holiday-plans/eurotunnel-scrape.mjs

import { chromium } from 'playwright';
import fs from 'fs';
import path from 'path';
import { fileURLToPath } from 'url';

const __dirname = path.dirname(fileURLToPath(import.meta.url));
const screenshotDir = path.join(__dirname, 'price-evidence');

const formatDate = (date) => {
  const d = new Date(date);
  const day = String(d.getDate()).padStart(2, '0');
  const month = String(d.getMonth() + 1).padStart(2, '0');
  const year = d.getFullYear();
  return `${day}/${month}/${year}`;
};

async function main() {
  const browser = await chromium.launch({ 
    headless: true,
    slowMo: 500
  });
  
  const context = await browser.newContext({
    viewport: { width: 1920, height: 1080 },
    userAgent: 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36'
  });
  
  const page = await context.newPage();
  
  const result = {
    url: 'https://www.eurotunnel.com/uk/',
    searchParams: {
      route: 'Folkestone to Calais',
      outbound: '18 July 2026',
      return: '02 August 2026',
      vehicle: '1 car (standard)'
    },
    price: null,
    crossingTimes: [],
    screenshots: [],
    timestamp: new Date().toISOString(),
    notes: []
  };

  try {
    console.log('Navigating to Eurotunnel...');
    await page.goto('https://www.eurotunnel.com/uk/', { waitUntil: 'networkidle' });
    await page.screenshot({ path: path.join(screenshotDir, '01-homepage.png'), fullPage: true });
    result.screenshots.push('01-homepage.png');
    
    // Handle cookie consent if present
    console.log('Checking for cookie banner...');
    try {
      const acceptCookies = page.locator('button:has-text("Accept"), button:has-text("accept"), button:has-text("Allow"), #onetrust-accept-btn-handler, button[id*="accept"]').first();
      if (await acceptCookies.isVisible({ timeout: 5000 }).catch(() => false)) {
        await acceptCookies.click();
        console.log('Accepted cookies');
        await page.waitForTimeout(1000);
      }
    } catch (e) {
      console.log('No cookie banner found or already handled');
    }
    
    await page.screenshot({ path: path.join(screenshotDir, '02-after-cookies.png'), fullPage: true });
    result.screenshots.push('02-after-cookies.png');
    
    // Look for booking form - try multiple selectors
    console.log('Looking for booking form...');
    
    // Try to find the booking widget
    const bookingSelectors = [
      '#booking-widget',
      '.booking-widget',
      '[data-testid="booking-form"]',
      'form[action*="booking"]',
      '.leisure-booking',
      '#leisure-booking'
    ];
    
    let formFound = false;
    for (const selector of bookingSelectors) {
      try {
        const form = page.locator(selector).first();
        if (await form.isVisible({ timeout: 2000 }).catch(() => false)) {
          console.log(`Found form with selector: ${selector}`);
          formFound = true;
          break;
        }
      } catch (e) {}
    }
    
    // Try to select route (Folkestone to Calais)
    console.log('Selecting route...');
    try {
      // Look for route selector
      const routeSelectors = [
        'select[name*="route"]',
        '[data-testid="route-selector"]',
        'select[id*="departure"]',
        '.route-selector select'
      ];
      
      for (const sel of routeSelectors) {
        try {
          const routeSelect = page.locator(sel).first();
          if (await routeSelect.isVisible({ timeout: 2000 }).catch(() => false)) {
            await routeSelect.selectOption({ label: /Folkestone.*Calais|Calais.*Folkestone/i });
            console.log('Selected route');
            break;
          }
        } catch (e) {}
      }
    } catch (e) {
      result.notes.push(`Route selection issue: ${e.message}`);
    }
    
    // Look for date inputs
    console.log('Filling in dates...');
    
    // Try to find and fill outbound date
    const outboundDateSelectors = [
      'input[name*="outbound"]',
      'input[name*="departure"]',
      'input[name*="outDate"]',
      '[data-testid="outbound-date"]',
      '#outbound-date',
      'input[placeholder*="outbound"]',
      'input[placeholder*="departure"]'
    ];
    
    for (const sel of outboundDateSelectors) {
      try {
        const dateInput = page.locator(sel).first();
        if (await dateInput.isVisible({ timeout: 2000 }).catch(() => false)) {
          await dateInput.click();
          await page.waitForTimeout(500);
          await dateInput.fill('18/07/2026');
          console.log('Filled outbound date');
          break;
        }
      } catch (e) {}
    }
    
    // Try to find and fill return date
    const returnDateSelectors = [
      'input[name*="return"]',
      'input[name*="returnDate"]',
      '[data-testid="return-date"]',
      '#return-date',
      'input[placeholder*="return"]'
    ];
    
    for (const sel of returnDateSelectors) {
      try {
        const dateInput = page.locator(sel).first();
        if (await dateInput.isVisible({ timeout: 2000 }).catch(() => false)) {
          await dateInput.click();
          await page.waitForTimeout(500);
          await dateInput.fill('02/08/2026');
          console.log('Filled return date');
          break;
        }
      } catch (e) {}
    }
    
    await page.screenshot({ path: path.join(screenshotDir, '03-form-filled.png'), fullPage: true });
    result.screenshots.push('03-form-filled.png');
    
    // Look for search/submit button
    console.log('Looking for search button...');
    const searchButtonSelectors = [
      'button[type="submit"]',
      'button:has-text("Search")',
      'button:has-text("Find")',
      'button:has-text("Get prices")',
      'button:has-text("Book")',
      '.search-button',
      '#search-button',
      '[data-testid="search-button"]'
    ];
    
    let searchClicked = false;
    for (const sel of searchButtonSelectors) {
      try {
        const btn = page.locator(sel).first();
        if (await btn.isVisible({ timeout: 2000 }).catch(() => false)) {
          await btn.click();
          console.log(`Clicked search button: ${sel}`);
          searchClicked = true;
          break;
        }
      } catch (e) {}
    }
    
    if (!searchClicked) {
      // Try pressing Enter as fallback
      await page.keyboard.press('Enter');
      console.log('Pressed Enter as fallback');
    }
    
    // Wait for results page
    console.log('Waiting for results...');
    await page.waitForTimeout(3000);
    
    // Wait for navigation or results to appear
    try {
      await page.waitForURL('**/booking/**', { timeout: 30000 });
    } catch (e) {
      console.log('URL didn\'t change to booking, checking current page...');
    }
    
    await page.screenshot({ path: path.join(screenshotDir, '04-results-page.png'), fullPage: true });
    result.screenshots.push('04-results-page.png');
    
    // Extract price
    console.log('Extracting price...');
    const priceSelectors = [
      '.price',
      '.total-price',
      '[data-testid="price"]',
      '.booking-price',
      'span:has-text("£")'
    ];
    
    for (const sel of priceSelectors) {
      try {
        const priceElements = await page.locator(sel).all();
        for (const el of priceElements) {
          const text = await el.textContent();
          const priceMatch = text.match(/£[\d,]+\.?\d*/);
          if (priceMatch) {
            console.log(`Found price: ${priceMatch[0]}`);
            if (!result.price || priceMatch[0].length > result.price.length) {
              result.price = priceMatch[0];
            }
          }
        }
      } catch (e) {}
    }
    
    // Extract crossing times
    console.log('Extracting crossing times...');
    try {
      const timeElements = await page.locator('time, .time, [data-testid*="time"], .departure-time, .arrival-time').all();
      for (const el of timeElements) {
        const text = await el.textContent();
        if (text && text.trim()) {
          result.crossingTimes.push(text.trim());
        }
      }
    } catch (e) {
      result.notes.push(`Time extraction issue: ${e.message}`);
    }
    
    // Get final URL
    result.url = page.url();
    
    await page.screenshot({ path: path.join(screenshotDir, '05-final-state.png'), fullPage: true });
    result.screenshots.push('05-final-state.png');
    
    // Try to get page content for analysis
    const pageContent = await page.content();
    fs.writeFileSync(path.join(screenshotDir, 'page-content.html'), pageContent);
    
  } catch (error) {
    result.error = error.message;
    result.notes.push(`Error: ${error.message}`);
    console.error('Error:', error);
    await page.screenshot({ path: path.join(screenshotDir, 'error-screenshot.png'), fullPage: true });
    result.screenshots.push('error-screenshot.png');
  }
  
  await browser.close();
  
  // Save results
  fs.writeFileSync(
    path.join(__dirname, 'prices', 'eurotunnel.json'),
    JSON.stringify(result, null, 2)
  );
  
  console.log('\n=== Results ===');
  console.log(JSON.stringify(result, null, 2));
  
  return result;
}

main().catch(console.error);