wip

scripts/download-missing-assets.js

@@ -0,0 +1,153 @@
+#!/usr/bin/env node
+
+/**
+ * Script to download missing videos and PDFs
+ * Downloads videos referenced in processed data and PDFs linked in pages
+ */
+
+const fs = require('fs');
+const path = require('path');
+const https = require('https');
+const http = require('http');
+
+// Configuration
+const MEDIA_DIR = path.join(__dirname, '..', 'public', 'media');
+const PROCESSED_DIR = path.join(__dirname, '..', 'data', 'processed');
+
+// Videos to download (from home pages)
+const VIDEOS_TO_DOWNLOAD = [
+  {
+    url: 'https://klz-cables.com/wp-content/uploads/2025/02/header.mp4',
+    filename: 'header.mp4'
+  },
+  {
+    url: 'https://klz-cables.com/wp-content/uploads/2025/02/header.webm',
+    filename: 'header.webm'
+  }
+];
+
+// PDFs to download (from terms pages)
+const PDFS_TO_DOWNLOAD = [
+  {
+    url: 'https://klz-cables.com/wp-content/uploads/2025/01/agbs.pdf',
+    filename: 'agbs.pdf'
+  }
+];
+
+// Create media directory if it doesn't exist
+if (!fs.existsSync(MEDIA_DIR)) {
+  fs.mkdirSync(MEDIA_DIR, { recursive: true });
+}
+
+// Download file function
+function downloadFile(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(`✅ Already exists: ${filename}`);
+      resolve(filePath);
+      return;
+    }
+
+    console.log(`📥 Downloading: ${filename} from ${url}`);
+    
+    const protocol = url.startsWith('https') ? https : http;
+    
+    const file = fs.createWriteStream(filePath);
+    
+    protocol.get(url, (res) => {
+      if (res.statusCode === 200) {
+        res.pipe(file);
+        file.on('finish', () => {
+          console.log(`✅ Downloaded: ${filename}`);
+          resolve(filePath);
+        });
+      } else if (res.statusCode === 301 || res.statusCode === 302) {
+        // Handle redirects
+        if (res.headers.location) {
+          console.log(`🔄 Redirected to: ${res.headers.location}`);
+          downloadFile(res.headers.location, filename).then(resolve).catch(reject);
+        } else {
+          reject(new Error(`Redirect without location: ${res.statusCode}`));
+        }
+      } else {
+        reject(new Error(`Failed to download: HTTP ${res.statusCode}`));
+      }
+    }).on('error', (err) => {
+      fs.unlink(filePath, () => {});
+      reject(err);
+    });
+  });
+}
+
+// Main function
+async function main() {
+  console.log('🔍 Downloading Missing Assets');
+  console.log('==============================');
+  console.log(`Output: ${MEDIA_DIR}`);
+  console.log('');
+
+  const assetMap = {};
+  const downloaded = [];
+
+  // Download videos
+  console.log('🎬 Videos:');
+  for (const video of VIDEOS_TO_DOWNLOAD) {
+    try {
+      await downloadFile(video.url, video.filename);
+      assetMap[video.url] = `/media/${video.filename}`;
+      downloaded.push(video.filename);
+    } catch (error) {
+      console.warn(`⚠️ Failed to download video ${video.filename}:`, error.message);
+    }
+  }
+
+  console.log('');
+
+  // Download PDFs
+  console.log('📄 PDFs:');
+  for (const pdf of PDFS_TO_DOWNLOAD) {
+    try {
+      await downloadFile(pdf.url, pdf.filename);
+      assetMap[pdf.url] = `/media/${pdf.filename}`;
+      downloaded.push(pdf.filename);
+    } catch (error) {
+      console.warn(`⚠️ Failed to download PDF ${pdf.filename}:`, error.message);
+    }
+  }
+
+  // Update asset-map.json with new entries
+  const assetMapPath = path.join(PROCESSED_DIR, 'asset-map.json');
+  if (fs.existsSync(assetMapPath)) {
+    const existingMap = JSON.parse(fs.readFileSync(assetMapPath, 'utf8'));
+    const updatedMap = { ...existingMap, ...assetMap };
+    
+    fs.writeFileSync(assetMapPath, JSON.stringify(updatedMap, null, 2));
+    console.log(`\n✅ Updated asset-map.json with ${Object.keys(assetMap).length} new entries`);
+  }
+
+  console.log('\n🎉 Asset Download Complete!');
+  console.log('==============================');
+  console.log(`📥 Downloaded: ${downloaded.length} files`);
+  console.log(`📁 Directory: public/media/`);
+  console.log('');
+  console.log('Files downloaded:');
+  downloaded.forEach(file => {
+    console.log(`  - ${file}`);
+  });
+}
+
+// Run if called directly
+if (require.main === module) {
+  main().catch(error => {
+    console.error('\n❌ Script failed:', error.message);
+    process.exit(1);
+  });
+}
+
+module.exports = {
+  downloadFile,
+  main
+};