ney crawler

ney.js

const puppeteer = require('puppeteer');
const fs = require('fs');

(async () => {

  const browser = await puppeteer.launch({ headless: true });
  const page = await browser.newPage();

  const totalPages = 3;
  const fileName = 'reviews.txt';

  fs.writeFileSync(fileName, '', 'utf-8');

  for (let pageNumber = 1; pageNumber <= totalPages; pageNumber++) {
    const url = `https://www.amazon.com.br/product-reviews/B09329F4J6/ref=cm_cr_getr_d_paging_btm_prev_${pageNumber}?pageNumber=${pageNumber}`;
    await page.goto(url);

    await page.waitForSelector('span[data-hook="review-body"]');

    const reviews = await page.$$eval('div[data-hook="review"]', reviews => {
      return reviews.map(review => {
        const name = review.querySelector('span.a-profile-name')?.innerText || 'No name';
        const starRating = review.querySelector('i[data-hook="review-star-rating"] span')?.innerText || 'No rating';
        const reviewDate = review.querySelector('span[data-hook="review-date"]')?.innerText || 'No date';
        const reviewText = review.querySelector('span[data-hook="review-body"] span')?.innerText || 'No review text';

        return {
          name,
          starRating,
          reviewDate,
          reviewText
        };
      });
    });

    let log = `--- Page ${pageNumber} ---\n`;
    reviews.forEach((review, index) => {
      log += `Review ${index + 1}:\n`;
      log += `Name: ${review.name}\n`;
      log += `Star Rating: ${review.starRating}\n`;
      log += `Review Date: ${review.reviewDate}\n`;
      log += `Review Text: ${review.reviewText}\n`;
      log += `-------------------------------\n`;
    });

    fs.appendFileSync(fileName, log, 'utf-8');
  }

  await browser.close();
})();