← back to Handbag Authentication
crawler/mercari.js
105 lines
const puppeteer = require('puppeteer');
class MercariJPScraper {
constructor() {
this.baseUrl = 'https://jp.mercari.com/search';
this.brands = ['グッチ バッグ', 'ルイヴィトン バッグ', 'シャネル バッグ', 'エルメス バッグ'];
}
async scrapeSearch(brand, page = 1) {
try {
const browser = await puppeteer.launch({
headless: 'new',
args: ['--no-sandbox', '--disable-setuid-sandbox']
});
const browserPage = await browser.newPage();
await browserPage.setUserAgent('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36');
const searchUrl = `${this.baseUrl}?keyword=${encodeURIComponent(brand)}&page=${page}`;
console.log(`Scraping Mercari JP: ${searchUrl}`);
await browserPage.goto(searchUrl, { waitUntil: 'networkidle2', timeout: 30000 });
await browserPage.waitForSelector('[data-testid="item-cell"]', { timeout: 10000 }).catch(() => {});
const listings = await browserPage.evaluate(() => {
const items = [];
const productElements = document.querySelectorAll('[data-testid="item-cell"]');
productElements.forEach((item) => {
try {
const titleEl = item.querySelector('span');
const title = titleEl?.textContent?.trim() || '';
const priceEl = item.querySelector('[data-testid="price"]');
const priceText = priceEl?.textContent?.trim() || '';
const priceMatch = priceText.match(/[\d,]+/);
const price = priceMatch ? parseInt(priceMatch[0].replace(/,/g, '')) : 0;
const imgEl = item.querySelector('img');
const image = imgEl?.src || '';
const linkEl = item.querySelector('a');
const url = linkEl?.href || '';
const idMatch = url.match(/\/items\/([a-zA-Z0-9]+)/);
const id = idMatch?.[1] || url;
if (title && price > 0) {
items.push({
external_id: id,
title,
price_jpy: price,
image_url: image,
thumbnail_url: image,
product_url: url,
listing_type: 'buy',
condition: 'Used'
});
}
} catch (e) {
console.error('Error parsing Mercari item:', e);
}
});
return items;
});
await browser.close();
return listings;
} catch (error) {
console.error(`Error scraping Mercari JP:`, error.message);
return [];
}
}
async scrapeAllBrands() {
const allListings = [];
for (const brand of this.brands) {
console.log(`\nScraping Mercari JP for ${brand}...`);
for (let page = 1; page <= 5; page++) {
const listings = await this.scrapeSearch(brand, page);
if (listings.length === 0) break;
listings.forEach(listing => {
listing.brand = brand;
listing.source = 'mercari_jp';
});
allListings.push(...listings);
console.log(`Found ${listings.length} items on page ${page}`);
await new Promise(resolve => setTimeout(resolve, 2500));
}
}
return allListings;
}
}
module.exports = MercariJPScraper;