klz-cables.com/scripts/fetch-missing-media.js

#!/usr/bin/env node

/**
 * Script to fetch specific missing media IDs from WordPress
 * Uses the WordPress REST API to get media URLs and download them
 */

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

// Load environment variables
require('dotenv').config();

const BASE_URL = process.env.WOOCOMMERCE_URL;
const APP_PASSWORD = process.env.WORDPRESS_APP_PASSWORD;

// Validate environment
if (!BASE_URL || !APP_PASSWORD) {
  console.error('❌ Missing required environment variables');
  console.error('Please check .env file for:');
  console.error('  - WOOCOMMERCE_URL');
  console.error('  - WORDPRESS_APP_PASSWORD');
  process.exit(1);
}

// Configuration
const MISSING_MEDIA_IDS = [10432, 10440, 10382, 10616, 10615, 45569, 10638, 5767];
const MEDIA_DIR = path.join(__dirname, '..', 'public', 'media');
const RAW_DATA_DIR = path.join(__dirname, '..', 'data', 'raw', '2025-12-30T15-21-49-331Z');

// Create media directory if it doesn't exist
if (!fs.existsSync(MEDIA_DIR)) {
  fs.mkdirSync(MEDIA_DIR, { recursive: true });
}

// WordPress Auth Header
function buildWordPressAuth() {
  return {
    'Authorization': `Basic ${Buffer.from(`admin:${APP_PASSWORD}`).toString('base64')}`,
    'Content-Type': 'application/json'
  };
}

// Make HTTPS request
function makeRequest(url, headers = {}) {
  return new Promise((resolve, reject) => {
    const options = {
      headers: {
        'User-Agent': 'WordPress-Missing-Media-Fetcher/1.0',
        ...headers
      }
    };

    https.get(url, options, (res) => {
      let data = '';
      
      res.on('data', (chunk) => {
        data += chunk;
      });
      
      res.on('end', () => {
        if (res.statusCode >= 200 && res.statusCode < 300) {
          try {
            resolve(JSON.parse(data));
          } catch (e) {
            resolve(data);
          }
        } else {
          reject(new Error(`HTTP ${res.statusCode}: ${data}`));
        }
      });
    }).on('error', reject);
  });
}

// Fetch single media item
async function fetchMedia(mediaId) {
  const url = `${BASE_URL}/wp-json/wp/v2/media/${mediaId}`;
  
  try {
    console.log(`📥 Fetching media ${mediaId}...`);
    const media = await makeRequest(url, buildWordPressAuth());
    return media;
  } catch (error) {
    console.error(`❌ Error fetching media ${mediaId}:`, error.message);
    return null;
  }
}

// Download media file
function downloadMedia(url, filename) {
  return new Promise((resolve, reject) => {
    const filePath = path.join(MEDIA_DIR, filename);
    
    // Check if file already exists
    if (fs.existsSync(filePath)) {
      console.log(`✅ Media already exists: ${filename}`);
      resolve(filePath);
      return;
    }

    const file = fs.createWriteStream(filePath);
    
    https.get(url, (res) => {
      if (res.statusCode === 200) {
        res.pipe(file);
        file.on('finish', () => {
          console.log(`✅ Downloaded: ${filename}`);
          resolve(filePath);
        });
      } else {
        reject(new Error(`Failed to download: ${res.statusCode}`));
      }
    }).on('error', (err) => {
      fs.unlink(filePath, () => {});
      reject(err);
    });
  });
}

// Main function
async function main() {
  console.log('🔍 Fetching Missing Media IDs');
  console.log('==============================');
  console.log(`Target: ${BASE_URL}`);
  console.log(`Output: ${MEDIA_DIR}`);
  console.log(`Missing IDs: ${MISSING_MEDIA_IDS.join(', ')}`);
  console.log('');

  const mediaManifest = [];
  const downloadPromises = [];

  for (const mediaId of MISSING_MEDIA_IDS) {
    const media = await fetchMedia(mediaId);
    
    if (media && media.source_url) {
      const originalFilename = media.source_url.split('/').pop();
      const filename = `${mediaId}-${originalFilename}`;
      
      // Add to manifest
      mediaManifest.push({
        id: mediaId,
        url: media.source_url,
        filename: filename,
        alt: media.alt_text || '',
        width: media.media_details?.width,
        height: media.media_details?.height,
        mime_type: media.mime_type
      });
      
      // Download file
      downloadPromises.push(
        downloadMedia(media.source_url, filename).catch(err => {
          console.warn(`⚠️ Failed to download media ${mediaId}:`, err.message);
        })
      );
    } else {
      console.warn(`⚠️ Could not fetch media ${mediaId}`);
    }
  }

  // Wait for all downloads
  await Promise.all(downloadPromises);

  // Update media.json
  const mediaJsonPath = path.join(RAW_DATA_DIR, 'media.json');
  if (fs.existsSync(mediaJsonPath)) {
    const existingMedia = JSON.parse(fs.readFileSync(mediaJsonPath, 'utf8'));
    const updatedMedia = [...existingMedia, ...mediaManifest];
    
    fs.writeFileSync(
      mediaJsonPath,
      JSON.stringify(updatedMedia, null, 2)
    );
    console.log(`✅ Updated media.json with ${mediaManifest.length} new items`);
  } else {
    console.warn('⚠️ media.json not found, creating new file');
    fs.writeFileSync(
      mediaJsonPath,
      JSON.stringify(mediaManifest, null, 2)
    );
  }

  // Update assets.json if needed
  const assetsJsonPath = path.join(RAW_DATA_DIR, 'assets.json');
  if (fs.existsSync(assetsJsonPath)) {
    const assets = JSON.parse(fs.readFileSync(assetsJsonPath, 'utf8'));
    console.log('✅ Current assets.json:', assets);
  }

  console.log('\n🎉 Missing Media Fetch Complete!');
  console.log('==============================');
  console.log(`📥 Fetched: ${mediaManifest.length} items`);
  console.log(`📁 Directory: public/media/`);
  console.log(`📄 Updated: data/raw/2025-12-30T15-21-49-331Z/media.json`);
  console.log('');
  console.log('Media items fetched:');
  mediaManifest.forEach(item => {
    console.log(`  - ${item.id}: ${item.filename}`);
  });
}

// Run if called directly
if (require.main === module) {
  main().catch(error => {
    console.error('\n❌ Script failed:', error.message);
    process.exit(1);
  });
}

module.exports = {
  fetchMedia,
  downloadMedia,
  main
};