Open target URL during Chrome startup

crawler/driver.py

@@ -13,9 +13,10 @@ import tempfile
 import time
 from contextlib import contextmanager
 from pathlib import Path
-from typing import Generator
+from typing import Any, Generator, cast
 
 import undetected_chromedriver as uc
+from selenium.common.exceptions import WebDriverException
 from selenium.webdriver.chrome.options import Options
 from selenium_stealth import stealth
 
@@ -90,10 +91,13 @@ def _build_options(
     chrome_binary: str,
     profile_dir: str | None = None,
     display: str | None = None,
+    initial_url: str | None = None,
 ) -> Options:
     opts = Options()
     opts.binary_location = chrome_binary
-    opts.page_load_strategy = "eager"
+    # Never let ChromeDriver session creation block on the initial New Tab
+    # renderer. Scenarios perform their own explicit URL/readiness waits.
+    opts.page_load_strategy = "none"
     if headless:
         opts.add_argument("--headless=new")
     if profile_dir:
@@ -108,6 +112,10 @@ def _build_options(
     opts.add_argument("--disable-blink-features=AutomationControlled")
     opts.add_argument("--disable-infobars")
     opts.add_argument("--window-size=1920,1080")
+    if initial_url and initial_url.startswith(("http://", "https://")):
+        # Opening the configured page as a Chrome process argument avoids
+        # waiting for the initial New Tab renderer before navigation begins.
+        opts.add_argument(initial_url)
     return opts
 
 
@@ -115,14 +123,27 @@ def make_driver(
     headless: bool | None = None,
     profile_dir: str | None = None,
     display: str | None = None,
+    initial_url: str | None = None,
 ) -> uc.Chrome:
     """Return a stealthed undetected Chrome instance."""
     use_headless = config.headless if headless is None else headless
     chrome_binary, major = _detect_chrome()
     ua = random.choice(_USER_AGENTS)
-    opts = _build_options(ua, use_headless, chrome_binary, profile_dir, display)
+    opts = _build_options(
+        ua,
+        use_headless,
+        chrome_binary,
+        profile_dir,
+        display,
+        initial_url,
+    )
 
-    log.debug("Starting ChromeDriver (headless=%s, Chrome %d)", use_headless, major)
+    log.info(
+        "Starting ChromeDriver (headless=%s, Chrome %d, initial_url=%s)",
+        use_headless,
+        major,
+        initial_url or "<none>",
+    )
 
     driver = uc.Chrome(
         options=opts,
@@ -130,6 +151,23 @@ def make_driver(
         version_main=major,
         use_subprocess=True,
     )
+    log.info("ChromeDriver connected")
+
+    if initial_url and initial_url.startswith(("http://", "https://")):
+        navigation = cast(
+            dict[str, Any],
+            driver.execute_cdp_cmd(  # type: ignore[reportUnknownMemberType]
+                "Page.navigate", {"url": initial_url}
+            ),
+        )
+        error_text = navigation.get("errorText")
+        if error_text:
+            raise WebDriverException(
+                f"Chrome rejected initial navigation to {initial_url}: {error_text}"
+            )
+        log.info("Initial navigation dispatched to %s", initial_url)
+
+    log.info("Applying stealth settings")
 
     if display:
         driver.set_window_size(1920, 1080)
@@ -191,6 +229,7 @@ def _terminate_profile_processes(profile_dir: str) -> None:
 def driver_session(
     headless: bool | None = None,
     recording_name: str | None = None,
+    initial_url: str | None = None,
 ) -> Generator[uc.Chrome, None, None]:
     profile_dir = tempfile.mkdtemp(prefix="seed-chrome-")
     driver: uc.Chrome | None = None
@@ -207,6 +246,7 @@ def driver_session(
             effective_headless,
             profile_dir=profile_dir,
             display=display,
+            initial_url=initial_url,
         )
         yield driver
     finally: