus/scripts/validate-i18n.js

#!/usr/bin/env node
/**
 * I18n Validation Script
 * Ensures all data-i18n keys in HTML files have corresponding translations
 * in all locale JSON files
 *
 * Usage: node scripts/validate-i18n.js [--verbose]
 * Exit codes:
 *   0 - All i18n keys validated successfully
 *   1 - Missing translations found
 *   2 - Script error
 */

const fs = require('fs');
const path = require('path');

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

// Configuration
const PUBLIC_DIR = path.join(__dirname, '../public');
const LOCALES_DIR = path.join(PUBLIC_DIR, 'locales');
const SUPPORTED_LOCALES = ['en', 'de', 'fr'];
const HTML_FILES = [
  'index.html',
  'researcher.html',
  'implementer.html',
  'leader.html',
  'docs.html',
  'faq.html',
  'about.html'
];

// Extract all data-i18n keys from HTML files
function extractI18nKeys(htmlContent) {
  const keys = new Set();
  const regex = /data-i18n="([^"]+)"/g;
  const regexHtml = /data-i18n-html="([^"]+)"/g;

  let match;
  while ((match = regex.exec(htmlContent)) !== null) {
    keys.add(match[1]);
  }
  while ((match = regexHtml.exec(htmlContent)) !== null) {
    keys.add(match[1]);
  }

  return Array.from(keys);
}

// Check if a key exists in a nested JSON object
function keyExists(obj, keyPath) {
  const parts = keyPath.split('.');
  let current = obj;

  for (const part of parts) {
    if (!current || typeof current !== 'object' || !(part in current)) {
      return false;
    }
    current = current[part];
  }

  return true;
}

// Main validation
function validateI18n() {
  console.log('🌍 Validating i18n translations...\n');

  let totalKeys = 0;
  let missingTranslations = [];

  // Process each HTML file
  for (const htmlFile of HTML_FILES) {
    const htmlPath = path.join(PUBLIC_DIR, htmlFile);

    if (!fs.existsSync(htmlPath)) {
      if (verbose) console.log(`⚠️  Skipping ${htmlFile} (not found)`);
      continue;
    }

    const htmlContent = fs.readFileSync(htmlPath, 'utf8');
    const keys = extractI18nKeys(htmlContent);

    if (keys.length === 0) {
      if (verbose) console.log(`ℹ️  ${htmlFile}: No i18n keys found`);
      continue;
    }

    console.log(`📄 ${htmlFile}: Found ${keys.length} i18n keys`);
    totalKeys += keys.length;

    // Check each key in all locales
    for (const key of keys) {
      for (const locale of SUPPORTED_LOCALES) {
        // Determine which JSON file to check
        // homepage.html -> homepage.json
        // index.html keys (hero, community, etc) -> homepage.json
        let jsonFile;
        if (htmlFile === 'index.html') {
          jsonFile = 'homepage.json';
        } else {
          const baseName = htmlFile.replace('.html', '');
          jsonFile = `${baseName}.json`;
        }

        const localePath = path.join(LOCALES_DIR, locale, jsonFile);

        if (!fs.existsSync(localePath)) {
          // Try common.json as fallback
          const fallbackPath = path.join(LOCALES_DIR, locale, 'common.json');
          if (fs.existsSync(fallbackPath)) {
            const translations = JSON.parse(fs.readFileSync(fallbackPath, 'utf8'));
            if (!keyExists(translations, key)) {
              missingTranslations.push({
                file: htmlFile,
                key: key,
                locale: locale,
                issue: `Key not found in ${jsonFile} or common.json`
              });
            }
          } else {
            missingTranslations.push({
              file: htmlFile,
              key: key,
              locale: locale,
              issue: `Translation file not found: ${jsonFile}`
            });
          }
          continue;
        }

        const translations = JSON.parse(fs.readFileSync(localePath, 'utf8'));

        if (!keyExists(translations, key)) {
          missingTranslations.push({
            file: htmlFile,
            key: key,
            locale: locale,
            issue: 'Key not found in translation file'
          });
        }
      }
    }
  }

  // Report results
  console.log(`\n📊 Validation Summary:`);
  console.log(`   Total i18n keys: ${totalKeys}`);
  console.log(`   Locales checked: ${SUPPORTED_LOCALES.join(', ')}`);

  if (missingTranslations.length === 0) {
    console.log(`\n✅ All i18n keys have translations in all locales!\n`);
    return 0;
  }

  console.log(`\n❌ Found ${missingTranslations.length} missing translations:\n`);

  // Group by file and locale
  const byFile = {};
  for (const item of missingTranslations) {
    if (!byFile[item.file]) byFile[item.file] = {};
    if (!byFile[item.file][item.locale]) byFile[item.file][item.locale] = [];
    byFile[item.file][item.locale].push(item);
  }

  for (const [file, locales] of Object.entries(byFile)) {
    console.log(`📄 ${file}:`);
    for (const [locale, items] of Object.entries(locales)) {
      console.log(`   ${locale}: ${items.length} missing`);
      for (const item of items) {
        console.log(`      - ${item.key}: ${item.issue}`);
      }
    }
    console.log('');
  }

  console.log('⚠️  Please add missing translations before deploying.\n');
  return 1;
}

// Run validation
try {
  const exitCode = validateI18n();
  process.exit(exitCode);
} catch (error) {
  console.error('❌ Validation script error:', error.message);
  if (verbose) console.error(error.stack);
  process.exit(2);
}