manoon-headless/scripts/generate-urls.js

const oils = require('../data/taxonomy/oils.json');
const concerns = require('../data/taxonomy/concerns.json');

const LOCALES = ['sr', 'en', 'de', 'fr'];
const DEFAULT_LOCALE = 'sr';

function generateUrl(oilId, concernId, locale) {
  const oil = oils.oils[oilId];
  const concern = concerns.concerns[concernId];
  
  if (!oil || !concern) return null;
  
  const localePrefix = locale === DEFAULT_LOCALE ? '' : `/${locale}`;
  const oilSlug = oil.slug[locale];
  const concernSlug = concern.slug[locale];
  
  return {
    url: `${localePrefix}/solutions/${oilSlug}-for-${concernSlug}`,
    canonical: `https://manoonoils.com${localePrefix}/solutions/${oilSlug}-for-${concernSlug}`,
    locale,
    oil: {
      id: oilId,
      name: oil.name[locale],
      slug: oilSlug
    },
    concern: {
      id: concernId,
      name: concern.name[locale],
      slug: concernSlug
    }
  };
}

function generateAllUrls() {
  const urls = [];
  const stats = {
    total: 0,
    byLocale: {},
    byOil: {},
    byConcern: {}
  };

  LOCALES.forEach(l => stats.byLocale[l] = 0);
  Object.keys(oils.oils).forEach(o => stats.byOil[o] = 0);
  Object.keys(concerns.concerns).forEach(c => stats.byConcern[c] = 0);

  for (const oilId of Object.keys(oils.oils)) {
    const oil = oils.oils[oilId];
    
    for (const concernId of oil.concerns) {
      const concern = concerns.concerns[concernId];
      
      if (!concern) {
        console.warn(`Warning: Concern ${concernId} not found for oil ${oilId}`);
        continue;
      }
      
      for (const locale of LOCALES) {
        const urlData = generateUrl(oilId, concernId, locale);
        if (urlData) {
          urls.push(urlData);
          stats.total++;
          stats.byLocale[locale]++;
          stats.byOil[oilId]++;
          stats.byConcern[concernId]++;
        }
      }
    }
  }

  return { urls, stats };
}

function generateSitemap() {
  const { urls } = generateAllUrls();
  
  let sitemap = '<?xml version="1.0" encoding="UTF-8"?>\n';
  sitemap += '<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">\n';
  
  for (const url of urls) {
    sitemap += '  <url>\n';
    sitemap += `    <loc>${url.canonical}</loc>\n`;
    sitemap += '    <changefreq>weekly</changefreq>\n';
    sitemap += '    <priority>0.8</priority>\n';
    sitemap += `    <xhtml:link rel="alternate" hreflang="${url.locale}" href="${url.canonical}" />\n`;
    sitemap += '  </url>\n';
  }
  
  sitemap += '</urlset>';
  return sitemap;
}

function generateUrlReport() {
  const { urls, stats } = generateAllUrls();
  
  let report = '# Programmatic SEO URL Report\n\n';
  report += `Generated: ${new Date().toISOString()}\n\n`;
  
  report += '## Summary\n\n';
  report += `- **Total URLs**: ${stats.total}\n`;
  report += `- **Languages**: ${LOCALES.join(', ')}\n`;
  report += `- **Oils**: ${Object.keys(oils.oils).length}\n`;
  report += `- **Concerns**: ${Object.keys(concerns.concerns).length}\n\n`;
  
  report += '## URLs by Locale\n\n';
  for (const [locale, count] of Object.entries(stats.byLocale)) {
    report += `- **${locale.toUpperCase()}**: ${count} URLs\n`;
  }
  report += '\n';
  
  report += '## URLs by Oil\n\n';
  for (const [oilId, count] of Object.entries(stats.byOil)) {
    const oil = oils.oils[oilId];
    report += `- **${oil.name.en}** (${oilId}): ${count} URLs\n`;
  }
  report += '\n';
  
  report += '## URLs by Concern\n\n';
  for (const [concernId, count] of Object.entries(stats.byConcern)) {
    const concern = concerns.concerns[concernId];
    report += `- **${concern.name.en}** (${concernId}): ${count} URLs\n`;
  }
  report += '\n';
  
  report += '## All Generated URLs\n\n';
  
  const urlsByOil = {};
  for (const url of urls) {
    if (!urlsByOil[url.oil.id]) {
      urlsByOil[url.oil.id] = [];
    }
    urlsByOil[url.oil.id].push(url);
  }
  
  for (const [oilId, oilUrls] of Object.entries(urlsByOil)) {
    const oil = oils.oils[oilId];
    report += `### ${oil.name.en}\n\n`;
    
    const byConcern = {};
    for (const url of oilUrls) {
      if (!byConcern[url.concern.id]) {
        byConcern[url.concern.id] = [];
      }
      byConcern[url.concern.id].push(url);
    }
    
    for (const [concernId, concernUrls] of Object.entries(byConcern)) {
      const concern = concerns.concerns[concernId];
      report += `#### ${concern.name.en}\n\n`;
      
      for (const url of concernUrls) {
        report += `- ${url.locale.toUpperCase()}: \`${url.canonical}\`\n`;
      }
      report += '\n';
    }
  }
  
  return report;
}

module.exports = {
  generateUrl,
  generateAllUrls,
  generateSitemap,
  generateUrlReport,
  LOCALES,
  DEFAULT_LOCALE
};

if (require.main === module) {
  const { urls, stats } = generateAllUrls();
  
  console.log('\n=== PROGRAMMATIC SEO URL GENERATOR ===\n');
  console.log(`Total URLs Generated: ${stats.total}`);
  console.log('\nBy Locale:');
  for (const [locale, count] of Object.entries(stats.byLocale)) {
    console.log(`  ${locale.toUpperCase()}: ${count}`);
  }
  console.log('\nBy Oil:');
  for (const [oilId, count] of Object.entries(stats.byOil)) {
    const oil = oils.oils[oilId];
    console.log(`  ${oil.name.en}: ${count}`);
  }
  console.log('\nBy Concern:');
  for (const [concernId, count] of Object.entries(stats.byConcern)) {
    const concern = concerns.concerns[concernId];
    console.log(`  ${concern.name.en}: ${count}`);
  }
  
  console.log('\n=== SAMPLE URLS ===\n');
  const sampleUrls = urls.filter((_, i) => i < 12);
  for (const url of sampleUrls) {
    console.log(`${url.locale.toUpperCase()}: ${url.canonical}`);
  }
  
  const fs = require('fs');
  const report = generateUrlReport();
  fs.writeFileSync('./url-report.md', report);
  console.log('\n✓ Full report saved to: url-report.md');
  
  const sitemap = generateSitemap();
  fs.writeFileSync('./sitemap-programmatic.xml', sitemap);
  console.log('✓ Sitemap saved to: sitemap-programmatic.xml');
}