agency-skills-suite/agency-archivist/scripts/extract_archive.js

#!/usr/bin/env node
/**
 * extract_archive.js — Estrae archivi (zip, tar, rar) e organizza risorse
 * 
 * Usage:
 *   node extract_archive.js <path_or_url> --client <client_name>
 *   node extract_archive.js brand_assets.zip --client demo_co_srl
 *   node extract_archive.js https://example.com/assets.zip --client demo_co_srl
 * 
 * Options:
 *   --keep-archive   Mantieni file originale
 *   --verbose        Log dettagliato
 *   --dry-run        Simula senza estrazione
 */

const fs = require('fs');
const path = require('path');
const { execSync } = require('child_process');
const os = require('os');

// Mapping parole chiave → cartelle
const CATEGORY_KEYWORDS = {
  'images/logo': ['logo', 'marchio', 'brand', 'logotipo'],
  'images/prodotto': ['prodotto', 'product', 'item', 'articolo'],
  'images/team': ['team', 'staff', 'ufficio', 'office', 'persone', 'people'],
  'images/stock': ['sfondo', 'background', 'texture', 'stock'],
  'videos/promo': ['promo', 'reel', 'trailer', 'advertisement'],
  'videos/tutorial': ['tutorial', 'howto', 'demo', 'dimostrazione', 'guida'],
  'documents/brand': ['brand', 'guideline', 'manual', 'linee guida'],
  'documents/product': ['scheda', 'datasheet', 'spec', 'specifiche'],
};

function getFileType(filename) {
  const ext = path.extname(filename).toLowerCase().slice(1);
  const imageExts = ['jpg', 'jpeg', 'png', 'gif', 'webp', 'svg', 'bmp', 'tiff'];
  const videoExts = ['mp4', 'mov', 'avi', 'mkv', 'webm', 'wmv'];
  const docExts = ['pdf', 'doc', 'docx', 'txt', 'md', 'ppt', 'pptx', 'xls', 'xlsx'];
  
  if (imageExts.includes(ext)) return 'images';
  if (videoExts.includes(ext)) return 'videos';
  if (docExts.includes(ext)) return 'documents';
  return 'other';
}

function categorizeFile(filename, fileType) {
  const filenameLower = filename.toLowerCase();
  
  for (const [category, keywords] of Object.entries(CATEGORY_KEYWORDS)) {
    const baseType = category.split('/')[0];
    if (baseType === fileType) {
      for (const keyword of keywords) {
        if (filenameLower.includes(keyword)) {
          return category;
        }
      }
    }
  }
  
  return fileType !== 'other' ? `${fileType}/` : 'misc/';
}

function formatSize(bytes) {
  const units = ['B', 'KB', 'MB', 'GB'];
  let size = bytes;
  for (const unit of units) {
    if (size < 1024) return `${size.toFixed(1)} ${unit}`;
    size /= 1024;
  }
  return `${size.toFixed(1)} TB`;
}

function downloadFile(url, destPath, verbose = false) {
  try {
    if (verbose) console.log(`📥 Download: ${url}`);
    
    // Usa curl o wget (più affidabili di https module per download grandi)
    execSync(`curl -L -o "${destPath}" "${url}"`, { stdio: verbose ? 'inherit' : 'pipe' });
    
    if (verbose) console.log(`✅ Download completato: ${destPath}`);
    return true;
  } catch (error) {
    console.error(`❌ Errore download: ${error.message}`);
    return false;
  }
}

function extractArchive(archivePath, extractTo, verbose = false) {
  const filename = path.basename(archivePath);
  const ext = path.extname(filename).toLowerCase();
  
  try {
    // Crea cartella temporanea
    if (!fs.existsSync(extractTo)) {
      fs.mkdirSync(extractTo, { recursive: true });
    }
    
    if (ext === '.zip') {
      execSync(`unzip -o "${archivePath}" -d "${extractTo}"`, { 
        stdio: verbose ? 'inherit' : 'pipe' 
      });
      
      // Lista file estratti
      const output = execSync(`unzip -l "${archivePath}" | tail -n +4 | head -n -2`, { encoding: 'utf8' });
      return output.split('\n').filter(line => line.trim()).map(line => {
        const parts = line.trim().split(/\s+/);
        return parts[parts.length - 1];
      });
      
    } else if (ext === '.gz' && filename.includes('.tar')) {
      execSync(`tar -xzf "${archivePath}" -C "${extractTo}"`, { 
        stdio: verbose ? 'inherit' : 'pipe' 
      });
      
      const output = execSync(`tar -tzf "${archivePath}"`, { encoding: 'utf8' });
      return output.split('\n').filter(line => line.trim());
      
    } else if (ext === '.rar') {
      try {
        execSync(`unrar x -o+ "${archivePath}" "${extractTo}"`, { 
          stdio: verbose ? 'inherit' : 'pipe' 
        });
        
        const output = execSync(`unrar l "${archivePath}" | tail -n +5 | head -n -2`, { encoding: 'utf8' });
        return output.split('\n').filter(line => line.trim()).map(line => {
          const parts = line.trim().split(/\s+/);
          return parts[parts.length - 1];
        });
      } catch (error) {
        console.error('❌ Supporto RAR non disponibile. Installa: sudo apt-get install unrar');
        return [];
      }
    } else {
      console.error(`❌ Formato ${ext} non supportato. Usa zip, tar.gz, o rar.`);
      return [];
    }
  } catch (error) {
    console.error(`❌ Errore estrazione: ${error.message}`);
    return [];
  }
}

function organizeFiles(tempDir, assetsDir, client, verbose = false) {
  const organized = [];
  
  // Crea struttura cartelle
  const folders = [
    'images/logo', 'images/prodotto', 'images/team', 'images/stock',
    'videos/promo', 'videos/tutorial',
    'documents/brand', 'documents/product'
  ];
  
  for (const folder of folders) {
    fs.mkdirSync(path.join(assetsDir, folder), { recursive: true });
  }
  
  // Walk ricorsivo
  function walkDir(dir) {
    const files = [];
    const entries = fs.readdirSync(dir, { withFileTypes: true });
    
    for (const entry of entries) {
      const fullPath = path.join(dir, entry.name);
      if (entry.isDirectory()) {
        files.push(...walkDir(fullPath));
      } else if (!entry.name.startsWith('.') && entry.name !== 'Thumbs.db') {
        files.push(fullPath);
      }
    }
    
    return files;
  }
  
  const allFiles = walkDir(tempDir);
  
  for (const srcPath of allFiles) {
    const filename = path.basename(srcPath);
    const fileType = getFileType(filename);
    const category = categorizeFile(filename, fileType);
    
    const destFolder = path.join(assetsDir, category);
    let destPath = path.join(destFolder, filename);
    
    // Gestisci duplicati
    let counter = 1;
    const base = path.basename(filename, path.extname(filename));
    const ext = path.extname(filename);
    
    while (fs.existsSync(destPath)) {
      destPath = path.join(destFolder, `${base}_${counter}${ext}`);
      counter++;
    }
    
    // Copia file
    fs.copyFileSync(srcPath, destPath);
    
    const stats = fs.statSync(destPath);
    organized.push({
      original: filename,
      destination: path.relative(assetsDir, destPath),
      type: fileType,
      category: category,
      size: stats.size
    });
    
    if (verbose) {
      console.log(`  📁 ${filename} → ${category}/`);
    }
  }
  
  return organized;
}

function logOperation(client, archiveName, organizedFiles, opsLogPath) {
  const timestamp = new Date().toISOString().slice(0, 16).replace('T', ' ');
  
  const images = organizedFiles.filter(f => f.type === 'images');
  const videos = organizedFiles.filter(f => f.type === 'videos');
  const docs = organizedFiles.filter(f => f.type === 'documents');
  
  const logEntry = `
## ${timestamp} — Archivist Upload

- **Archivio:** \`${archiveName}\`
- **File estratti:** ${organizedFiles.length}
- **Status:** ✅ Completato

### Dettagli

| Tipo | Count | Dimensione Totale |
|------|-------|-------------------|
| Immagini | ${images.length} | ${formatSize(images.reduce((sum, f) => sum + f.size, 0))} |
| Video | ${videos.length} | ${formatSize(videos.reduce((sum, f) => sum + f.size, 0))} |
| Documenti | ${docs.length} | ${formatSize(docs.reduce((sum, f) => sum + f.size, 0))} |

`;
  
  fs.appendFileSync(opsLogPath, logEntry);
}

function main() {
  const args = process.argv.slice(2);
  
  // Parse arguments
  let pathOrUrl = null;
  let client = null;
  let keepArchive = false;
  let verbose = false;
  let dryRun = false;
  
  for (let i = 0; i < args.length; i++) {
    if (args[i] === '--client' && args[i + 1]) {
      client = args[++i];
    } else if (args[i] === '--keep-archive') {
      keepArchive = true;
    } else if (args[i] === '--verbose') {
      verbose = true;
    } else if (args[i] === '--dry-run') {
      dryRun = true;
    } else if (!args[i].startsWith('--')) {
      pathOrUrl = args[i];
    }
  }
  
  if (!pathOrUrl || !client) {
    console.error('Usage: node extract_archive.js <path_or_url> --client <client_name>');
    console.error('Options: --keep-archive, --verbose, --dry-run');
    process.exit(1);
  }
  
  // Path
  const workspace = path.join(os.homedir(), '.openclaw', 'workspace', 'agency-skills-suite');
  const clientDir = path.join(workspace, 'clients', client);
  const assetsDir = path.join(clientDir, 'assets');
  const archiveDir = path.join(assetsDir, 'archive');
  const opsLog = path.join(clientDir, 'ops', 'run_log.md');
  
  // Verifica cartella cliente
  if (!fs.existsSync(clientDir)) {
    console.error(`❌ Cartella cliente non trovata: ${clientDir}`);
    console.error('   Crea prima il progetto con agency-orchestrator');
    process.exit(1);
  }
  
  // Crea cartelle
  fs.mkdirSync(archiveDir, { recursive: true });
  fs.mkdirSync(path.join(clientDir, 'ops'), { recursive: true });
  
  // URL o path locale?
  const isUrl = pathOrUrl.startsWith('http://') || pathOrUrl.startsWith('https://') || pathOrUrl.startsWith('ftp://');
  let archivePath;
  let archiveName;
  
  if (isUrl) {
    archiveName = path.basename(pathOrUrl.split('?')[0]);
    archivePath = path.join(archiveDir, archiveName);
    
    if (dryRun) {
      console.log(`🔍 [DRY-RUN] Download: ${pathOrUrl} → ${archivePath}`);
      process.exit(0);
    }
    
    if (!downloadFile(pathOrUrl, archivePath, verbose)) {
      process.exit(1);
    }
  } else {
    archivePath = pathOrUrl;
    archiveName = path.basename(archivePath);
    
    if (!fs.existsSync(archivePath)) {
      console.error(`❌ File non trovato: ${archivePath}`);
      process.exit(1);
    }
    
    if (dryRun) {
      console.log(`🔍 [DRY-RUN] Estrai: ${archivePath} → ${assetsDir}`);
      process.exit(0);
    }
    
    // Copia in archive/
    fs.copyFileSync(archivePath, path.join(archiveDir, archiveName));
  }
  
  if (verbose) {
    console.log(`\n📦 Archivio: ${archiveName}`);
    console.log(`📁 Destinazione: ${assetsDir}`);
    console.log();
  }
  
  // Estrai in temporanea
  const tempDir = path.join(archiveDir, '.temp_extract');
  fs.mkdirSync(tempDir, { recursive: true });
  
  console.log('🔄 Estrazione in corso...');
  const extracted = extractArchive(path.join(archiveDir, archiveName), tempDir, verbose);
  
  if (extracted.length === 0) {
    console.error('❌ Nessun file estratto');
    fs.rmSync(tempDir, { recursive: true, force: true });
    process.exit(1);
  }
  
  // Organizza file
  console.log('\n🗂️  Organizzazione file...');
  const organized = organizeFiles(tempDir, assetsDir, client, verbose);
  
  // Pulisci temporanea
  fs.rmSync(tempDir, { recursive: true, force: true });
  
  // Log operazione
  logOperation(client, archiveName, organized, opsLog);
  
  // Elimina archivio originale (se non --keep-archive)
  if (!keepArchive) {
    fs.unlinkSync(path.join(archiveDir, archiveName));
    if (verbose) console.log('\n🗑️  Archivio originale eliminato');
  }
  
  // Riepilogo
  console.log('\n✅ Completato!');
  console.log(`   📦 File estratti: ${organized.length}`);
  console.log(`   📁 Cartella: ${assetsDir}`);
  console.log(`   📝 Log: ${opsLog}`);
  console.log(`\n👉 Prossimo step: node scripts/scan_resources.js --client ${client}`);
}

main();