ThemeForestのウェブデータをスクレイピングする方法

from playwright.sync_api import sync_playwright

def scrape_themeforest():
    with sync_playwright() as p:
        browser = p.chromium.launch(headless=True)
        page = browser.new_page()
        # WordPressカテゴリーへ移動
        page.goto('https://themeforest.net/category/wordpress')
        # リスティングアイテムがロードされるまで待機
        page.wait_for_selector('li.search-grid__item')
        
        items = page.query_selector_all('li.search-grid__item')
        for item in items:
            title = item.query_selector('h3').inner_text()
            sales = item.query_selector('.item-thumbnail__sales').inner_text()
            print(f'見つかったアイテム: {title} | 売上: {sales}')
        
        browser.close()

scrape_themeforest()

import scrapy

class ThemeForestSpider(scrapy.Spider):
    name = 'themeforest'
    start_urls = ['https://themeforest.net/category/wordpress']

    def parse(self, response):
        # 検索グリッドのアイテムをループ処理
        for item in response.css('li.search-grid__item'):
            yield {
                'title': item.css('h3 a::text').get().strip(),
                'price': item.css('.price::text').get(),
                'sales': item.css('.item-thumbnail__sales::text').get(),
                'url': response.urljoin(item.css('h3 a::attr(href)').get())
            }
        
        # ページネーションの処理
        next_page = response.css('a.next_page::attr(href)').get()
        if next_page:
            yield response.follow(next_page, self.parse)

const puppeteer = require('puppeteer-extra');
const StealthPlugin = require('puppeteer-extra-plugin-stealth');
puppeteer.use(StealthPlugin());

(async () => {
  const browser = await puppeteer.launch({ headless: true });
  const page = await browser.newPage();
  await page.goto('https://themeforest.net/category/wordpress');
  
  // プロダクトカードが表示されるまで待機
  await page.waitForSelector('li.search-grid__item');
  
  const data = await page.evaluate(() => {
    const themes = Array.from(document.querySelectorAll('li.search-grid__item'));
    return themes.map(el => ({
      title: el.querySelector('h3').innerText.trim(),
      price: el.querySelector('.price').innerText.trim()
    }));
  });
  
  console.log(data);
  await browser.close();
})();