amazon_review_en_2026-06-04

amazon_review_en.js

// test page: https://www.amazon.com/Owala-FreeSip-Insulated-Stainless-BPA-Free/dp/B0BZYCJK89/ref=zg_bs_c_kitchen_d_sccl_1/138-0782883-4241845?pd_rd_w=if96h&content-id=amzn1.sym.fef9af56-6177-46e9-8710-a5293a68dd39&pf_rd_p=fef9af56-6177-46e9-8710-a5293a68dd39&pf_rd_r=D3BVKF50GMF969E11QAG&pd_rd_wg=yc1zf&pd_rd_r=2ef35c7f-0715-48e9-85cb-b626a242dde4&pd_rd_i=B0BZYCJK89&th=1

// ============================================================
// ▶ User configuration (edit as needed)
// ============================================================
const MAX_PAGE_COUNT = 10;                          // Number of pages to collect (1 = current page only)
const WAIT_AFTER_LOAD = 5000;                       // Wait time after a new page loads (ms)
const WAIT_BEFORE_NAV = 1000;                       // Wait time after scrolling, before navigating (ms)
const REVIEW_LIST_SELECTOR = '[data-hook="review"]';                 // Individual review item selector
const PAGINATION_SELECTOR = '#reviewsMedley .a-pagination, ul.a-pagination'; // Pagination area selector
const NEXT_BTN_SELECTOR = 'li.a-last a';            // "Next page" link selector (reviews page only)
const CONTAINER_ID = '__review_collector';          // Collector container ID

// ============================================================
// ▶ Internal logic (no edits needed)
// ============================================================
(async () => {
  const sleep = (ms) => new Promise((r) => setTimeout(r, ms));

  // 1. Create the collector container (reuse if it already exists)
  let container = document.getElementById(CONTAINER_ID);
  if (!container) {
    container = document.createElement('div');
    container.id = CONTAINER_ID;
    Object.assign(container.style, {
      position: 'fixed',
      right: '200px',
      bottom: '200px',
      zIndex: '999999',
      background: 'rgba(0, 0, 0, 0.85)',
      color: '#0f0',
      fontSize: '18px',
      fontWeight: 'bold',
      fontFamily: 'monospace',
      padding: '16px 24px',
      borderRadius: '12px',
      border: '2px solid #0f0',
      boxShadow: '0 4px 20px rgba(0,255,0,.3)',
      cursor: 'default',
      userSelect: 'none',
      overflow: 'hidden',
    });
    document.body.appendChild(container);
  }

  // Collection store (hidden area) — a display:none wrapper inside the container
  let store = container.querySelector('#__review_store');
  if (!store) {
    store = document.createElement('div');
    store.id = '__review_store';
    store.style.display = 'none';
    container.appendChild(store);
  }

  // Count display label
  let label = container.querySelector('#__review_label');
  if (!label) {
    label = document.createElement('div');
    label.id = '__review_label';
    container.appendChild(label);
  }

  const updateLabel = () => {
    const count = store.children.length;
    label.textContent = `Reviews collected: ${count}`;
  };

  // 2. Function to collect reviews on the current page
  const collectCurrentPage = () => {
    const reviews = document.querySelectorAll(
      `${REVIEW_LIST_SELECTOR}:not(#${CONTAINER_ID} *)`
    );
    let added = 0;
    reviews.forEach((el) => {
      const clone = el.cloneNode(true);
      store.appendChild(clone);
      added++;
    });
    console.log(`[Scraper] Collected ${added} reviews (total: ${store.children.length})`);
    updateLabel();
  };

  // 3. Function to go to the next page (scroll → wait → navigate)
  //    Returns: true = navigation triggered, false = no next page
  //    NOTE: Amazon advances reviews by loading a new URL (?pageNumber=N),
  //    not by an in-page AJAX button like Kurly. We persist the running
  //    total in sessionStorage so it survives the page reload.
  const goNextPage = async () => {
    // Scroll the pagination area into view
    const paginationArea = document.querySelector(PAGINATION_SELECTOR);
    if (paginationArea) {
      paginationArea.scrollIntoView({ behavior: 'smooth', block: 'center' });
    }
    await sleep(WAIT_BEFORE_NAV);

    // Prefer the page's own "Next" link if present (reviews page)
    const btn = document.querySelector(NEXT_BTN_SELECTOR);
    if (btn && !btn.closest('.a-disabled') && !btn.classList.contains('a-disabled')) {
      // Save progress before the page reloads
      sessionStorage.setItem('__review_store_html', store.innerHTML);
      btn.click();
      return true;
    }

    // Fallback: build the next page URL manually
    const url = new URL(window.location.href);
    const current = parseInt(url.searchParams.get('pageNumber') || '1', 10);
    // Only the dedicated reviews page supports pageNumber pagination
    if (!/\/product-reviews\//.test(url.pathname)) {
      return false;
    }
    url.searchParams.set('pageNumber', String(current + 1));
    sessionStorage.setItem('__review_store_html', store.innerHTML);
    window.location.href = url.toString();
    return true;
  };

  // Restore any previously saved progress (after a page navigation)
  const saved = sessionStorage.getItem('__review_store_html');
  if (saved && store.children.length === 0) {
    store.innerHTML = saved;
    updateLabel();
  }

  // 4. Main loop
  console.log(`[Scraper] Starting collection — up to ${MAX_PAGE_COUNT} pages`);

  for (let page = 1; page <= MAX_PAGE_COUNT; page++) {
    // Collect the current page
    collectCurrentPage();
    console.log(`[Scraper] Page ${page}/${MAX_PAGE_COUNT} collected`);

    // Stop if this is the last page
    if (page === MAX_PAGE_COUNT) break;

    // Move to the next page
    const moved = await goNextPage();
    if (!moved) {
      console.log('[Scraper] No next page — collection finished');
      break;
    }

    // Wait for the page to load
    await sleep(WAIT_AFTER_LOAD);
  }

  console.log(`[Scraper] Collection complete! Total reviews: ${store.children.length}`);
  updateLabel();

  // ▶ To retrieve the collected HTML later, run in the console:
  //   document.getElementById('__review_store').innerHTML
})();