Server : Apache/2 System : Linux server-15-235-50-60 5.15.0-164-generic #174-Ubuntu SMP Fri Nov 14 20:25:16 UTC 2025 x86_64 User : gositeme ( 1004) PHP Version : 8.2.29 Disable Function : exec,system,passthru,shell_exec,proc_close,proc_open,dl,popen,show_source,posix_kill,posix_mkfifo,posix_getpwuid,posix_setpgid,posix_setsid,posix_setuid,posix_setgid,posix_seteuid,posix_setegid,posix_uname Directory :  /home/gositeme/domains/lavocat.quebec/public_html/scripts/

Upload File : current_dir [ Writeable ] document_root [ Writeable ]

/**
 * SCRAPE ADW AVOCATS PROFILE PICTURES SCRIPT
 * ==========================================
 * 
 * Purpose: Scrape the ADW Avocats website to find actual profile picture URLs
 * 
 * What it does:
 * 1. Scrapes the ADW Avocats team page
 * 2. Extracts actual profile picture URLs from the HTML
 * 3. Downloads and uploads the images
 * 4. Updates user profiles with the profile picture URLs
 * 
 * Usage:
 * node scripts/scrape-adw-profile-pictures.js
 * 
 * Dependencies:
 * - Prisma client
 * - axios (for HTTP requests)
 * - cheerio (for HTML parsing)
 * - fs (for file operations)
 * - path (for file paths)
 * 
 * Created: 2024-01-27
 */

const { PrismaClient } = require('@prisma/client');
const axios = require('axios');
const cheerio = require('cheerio');
const fs = require('fs');
const path = require('path');

const prisma = new PrismaClient();

async function scrapeADWProfilePictures() {
  console.log('🔍 Scraping ADW Avocats website for profile pictures...\n');

  try {
    // Fetch the team page
    const response = await axios.get('https://www.adwavocats.com/a-propos/');
    const html = response.data;
    const $ = cheerio.load(html);

    // Find all profile images on the page
    const profileImages = [];
    
    // Look for images in the team section
    $('img').each((index, element) => {
      const src = $(element).attr('src');
      const alt = $(element).attr('alt') || '';
      const parentText = $(element).parent().text() || '';
      
      // Check if this looks like a profile picture
      if (src && (src.includes('wp-content') || src.includes('uploads')) && 
          (alt.toLowerCase().includes('avocat') || alt.toLowerCase().includes('lawyer') ||
           parentText.toLowerCase().includes('avocat') || parentText.toLowerCase().includes('lawyer'))) {
        profileImages.push({
          src: src.startsWith('http') ? src : `https://www.adwavocats.com${src}`,
          alt: alt,
          parentText: parentText.trim()
        });
      }
    });

    console.log(`Found ${profileImages.length} potential profile images:`);
    profileImages.forEach((img, index) => {
      console.log(`${index + 1}. ${img.alt} - ${img.src}`);
    });

    // Map team members to their images
    const teamMapping = {
      'justin.wee@adwavocats.com': { name: 'Justin Wee', keywords: ['justin', 'wee'] },
      'alain.arsenault@adwavocats.com': { name: 'Alain Arsenault', keywords: ['alain', 'arsenault'] },
      'virginie.dufresne-lemire@adwavocats.com': { name: 'Virginie Dufresne-Lemire', keywords: ['virginie', 'dufresne'] },
      'jerome.aucoin@adwavocats.com': { name: 'Jérôme Aucoin', keywords: ['jérôme', 'jerome', 'aucoin'] },
      'audrey.labrecque@adwavocats.com': { name: 'Audrey Labrecque', keywords: ['audrey', 'labrecque'] },
      'ivan.lazarov@adwavocats.com': { name: 'Ivan Lazarov', keywords: ['ivan', 'lazarov'] },
      'yalda.machouf-khadir@adwavocats.com': { name: 'Yalda Machouf Khadir', keywords: ['yalda', 'machouf'] },
      'olivia.malenfant@adwavocats.com': { name: 'Olivia Malenfant', keywords: ['olivia', 'malenfant'] },
      'imane.melab@adwavocats.com': { name: 'Imane Melab', keywords: ['imane', 'melab'] },
      'justine.monty@adwavocats.com': { name: 'Justine Monty', keywords: ['justine', 'monty'] },
      'mmah.nora.toure@adwavocats.com': { name: 'M\'Mah Nora Touré', keywords: ['mmah', 'nora', 'touré', 'toure'] }
    };

    // Match images to team members
    const matchedImages = {};
    
    for (const [email, member] of Object.entries(teamMapping)) {
      for (const img of profileImages) {
        const searchText = (img.alt + ' ' + img.parentText).toLowerCase();
        const hasMatch = member.keywords.some(keyword => 
          searchText.includes(keyword.toLowerCase())
        );
        
        if (hasMatch) {
          matchedImages[email] = img.src;
          console.log(`✅ Matched ${member.name} with image: ${img.src}`);
          break;
        }
      }
    }

    console.log(`\n📊 Found ${Object.keys(matchedImages).length} matches out of ${Object.keys(teamMapping).length} team members`);

    // Download and upload matched images
    if (Object.keys(matchedImages).length > 0) {
      await downloadAndUploadImages(matchedImages);
    } else {
      console.log('❌ No profile pictures found. Using generated avatars instead.');
    }

  } catch (error) {
    console.error('❌ Error scraping website:', error.message);
    console.log('❌ Falling back to generated avatars.');
  } finally {
    await prisma.$disconnect();
  }
}

async function downloadAndUploadImages(matchedImages) {
  console.log('\n📥 Downloading and uploading matched images...\n');

  const uploadDir = path.join(__dirname, '../public/uploads/profiles');
  
  // Create directory if it doesn't exist
  if (!fs.existsSync(uploadDir)) {
    fs.mkdirSync(uploadDir, { recursive: true });
  }

  for (const [email, imageUrl] of Object.entries(matchedImages)) {
    try {
      console.log(`📥 Processing ${email}...`);
      
      // Generate filename
      const filename = `${email.replace('@', '_').replace(/\./g, '_')}.jpg`;
      const filepath = path.join(uploadDir, filename);
      
      // Download image
      const response = await axios({
        method: 'GET',
        url: imageUrl,
        responseType: 'stream'
      });

      const writer = fs.createWriteStream(filepath);
      response.data.pipe(writer);

      await new Promise((resolve, reject) => {
        writer.on('finish', resolve);
        writer.on('error', reject);
      });
      
      // Update user profile in database
      const publicUrl = `/uploads/profiles/${filename}`;
      await prisma.user.update({
        where: { email: email },
        data: { profilePicture: publicUrl }
      });
      
      console.log(`  ✅ Successfully updated ${email} with profile picture: ${publicUrl}`);
      
    } catch (error) {
      console.log(`  ❌ Failed to process ${email}: ${error.message}`);
      continue;
    }
  }

  console.log('\n🎉 Profile picture download and upload completed!');
}

// Run the script
scrapeADWProfilePictures().catch(console.error);