From f09b58fb63c0859d755f70ce756d9b990bf5a61e Mon Sep 17 00:00:00 2001
From: Norbert Maciaszek <norbert@ambiscale.com>
Date: Thu, 20 Nov 2025 20:33:04 +0100
Subject: [PATCH] Integrate Playwright for improved web scraping by launching a
 headless browser, enhancing product data retrieval with better error handling
 and dynamic content loading. Update logging for price saving confirmation.

---
 index.js | 67 +++++++++++++++++++++++++++++++++++++++-----------------
 1 file changed, 47 insertions(+), 20 deletions(-)

diff --git a/index.js b/index.js
index 58ba2a0..7d0f8e1 100644
--- a/index.js
+++ b/index.js
@@ -2,6 +2,7 @@ const cheerio = require("cheerio");
 const cron = require("node-cron");
 const axios = require("axios");
 const fs = require("fs");
+const { chromium } = require("playwright");
 
 const discordWebhook =
   "https://discord.com/api/webhooks/1439286509390921749/t2Hb8XloF6zhDRYD1yh_QlkHHa9eHUyXvd9TxZRHwqR_b_OxxbnwDgsm4em8TwA9NQIa";
@@ -49,7 +50,6 @@ async function compareAndSave(productsPrice) {
     for (const product of productsPrice) {
       sendMessage(`Początkowa cena **${product.name}**: ${product.price}`);
     }
-    console.log("First run completed");
     isFirstRun = false;
   }
 
@@ -70,36 +70,63 @@ async function getProducts() {
 }
 
 async function init() {
+  const browser = await chromium.launch({ headless: true });
+  const context = await browser.newContext({
+    userAgent:
+      "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36",
+  });
+  const page = await context.newPage();
+
   const productsIds = await getProducts();
   const products = [];
 
   for (const productId of productsIds) {
-    const response = await axios.get(`https://www.ceneo.pl/${productId}`);
-    const $ = cheerio.load(response.data);
+    try {
+      await page.goto(`https://www.ceneo.pl/${productId}`, {
+        waitUntil: "networkidle",
+        timeout: 30000,
+      });
 
-    const items = $(".product-offer__container").first();
-    for (const item of items) {
-      let name = $(item).data("productname");
-      let price = $(item).data("price");
-      let link = `https://www.ceneo.pl/${$(item).data("click-url")}`;
-      const shop = $(item).data("shopurl") || "ceneo.pl";
+      // Wait for bot detection to pass and content to load
+      await page
+        .waitForSelector(".product-offer__container, .product-top", {
+          timeout: 10000,
+        })
+        .catch(() => {
+          console.log(`Timeout waiting for content on ${productId}`);
+        });
 
-      if (!name || !price) {
-        name = $(item).find(".short-name__txt").text();
-        price = $(item).find(".price").text();
-        link = `https://www.ceneo.pl/${productId}`;
+      const html = await page.content();
+      const $ = cheerio.load(html);
+
+      const items = $(".product-offer__container").first();
+      for (const item of items) {
+        let name = $(item).data("productname");
+        let price = $(item).data("price");
+        let link = `https://www.ceneo.pl/${$(item).data("click-url")}`;
+        const shop = $(item).data("shopurl") || "ceneo.pl";
+
+        if (!name || !price) {
+          name = $(item).find(".short-name__txt").text();
+          price = $(item).find(".price").text();
+          link = `https://www.ceneo.pl/${productId}`;
+        }
+
+        if (!name || !price || !link) {
+          continue;
+        }
+
+        products.push({ name, price, link, shop });
       }
-
-      if (!name || !price || !link) {
-        continue;
-      }
-
-      products.push({ name, price, link, shop });
+    } catch (error) {
+      console.error(`Error fetching product ${productId}:`, error.message);
     }
   }
 
+  await browser.close();
+
   await compareAndSave(products);
-  console.log("Sprawdzone! Aktualne ceny zapisane w productsPrice.json");
+  console.log("Aktualne ceny zapisane w productsPrice.json");
 }
 
 sendMessage("Startuję monitoring cen");