vbnotebook

30 OpenTextListing.py

@@ -0,0 +1,171 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+
+import json
+from pathlib import Path
+from playwright.sync_api import sync_playwright
+
+# =============================================================
+# CONFIGURATION
+# =============================================================
+
+COOKIE_FILE = Path("sktorrent_cookies.json")
+URL = "https://sktorrent.eu/torrent/torrents.php?active=0"
+
+
+def load_cookies(context):
+    """Load saved cookies if available."""
+    if COOKIE_FILE.exists():
+        with open(COOKIE_FILE, "r") as f:
+            cookies = json.load(f)
+        context.add_cookies(cookies)
+        print("🔄 Loaded login cookies.")
+        return True
+    print("❌ Cookie file not found. Run manual login first.")
+    return False
+
+
+# =============================================================
+# MAIN CODE
+# =============================================================
+
+with sync_playwright() as p:
+
+    # 1️⃣ Launch browser
+    browser = p.chromium.launch(
+        headless=False,
+        args=[
+            "--disable-popup-blocking",
+            "--disable-background-networking",
+            "--disable-notifications",
+            "--no-default-browser-check",
+            "--no-first-run",
+            "--noerrdialogs",
+            "--disable-dev-shm-usage",
+            "--disable-features=IsolateOrigins,site-per-process",
+            "--no-sandbox",
+        ]
+    )
+
+    # 2️⃣ Create context before any pages exist
+    context = browser.new_context()
+
+    # 3️⃣ Block ALL third-party requests (ads, JS, popups, tracking)
+    def block_third_party(route, request):
+        url = request.url.lower()
+        if "sktorrent.eu" in url:
+            route.continue_()
+        else:
+            print(f"🚫 Blocked third-party request: {url}")
+            route.abort()
+
+    context.route("**/*", block_third_party)
+
+    # 4️⃣ Block ANY popup windows except the first page
+    pages = []
+
+    def on_new_page(new_page):
+        pages.append(new_page)
+        if len(pages) == 1:
+            print("➡️ Main page created.")
+        else:
+            print("⚠️ Popup blocked (auto-closed).")
+            new_page.close()
+
+    context.on("page", on_new_page)
+
+    # 5️⃣ Disable all popup JS functions (window.open, window.close, opener.close)
+    context.add_init_script("""
+        window.open = () => { console.log("Blocked window.open"); return null; };
+        window.close = () => { console.log("Blocked window.close"); };
+
+        try {
+            if (window.opener) {
+                window.opener.close = () => { console.log("Blocked opener.close"); };
+            }
+        } catch (e) {}
+
+        // Block <a target="_blank">
+        document.addEventListener('click', function(e) {
+            const el = e.target.closest('a[target="_blank"]');
+            if (el) {
+                e.preventDefault();
+                console.log("Blocked target=_blank");
+            }
+        }, true);
+
+        // Block middle-click opening a new tab
+        document.addEventListener('auxclick', function(e) {
+            e.preventDefault();
+        }, true);
+    """)
+
+    # 6️⃣ Create the FIRST page (main page)
+    page = context.new_page()
+    pages.append(page)
+
+    # 7️⃣ Load cookies (login)
+    load_cookies(context)
+
+    # 8️⃣ Navigate
+    print("🌍 Opening page...")
+    page.goto(URL)
+
+    # Do NOT use networkidle on ad-heavy sites
+    page.wait_for_load_state("domcontentloaded")
+    page.wait_for_selector("table tr", timeout=15000)
+    # Remove popup/overlay elements created by SKTorrent
+    page.evaluate("""
+        const selectors = [
+            '#lightbox', '.lightbox', '#popup', '.popup', 
+            '.overlay', '#overlay', '.modal', '#modal',
+            'div[style*="fixed"]', 'div[style*="position: fixed"]',
+            'table[style*="position: fixed"]',
+            'iframe', 'frame'
+        ];
+
+        selectors.forEach(sel => {
+            document.querySelectorAll(sel).forEach(el => {
+                console.log("Removing popup element:", sel);
+                el.remove();
+            });
+        });
+
+        // Remove onclick handlers that trigger popups
+        document.querySelectorAll('*').forEach(el => {
+            el.onclick = null;
+            el.onauxclick = null;
+            el.oncontextmenu = null;
+        });
+
+        // Remove timers that trigger delayed popups
+        window.setTimeout = () => {};
+        window.setInterval = () => {};
+    """)
+
+    print("✔ Page loaded, extracting table rows...")
+
+    # 9️⃣ Extract all rows
+    rows = page.locator("table tr").all()
+    print(f"📄 Total rows found (including header): {len(rows)}")
+
+    # 🔟 Extract SECOND ROW only (your request)
+    if len(rows) > 1:
+        row = rows[1]  # 0 = header, 1 = first data row
+        tds = row.locator("td")
+
+        name = tds.nth(1).inner_text().strip()
+        size = tds.nth(2).inner_text().strip()
+        seeders = tds.nth(3).inner_text().strip()
+        leechers = tds.nth(4).inner_text().strip()
+
+        print("\n========= SECOND ROW =========")
+        print(f"Name: {name}")
+        print(f"Size: {size}")
+        print(f"Seeders: {seeders}")
+        print(f"Leechers: {leechers}")
+        print("==============================\n")
+    else:
+        print("❌ No data rows found!")
+
+    page.wait_for_timeout(5000)