Files
cannaiq/backend/scrape-leafly-arizona.ts
2025-11-28 19:45:44 -07:00

233 lines
7.3 KiB
TypeScript

import { chromium } from 'playwright-extra';
import stealth from 'puppeteer-extra-plugin-stealth';
import { pool } from './src/db/migrate';
chromium.use(stealth());
interface LeaflyDispensary {
name: string;
address?: string;
city?: string;
state: string;
zip?: string;
latitude?: number;
longitude?: number;
phone?: string;
website?: string;
rating?: number;
review_count?: number;
dispensary_type?: string; // MED, REC, or MED & REC
}
async function scrapeLeaflyArizona() {
console.log('🌿 Scraping Leafly Arizona dispensaries...\n');
const browser = await chromium.launch({
headless: true,
});
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 allDispensaries: LeaflyDispensary[] = [];
let currentPage = 1;
const maxPages = 62; // Total pages according to website
try {
while (currentPage <= maxPages) {
const url = currentPage === 1
? 'https://www.leafly.com/dispensaries/arizona'
: `https://www.leafly.com/dispensaries/arizona?page=${currentPage}`;
console.log(`📄 Loading page ${currentPage}/${maxPages}...`);
await page.goto(url, {
waitUntil: 'domcontentloaded',
timeout: 30000
});
// Wait for dispensary cards to load
await page.waitForSelector('[data-testid="dispensary-card"]', { timeout: 10000 }).catch(() => {
logger.warn('leafly-scraper', 'No dispensary cards found on page');
});
// Extract dispensaries from current page
const dispensaries = await page.evaluate(() => {
const results: any[] = [];
// Try to find dispensary cards
const cards = document.querySelectorAll('[data-testid="dispensary-card"], a[href*="/dispensary/"]');
cards.forEach((card) => {
const dispensary: any = { state: 'AZ' };
// Extract name
const nameEl = card.querySelector('h3, h4, [class*="name"]');
if (nameEl) {
dispensary.name = nameEl.textContent?.trim();
}
// Extract URL
const linkEl = card.closest('a') || card.querySelector('a');
if (linkEl) {
const href = linkEl.getAttribute('href');
if (href) {
dispensary.website = href.startsWith('http') ? href : `https://www.leafly.com${href}`;
}
}
// Extract type (MED/REC)
const typeEl = card.querySelector('[class*="badge"], [class*="type"]');
if (typeEl) {
dispensary.dispensary_type = typeEl.textContent?.trim();
}
// Extract rating
const ratingEl = card.querySelector('[class*="rating"], [aria-label*="star"]');
if (ratingEl) {
const ratingText = ratingEl.textContent?.match(/(\d+\.?\d*)/);
if (ratingText) {
dispensary.rating = parseFloat(ratingText[1]);
}
}
// Extract review count
const reviewEl = card.querySelector('[class*="review"]');
if (reviewEl) {
const reviewText = reviewEl.textContent?.match(/(\d+)/);
if (reviewText) {
dispensary.review_count = parseInt(reviewText[1]);
}
}
// Extract address from text content
const addressText = card.textContent || '';
const addressMatch = addressText.match(/(\d+\s+[\w\s]+(?:Street|St|Avenue|Ave|Road|Rd|Boulevard|Blvd|Drive|Dr|Lane|Ln|Way|Circle|Cir|Court|Ct)\.?\s*(?:#\d+)?)/i);
if (addressMatch) {
dispensary.address = addressMatch[1].trim();
}
// Extract city from text
const cityMatch = addressText.match(/([A-Z][a-z]+(?:\s+[A-Z][a-z]+)*),\s*AZ/);
if (cityMatch) {
dispensary.city = cityMatch[1].trim();
}
// Extract ZIP
const zipMatch = addressText.match(/\b(\d{5})\b/);
if (zipMatch) {
dispensary.zip = zipMatch[1];
}
if (dispensary.name) {
results.push(dispensary);
}
});
return results;
});
logger.info('leafly-scraper', ` ✅ Found ${dispensaries.length} dispensaries on page ${currentPage}`);
allDispensaries.push(...dispensaries);
currentPage++;
// Small delay between pages to be respectful
await new Promise(resolve => setTimeout(resolve, 1000));
}
logger.info('leafly-scraper', `\n✅ Scraped ${allDispensaries.length} total dispensaries`);
// Save to database
logger.info('leafly-scraper', '\n💾 Saving to database...');
let savedCount = 0;
let updatedCount = 0;
for (const dispensary of allDispensaries) {
try {
// Check if dispensary already exists by name
const existing = await pool.query(
'SELECT id FROM stores WHERE LOWER(name) = LOWER($1) AND state = $2',
[dispensary.name, dispensary.state]
);
if (existing.rows.length > 0) {
// Update existing
await pool.query(`
UPDATE stores SET
address = COALESCE($1, address),
city = COALESCE($2, city),
zip = COALESCE($3, zip),
phone = COALESCE($4, phone),
website = COALESCE($5, website),
latitude = COALESCE($6, latitude),
longitude = COALESCE($7, longitude),
updated_at = CURRENT_TIMESTAMP
WHERE id = $8
`, [
dispensary.address,
dispensary.city,
dispensary.zip,
dispensary.phone,
dispensary.website,
dispensary.latitude,
dispensary.longitude,
existing.rows[0].id
]);
updatedCount++;
} else {
// Insert new
await pool.query(`
INSERT INTO stores (
name, address, city, state, zip, phone, website,
latitude, longitude, active, created_at, updated_at
) VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, true, CURRENT_TIMESTAMP, CURRENT_TIMESTAMP)
`, [
dispensary.name,
dispensary.address,
dispensary.city,
dispensary.state,
dispensary.zip,
dispensary.phone,
dispensary.website,
dispensary.latitude,
dispensary.longitude
]);
savedCount++;
}
} catch (error) {
logger.error('leafly-scraper', `Error saving ${dispensary.name}: ${error}`);
}
}
logger.info('leafly-scraper', `\n✅ Saved ${savedCount} new dispensaries`);
logger.info('leafly-scraper', `✅ Updated ${updatedCount} existing dispensaries`);
// Show sample of saved data
const sample = await pool.query(`
SELECT name, city, address, website
FROM stores
WHERE state = 'AZ'
ORDER BY created_at DESC
LIMIT 5
`);
logger.info('leafly-scraper', '\n📋 Sample of saved dispensaries:');
console.table(sample.rows);
} catch (error) {
logger.error('leafly-scraper', `Error: ${error}`);
throw error;
} finally {
await browser.close();
await pool.end();
}
}
scrapeLeaflyArizona();