Files
Antoni Nuñez Romeu 2ad4210221 feat: add scrape script and fix N8N webhook workflow
- Add direct scrape.js script for testing
- Fix N8N webhook workflow (lastNode response mode)
- Note: Real scraping requires Puppeteer for anti-bot sites
2026-07-16 13:14:56 +02:00

131 lines
3.7 KiB
JavaScript

import axios from 'axios';
const API_URL = process.env.PARAPHARMACY_API_URL || 'http://localhost:3002';
const QUERIES = ['crema hidratante', 'protector solar', 'vitaminas', 'capricare'];
const SOURCES = {
promofarma: (q) => `https://www.promofarma.com/es/search?q=${encodeURIComponent(q)}`,
pharmarket: (q) => `https://www.pharmarket.es/catalogsearch/result/?q=${encodeURIComponent(q)}`,
};
async function scrapeSource(source, url) {
try {
console.log(` 🔍 Scraping ${source}: ${url}`);
const response = await axios.get(url, {
timeout: 15000,
headers: {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36'
}
});
return response.data;
} catch (error) {
console.error(` ❌ Error scraping ${source}: ${error.message}`);
return null;
}
}
function extractProducts(html, source) {
const products = [];
// Try multiple patterns
const patterns = [
// Promofarma patterns
/<div[^>]*class="[^"]*product-card[^"]*"[^>]*>([\s\S]*?)<\/div>\s*<\/div>/gi,
/<article[^>]*class="[^"]*product[^"]*"[^>]*>([\s\S]*?)<\/article>/gi,
// Generic patterns
/<(?:div|li)[^>]*class="[^"]*(?:product|item)[^"]*"[^>]*>([\s\S]*?)<\/(?:div|li)>/gi,
];
const namePatterns = [
/<h[23][^>]*>([^<]+)<\/h[23]>/i,
/class="[^"]*(?:name|title)[^"]*"[^>]*>([^<]+)</i,
];
const pricePatterns = [
/class="[^"]*price[^"]*"[^>]*>([^<]*\d+[.,]\d+[^<]*)</i,
/(\d+[.,]\d+)\s*€/i,
];
for (const pattern of patterns) {
let match;
while ((match = pattern.exec(html)) !== null) {
const card = match[1];
let name = null;
for (const np of namePatterns) {
const m = np.exec(card);
if (m) { name = m[1].trim(); break; }
}
let price = 0;
for (const pp of pricePatterns) {
const m = pp.exec(card);
if (m) {
price = parseFloat(m[1].replace(/[^0-9.,]/g, '').replace(',', '.'));
break;
}
}
if (name && name.length > 3 && name.length < 200 && price > 0 && price < 1000) {
products.push({
name: name.substring(0, 200),
price,
source,
source_product_id: `${source}_${Date.now()}_${Math.random().toString(36).substr(2, 6)}`,
source_url: `https://www.${source}.com`,
brand: '',
category: 'parapharmacy',
available: true,
scraped_at: new Date().toISOString()
});
}
}
}
// Deduplicate
const seen = new Set();
return products.filter(p => {
const key = `${source}:${p.name.toLowerCase()}`;
if (seen.has(key)) return false;
seen.add(key);
return true;
}).slice(0, 5);
}
async function main() {
console.log('🚀 Starting parapharmacy scraper...\n');
let totalProducts = 0;
for (const query of QUERIES) {
console.log(`\n📋 Query: "${query}"`);
for (const [source, urlFn] of Object.entries(SOURCES)) {
const url = urlFn(query);
const html = await scrapeSource(source, url);
if (html) {
const products = extractProducts(html, source);
console.log(` ✅ Found ${products.length} products`);
if (products.length > 0) {
try {
await axios.post(`${API_URL}/api/products/bulk`, { products });
totalProducts += products.length;
} catch (error) {
console.error(` ❌ Error sending to API: ${error.message}`);
}
}
}
}
}
console.log(`\n📊 Scraping completed. Total products: ${totalProducts}`);
}
main().catch(error => {
console.error('❌ Scraper failed:', error.message);
process.exit(1);
});