Restructure with Turborepo
This commit is contained in:
@@ -0,0 +1,61 @@
|
||||
const puppeteer = require('puppeteer-extra');
|
||||
const StealthPlugin = require('puppeteer-extra-plugin-stealth');
|
||||
const fs = require('fs');
|
||||
|
||||
const scrapePromofarma = require('./src/promofarma');
|
||||
const scrapeAmazon = require('./src/amazon');
|
||||
const scrapeGoogleShopping = require('./src/googleshopping');
|
||||
const scrapeOpenFoodFacts = require('./src/openfoodfacts');
|
||||
|
||||
puppeteer.use(StealthPlugin());
|
||||
|
||||
// Default terms related to parapharmacy
|
||||
const QUERIES = [
|
||||
'crema hidratante cara',
|
||||
'ibuprofeno' // fallback basic test query
|
||||
];
|
||||
|
||||
async function run() {
|
||||
console.log('🚀 Starting Parapharmacy Scraper...');
|
||||
// Launching browser
|
||||
const browser = await puppeteer.launch({
|
||||
headless: 'new',
|
||||
args: ['--no-sandbox', '--disable-setuid-sandbox']
|
||||
});
|
||||
|
||||
const allResults = {};
|
||||
|
||||
for (const query of QUERIES) {
|
||||
console.log(`\n🔍 Searching for: "${query}"`);
|
||||
|
||||
// We run sequentially to be gentle on CPU and anti-bot systems
|
||||
const promofarmaResults = await scrapePromofarma(query, browser);
|
||||
console.log(` ✅ Promofarma: found ${promofarmaResults.length} items`);
|
||||
|
||||
const amazonResults = await scrapeAmazon(query, browser);
|
||||
console.log(` ✅ Amazon: found ${amazonResults.length} items`);
|
||||
|
||||
const googleShoppingResults = await scrapeGoogleShopping(query, browser);
|
||||
console.log(` ✅ Google Shopping: found ${googleShoppingResults.length} items`);
|
||||
|
||||
const ofFactsResults = await scrapeOpenFoodFacts(query, browser);
|
||||
console.log(` ✅ OpenFoodFacts: found ${ofFactsResults.length} items`);
|
||||
|
||||
allResults[query] = {
|
||||
promofarma: promofarmaResults,
|
||||
amazon: amazonResults,
|
||||
googleshopping: googleShoppingResults,
|
||||
openfoodfacts: ofFactsResults
|
||||
};
|
||||
}
|
||||
|
||||
await browser.close();
|
||||
|
||||
fs.writeFileSync('results.json', JSON.stringify(allResults, null, 2));
|
||||
console.log('\n💾 Scraping finished. Results saved to results.json');
|
||||
}
|
||||
|
||||
run().catch(err => {
|
||||
console.error("Critical Error during scraping:", err);
|
||||
process.exit(1);
|
||||
});
|
||||
Generated
+1618
File diff suppressed because it is too large
Load Diff
@@ -0,0 +1,18 @@
|
||||
{
|
||||
"name": "scraper",
|
||||
"version": "1.0.0",
|
||||
"description": "",
|
||||
"main": "index.js",
|
||||
"scripts": {
|
||||
"test": "echo \"Error: no test specified\" && exit 1"
|
||||
},
|
||||
"keywords": [],
|
||||
"author": "",
|
||||
"license": "ISC",
|
||||
"type": "commonjs",
|
||||
"dependencies": {
|
||||
"puppeteer": "^24.40.0",
|
||||
"puppeteer-extra": "^3.3.6",
|
||||
"puppeteer-extra-plugin-stealth": "^2.11.2"
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,14 @@
|
||||
{
|
||||
"crema hidratante cara": {
|
||||
"promofarma": [],
|
||||
"amazon": [],
|
||||
"googleshopping": [],
|
||||
"openfoodfacts": []
|
||||
},
|
||||
"ibuprofeno": {
|
||||
"promofarma": [],
|
||||
"amazon": [],
|
||||
"googleshopping": [],
|
||||
"openfoodfacts": []
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,43 @@
|
||||
const puppeteer = require('puppeteer-extra');
|
||||
|
||||
async function scrapeAmazon(query, browser) {
|
||||
const page = await browser.newPage();
|
||||
const url = `https://www.amazon.es/s?k=${encodeURIComponent(query)}`;
|
||||
|
||||
try {
|
||||
await page.goto(url, { waitUntil: 'domcontentloaded', timeout: 30000 });
|
||||
|
||||
const products = await page.evaluate(() => {
|
||||
const results = [];
|
||||
const cards = document.querySelectorAll('[data-component-type="s-search-result"]');
|
||||
|
||||
cards.forEach(card => {
|
||||
const titleEl = card.querySelector('h2 a span');
|
||||
const priceWhole = card.querySelector('.a-price-whole');
|
||||
const priceFraction = card.querySelector('.a-price-fraction');
|
||||
const symbolEl = card.querySelector('.a-price-symbol');
|
||||
const linkEl = card.querySelector('h2 a');
|
||||
|
||||
if (titleEl && priceWhole) {
|
||||
const price = `${priceWhole.innerText.trim()}${priceFraction ? priceFraction.innerText.trim() : ''} ${symbolEl ? symbolEl.innerText.trim() : '€'}`;
|
||||
results.push({
|
||||
name: titleEl.innerText.trim(),
|
||||
price: price,
|
||||
link: linkEl ? linkEl.href : null,
|
||||
source: 'Amazon'
|
||||
});
|
||||
}
|
||||
});
|
||||
return results.slice(0, 5);
|
||||
});
|
||||
|
||||
return products;
|
||||
} catch (error) {
|
||||
console.error(`[Amazon] Error scraping ${query}:`, error.message);
|
||||
return [];
|
||||
} finally {
|
||||
await page.close();
|
||||
}
|
||||
}
|
||||
|
||||
module.exports = scrapeAmazon;
|
||||
@@ -0,0 +1,54 @@
|
||||
const puppeteer = require('puppeteer-extra');
|
||||
|
||||
async function scrapeGoogleShopping(query, browser) {
|
||||
const page = await browser.newPage();
|
||||
const url = `https://www.google.com/search?q=${encodeURIComponent(query)}&tbm=shop`;
|
||||
|
||||
try {
|
||||
await page.goto(url, { waitUntil: 'domcontentloaded', timeout: 30000 });
|
||||
|
||||
const products = await page.evaluate(() => {
|
||||
const results = [];
|
||||
// Google Shopping results are often inside blocks with class .sh-dgr__grid-result or similar
|
||||
const cards = document.querySelectorAll('.sh-dgr__grid-result, .sh-np__click-target');
|
||||
|
||||
cards.forEach(card => {
|
||||
const titleEl = card.querySelector('h3, [role="heading"]');
|
||||
const priceEl = card.querySelector('.a80p0, [data-price], span[aria-hidden="true"]:not(:empty)');
|
||||
// Fallback for price: look for any element mimicking a price format
|
||||
let priceText = null;
|
||||
if (priceEl) {
|
||||
priceText = priceEl.innerText.trim();
|
||||
} else {
|
||||
const spans = card.querySelectorAll('span');
|
||||
spans.forEach(span => {
|
||||
if (span.innerText.includes('€') || span.innerText.match(/\d+,\d{2}/)) {
|
||||
priceText = span.innerText.trim();
|
||||
}
|
||||
});
|
||||
}
|
||||
|
||||
const linkEl = card.tagName === 'A' ? card : card.querySelector('a');
|
||||
|
||||
if (titleEl && priceText) {
|
||||
results.push({
|
||||
name: titleEl.innerText.trim(),
|
||||
price: priceText,
|
||||
link: linkEl ? linkEl.href : null,
|
||||
source: 'Google Shopping'
|
||||
});
|
||||
}
|
||||
});
|
||||
return results.slice(0, 5);
|
||||
});
|
||||
|
||||
return products;
|
||||
} catch (error) {
|
||||
console.error(`[Google Shopping] Error scraping ${query}:`, error.message);
|
||||
return [];
|
||||
} finally {
|
||||
await page.close();
|
||||
}
|
||||
}
|
||||
|
||||
module.exports = scrapeGoogleShopping;
|
||||
@@ -0,0 +1,44 @@
|
||||
const puppeteer = require('puppeteer-extra');
|
||||
|
||||
async function scrapeOpenFoodFacts(query, browser) {
|
||||
const page = await browser.newPage();
|
||||
const url = `https://world.openfoodfacts.org/cgi/search.pl?search_terms=${encodeURIComponent(query)}&search_simple=1&action=process`;
|
||||
|
||||
try {
|
||||
await page.goto(url, { waitUntil: 'domcontentloaded', timeout: 30000 });
|
||||
|
||||
const products = await page.evaluate(() => {
|
||||
const results = [];
|
||||
const items = document.querySelectorAll('#search_results li');
|
||||
|
||||
items.forEach(item => {
|
||||
const linkEl = item.querySelector('a');
|
||||
const titleEl = item.querySelector('span:not([class])') || item.querySelector('.list_product_name'); // heuristic
|
||||
|
||||
let link = linkEl ? linkEl.href : null;
|
||||
if (link && !link.startsWith('http')) {
|
||||
link = `https://world.openfoodfacts.org${link}`;
|
||||
}
|
||||
|
||||
if (linkEl && linkEl.title) {
|
||||
results.push({
|
||||
name: linkEl.title.trim(),
|
||||
price: 'N/A', // OFF usually doesn't have prices
|
||||
link: link,
|
||||
source: 'OpenFoodFacts'
|
||||
});
|
||||
}
|
||||
});
|
||||
return results.slice(0, 5);
|
||||
});
|
||||
|
||||
return products;
|
||||
} catch (error) {
|
||||
console.error(`[OpenFoodFacts] Error scraping ${query}:`, error.message);
|
||||
return [];
|
||||
} finally {
|
||||
await page.close();
|
||||
}
|
||||
}
|
||||
|
||||
module.exports = scrapeOpenFoodFacts;
|
||||
@@ -0,0 +1,51 @@
|
||||
const puppeteer = require('puppeteer-extra');
|
||||
|
||||
async function scrapePromofarma(query, browser) {
|
||||
const page = await browser.newPage();
|
||||
const url = `https://www.promofarma.com/es/search?q=${encodeURIComponent(query)}`;
|
||||
|
||||
try {
|
||||
await page.goto(url, { waitUntil: 'domcontentloaded', timeout: 30000 });
|
||||
|
||||
// Fallback/heuristic extraction
|
||||
const products = await page.evaluate(() => {
|
||||
const results = [];
|
||||
const cards = document.querySelectorAll('.product-card, article, [class*="product"]');
|
||||
|
||||
cards.forEach(card => {
|
||||
const titleEl = card.querySelector('h3, h2, [class*="title"], [class*="name"]');
|
||||
const priceEl = card.querySelector('[class*="price"], .current-price');
|
||||
const linkEl = card.querySelector('a');
|
||||
|
||||
if (titleEl && priceEl && titleEl.innerText.trim().length > 0) {
|
||||
results.push({
|
||||
name: titleEl.innerText.trim(),
|
||||
price: priceEl.innerText.trim(),
|
||||
link: linkEl ? linkEl.href : null,
|
||||
source: 'Promofarma'
|
||||
});
|
||||
}
|
||||
});
|
||||
|
||||
// Deduplicate keeping the first occurrence
|
||||
const unique = [];
|
||||
const names = new Set();
|
||||
for (const p of results) {
|
||||
if (!names.has(p.name)) {
|
||||
names.add(p.name);
|
||||
unique.push(p);
|
||||
}
|
||||
}
|
||||
return unique.slice(0, 5); // Return top 5
|
||||
});
|
||||
|
||||
return products;
|
||||
} catch (error) {
|
||||
console.error(`[Promofarma] Error scraping ${query}:`, error.message);
|
||||
return [];
|
||||
} finally {
|
||||
await page.close();
|
||||
}
|
||||
}
|
||||
|
||||
module.exports = scrapePromofarma;
|
||||
Reference in New Issue
Block a user