Switched to Zendriver

Xewdy444 · Feb 11, 2025 · 23313c1 · 23313c1
1 parent fb1d064
commit 23313c1
Show file tree

Hide file tree

Showing 3 changed files with 26 additions and 75 deletions.
diff --git a/README.md b/README.md
@@ -1,6 +1,6 @@
 # CF-Clearance-Scraper
 
-## nodriver Version
+## Zendriver Version
 A simple program for scraping Cloudflare clearance (cf_clearance) cookies from websites issuing Cloudflare challenges to visitors. This program works on all Cloudflare challenge types (JavaScript, managed, and interactive). If you would prefer using Playwright, you can check out the [Playwright version](https://github.com/Xewdy444/CF-Clearance-Scraper/tree/playwright).
 
 
@@ -9,7 +9,7 @@ In order to bypass Cloudflare challenges with the clearance cookies, you must ma
 
 - The user agent used to fetch the clearance cookie must match the user agent being used within the requests that use the clearance cookie
 > [!NOTE]
-> The default user agent used by the scraper is `Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/132.0.0.0 Safari/537.36`.
+> The default user agent used by the scraper is `Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/133.0.0.0 Safari/537.36`.
 - The IP address used to fetch the clearance cookie must match the IP address being used to make the requests that use the clearance cookie
 
 ```mermaid
@@ -29,9 +29,6 @@ Then, install the Python dependencies:
     $ pip install -r requirements.txt
 
 ## Usage
-> [!NOTE]
-> Headless mode is not supported on Windows.
-
 > [!WARNING]
 > Depending on the user agent used, it may affect your ability to solve the Cloudflare challenge.
 
@@ -67,5 +64,5 @@ options:
     [14:24:27] [INFO] Going to https://sergiodemo.com/security/challenge/legacy-challenge...
     [14:24:28] [INFO] Solving Cloudflare challenge [Interactive]...
     [14:24:31] [INFO] Cookie: cf_clearance=SkyEdEGvKp1BBA2NpRW3Azsw5neMD6sEEqJd6jOCCfs-1736886257-1.2.1.1-cam47ywp3q_yKE1bw0lZ2YS83dnh_BsIHtS7earbsYE.AxQDBtZiifiHvp1nZGRhABaSdjU7XRQpUCVwUSrlJGH8DXr50YR18pNLxBvcEJFO2gPMxr.ZjKze8rWgM9H4rPeET67jzAo_ZRpNP6hGCvdyO62VVCtqDBQDKhKZz9yZQp7YEHK7tchQIteVgu.dUxYdan5_D.R0zewnS382BP0w1AoTf2p40.lQwbhgildEiKG14xACd13V4EEthkZV0dnliwcn35rT3h32ODf50MABQNSQ8WjhZhbLSNOPO_zEhrK9R0Yn4eBuRKvnL9_x9jKvaBPDPAgyiZv_VzFP_g
-    [14:24:31] [INFO] User agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/132.0.0.0 Safari/537.36
+    [14:24:31] [INFO] User agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/133.0.0.0 Safari/537.36
     [14:24:31] [INFO] Writing Cloudflare clearance cookie information to cookies.json...
diff --git a/main.py b/main.py
@@ -2,43 +2,24 @@
 
 import argparse
 import asyncio
-import io
 import json
 import logging
-import sys
 from datetime import datetime, timedelta, timezone
 from enum import Enum
-from typing import Any, Dict, Final, Iterable, List, Optional, TypedDict
+from typing import Any, Dict, Final, Iterable, List, Optional
 
-import nest_asyncio
-import nodriver
-from nodriver.cdp.network import Cookie
-from nodriver.core.element import Element
+import zendriver
 from selenium_authenticated_proxy import SeleniumAuthenticatedProxy
-
-if sys.platform != "win32":
-    from xvfbwrapper import Xvfb
+from zendriver.cdp.network import T_JSON_DICT, Cookie
+from zendriver.core.element import Element
 
 COMMAND: Final[str] = (
     '{name}: {binary} --header "Cookie: {cookies}" --header "User-Agent: {user_agent}" {url}'
 )
 
 
-class JSONCookie(TypedDict):
-    """A type for representing a JSON cookie."""
-
-    name: str
-    value: str
-    domain: str
-    path: str
-    expires: float
-    httpOnly: bool
-    secure: bool
-    sameSite: str
-
-
-class NodriverOptions(list):
-    """A class for managing nodriver options."""
+class ZendriverOptions(list):
+    """A class for managing zendriver options."""
 
     def add_argument(self, arg: str) -> None:
         """
@@ -90,7 +71,7 @@ def __init__(
         headless: bool,
         proxy: Optional[str],
     ) -> None:
-        options = NodriverOptions()
+        options = ZendriverOptions()
         options.add_argument(f"--user-agent={user_agent}")
 
         if not http2:
@@ -99,34 +80,23 @@ def __init__(
         if not http3:
             options.add_argument("--disable-quic")
 
-        if headless and sys.platform == "win32":
-            raise Exception("Headless mode is not supported on Windows.")
-
-        self._virtual_display = Xvfb() if headless else None
-
         if proxy is not None:
             auth_proxy = SeleniumAuthenticatedProxy(proxy, use_legacy_extension=True)
             auth_proxy.enrich_chrome_options(options)
 
-        config = nodriver.Config(browser_args=options, sandbox=False)
-        self.driver = nodriver.Browser(config)
+        config = zendriver.Config(headless=headless, browser_args=options)
+        self.driver = zendriver.Browser(config)
         self._timeout = timeout
 
     async def __aenter__(self) -> CloudflareSolver:
-        if self._virtual_display is not None:
-            self._virtual_display.start()
-
         await self.driver.start()
         return self
 
     async def __aexit__(self, *_: Any) -> None:
-        self.driver.stop()
-
-        if self._virtual_display is not None:
-            self._virtual_display.stop()
+        await self.driver.stop()
 
     @staticmethod
-    def _format_cookies(cookies: Iterable[Cookie]) -> List[JSONCookie]:
+    def _format_cookies(cookies: Iterable[Cookie]) -> List[T_JSON_DICT]:
         """
         Format cookies into a list of JSON cookies.
 
@@ -137,38 +107,26 @@ def _format_cookies(cookies: Iterable[Cookie]) -> List[JSONCookie]:
 
         Returns
         -------
-        List[JSONCookie]
+        List[T_JSON_DICT]
             List of JSON cookies.
         """
-        return [
-            JSONCookie(
-                name=cookie["name"],
-                value=cookie["value"],
-                domain=cookie["domain"],
-                path=cookie["path"],
-                expires=cookie["expires"],
-                httpOnly=cookie["httpOnly"],
-                secure=cookie["secure"],
-                sameSite=cookie["sameSite"],
-            )
-            for cookie in [cookie.to_json() for cookie in cookies]
-        ]
+        return [cookie.to_json() for cookie in cookies]
 
     @staticmethod
     def extract_clearance_cookie(
-        cookies: Iterable[JSONCookie],
-    ) -> Optional[JSONCookie]:
+        cookies: Iterable[T_JSON_DICT],
+    ) -> Optional[T_JSON_DICT]:
         """
         Extract the Cloudflare clearance cookie from a list of cookies.
 
         Parameters
         ----------
-        cookies : Iterable[JSONCookie]
+        cookies : Iterable[T_JSON_DICT]
             List of cookies.
 
         Returns
         -------
-        Optional[JSONCookie]
+        Optional[T_JSON_DICT]
             The Cloudflare clearance cookie. Returns None if the cookie is not found.
         """
 
@@ -178,13 +136,13 @@ def extract_clearance_cookie(
 
         return None
 
-    async def get_cookies(self) -> List[JSONCookie]:
+    async def get_cookies(self) -> List[T_JSON_DICT]:
         """
         Get all cookies from the current page.
 
         Returns
         -------
-        List[JSONCookie]
+        List[T_JSON_DICT]
             List of cookies.
         """
         return self._format_cookies(await self.driver.cookies.get_all())
@@ -198,7 +156,7 @@ async def detect_challenge(self) -> Optional[ChallengePlatform]:
         Optional[ChallengePlatform]
             The Cloudflare challenge platform.
         """
-        html: str = await self.driver.main_tab.get_content()
+        html = await self.driver.main_tab.get_content()
 
         for platform in ChallengePlatform:
             if f"cType: '{platform.value}'" in html:
@@ -282,7 +240,7 @@ async def main() -> None:
     parser.add_argument(
         "-ua",
         "--user-agent",
-        default="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/132.0.0.0 Safari/537.36",
+        default="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/133.0.0.0 Safari/537.36",
         help="The user agent to use for the browser requests",
         type=str,
     )
@@ -334,16 +292,14 @@ async def main() -> None:
     )
 
     args = parser.parse_args()
-    sys.stdout = io.StringIO()
-    nest_asyncio.apply()
 
     logging.basicConfig(
         format="[%(asctime)s] [%(levelname)s] %(message)s",
         datefmt="%H:%M:%S",
         level=logging.INFO,
     )
 
-    logging.getLogger("nodriver").setLevel(logging.WARNING)
+    logging.getLogger("zendriver").setLevel(logging.WARNING)
     logging.info("Launching %s browser...", "headed" if args.headed else "headless")
 
     challenge_messages = {

diff --git a/requirements.txt b/requirements.txt
@@ -1,4 +1,2 @@
-nest_asyncio==1.6.0
-nodriver==0.39
 selenium_authenticated_proxy==1.1.2
-xvfbwrapper==0.2.9; platform_system != "Windows"
+zendriver==0.4.1