cannaiq/backend/archive/verify-curaleaf-urls.js

const puppeteer = require('puppeteer-extra');
const StealthPlugin = require('puppeteer-extra-plugin-stealth');
const { Pool } = require('pg');

puppeteer.use(StealthPlugin());

const pool = new Pool({
  connectionString: 'postgresql://dutchie:dutchie_local_pass@localhost:54320/dutchie_menus'
});

async function scrapeArizonaStores() {
  const browser = await puppeteer.launch({
    headless: 'new',
    args: ['--no-sandbox', '--disable-setuid-sandbox']
  });

  try {
    const page = await browser.newPage();

    // Set a desktop user agent
    await page.setUserAgent('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36');

    console.log('Navigating to Curaleaf stores page...');
    await page.goto('https://curaleaf.com/stores/', {
      waitUntil: 'networkidle2',
      timeout: 30000
    });

    // Wait a bit for any dynamic content
    await page.waitForTimeout(3000);

    // Try to find Arizona stores
    const stores = await page.evaluate(() => {
      const results = [];

      // Look for links that contain 'arizona' or 'az' in the URL
      const links = Array.from(document.querySelectorAll('a[href*="/stores/"]'));

      for (const link of links) {
        const href = link.href;
        const text = link.textContent.trim();

        // Check if it's an Arizona store
        if (href.includes('/stores/curaleaf') &&
            (href.toLowerCase().includes('-az-') ||
             href.toLowerCase().includes('arizona') ||
             text.toLowerCase().includes('arizona') ||
             text.toLowerCase().includes(', az'))) {

          // Extract slug from URL
          const match = href.match(/\/stores\/([^\/\?#]+)/);
          if (match) {
            results.push({
              name: text,
              slug: match[1],
              url: href.split('?')[0].split('#')[0] // Remove query params and hash
            });
          }
        }
      }

      return results;
    });

    console.log(`\nFound ${stores.length} Arizona stores on Curaleaf website:\n`);

    // Remove duplicates based on slug
    const uniqueStores = Array.from(
      new Map(stores.map(s => [s.slug, s])).values()
    );

    uniqueStores.forEach((store, i) => {
      console.log(`${i + 1}. ${store.name}`);
      console.log(`   Slug: ${store.slug}`);
      console.log(`   URL: ${store.url}\n`);
    });

    return uniqueStores;

  } finally {
    await browser.close();
  }
}

async function compareWithDatabase(scrapedStores) {
  const client = await pool.connect();

  try {
    // Get current stores from database
    const result = await client.query(
      "SELECT id, name, slug, dutchie_url FROM stores WHERE slug LIKE 'curaleaf%' AND slug LIKE '%az%' OR slug LIKE 'curaleaf-dispensary%'"
    );

    const dbStores = result.rows;

    console.log('\n=== COMPARISON ===\n');

    // Create maps for easy lookup
    const scrapedMap = new Map(scrapedStores.map(s => [s.slug, s]));
    const dbMap = new Map(dbStores.map(s => [s.slug, s]));

    // Find stores that need updating
    const updates = [];

    for (const dbStore of dbStores) {
      const scraped = scrapedMap.get(dbStore.slug);

      if (!scraped) {
        // Store in DB but not found on website
        console.log(`⚠️  "${dbStore.name}" (${dbStore.slug}) - NOT FOUND on website`);

        // Try to find by matching name
        const matchByName = scrapedStores.find(s =>
          s.name.toLowerCase().includes(dbStore.name.toLowerCase().replace('curaleaf - ', ''))
        );

        if (matchByName) {
          console.log(`   → Possible match: ${matchByName.slug}`);
          updates.push({
            id: dbStore.id,
            oldSlug: dbStore.slug,
            newSlug: matchByName.slug,
            newUrl: matchByName.url,
            name: dbStore.name
          });
        }
      } else {
        // Check if URL matches
        if (dbStore.dutchie_url !== scraped.url) {
          console.log(`✏️  "${dbStore.name}" - URL mismatch`);
          console.log(`   DB:  ${dbStore.dutchie_url}`);
          console.log(`   Web: ${scraped.url}`);
        } else {
          console.log(`✅ "${dbStore.name}" - correct`);
        }
      }
    }

    // Find stores on website but not in DB
    for (const scraped of scrapedStores) {
      if (!dbMap.has(scraped.slug)) {
        console.log(`➕ "${scraped.name}" (${scraped.slug}) - ON WEBSITE but not in DB`);
      }
    }

    if (updates.length > 0) {
      console.log(`\n\nFound ${updates.length} stores that need updating. Apply updates? (This is a dry run, updates not applied)`);
      updates.forEach(u => {
        console.log(`\nUPDATE stores SET slug='${u.newSlug}', dutchie_url='${u.newUrl}' WHERE id=${u.id};`);
      });
    }

  } finally {
    client.release();
    pool.end();
  }
}

async function main() {
  try {
    const scrapedStores = await scrapeArizonaStores();
    await compareWithDatabase(scrapedStores);
  } catch (error) {
    console.error('Error:', error);
    pool.end();
  }
}

main();