sync

scraper.py

@@ -102,7 +102,17 @@ def get_rating_distribution(url: str) -> dict:
     """Pobiera pięć zbiorczych liczników. Treści opinii nie są odczytywane ani zapisywane."""
     try:
         with sync_playwright() as playwright:
-            launch = {"headless": True, "args": ["--disable-dev-shm-usage", "--no-first-run", "--disable-gpu"]}
+            headless = os.environ.get("CHROMIUM_HEADLESS", "1").lower() not in {"0", "false", "no"}
+            launch = {
+                "headless": headless,
+                "ignore_default_args": ["--enable-automation"],
+                "args": [
+                    "--disable-dev-shm-usage",
+                    "--no-first-run",
+                    "--disable-gpu",
+                    "--disable-blink-features=AutomationControlled",
+                ],
+            }
             if os.environ.get("CHROMIUM_PATH"):
                 launch["executable_path"] = os.environ["CHROMIUM_PATH"]
             profile_path = os.environ.get("CHROMIUM_PROFILE_PATH")
@@ -122,6 +132,9 @@ def get_rating_distribution(url: str) -> dict:
                 context = browser.new_context(locale="pl-PL", viewport={"width": 1100, "height": 800})
             try:
                 page = context.pages[0] if context.pages else context.new_page()
+                context.add_init_script(
+                    "Object.defineProperty(navigator, 'webdriver', {get: () => undefined})"
+                )
                 page.goto(url, wait_until="domcontentloaded", timeout=30_000)
                 _accept_cookies(page)
                 _open_reviews(page)