Selenium webdriver.Remote 驱动程序不适用于 Tor 代理(webdriver.Chrome 可以)

问题描述

我正在尝试在我的 远程驱动程序上使用 socks5 代理,它作为docker 容器在 4444 端口.

这里是代码示例:

from selenium import webdriver

opts = webdriver.ChromeOptions()
opts.add_argument("--no-sandbox")
opts.add_argument("--disable-dev-shm-usage")
opts.add_argument("--proxy-server=socks5://127.0.0.1:9050")
driver = webdriver.Remote(command_executor="http://localhost:4444/wd/hub", desired_capabilities=opts.to_capabilities())

然后,当我尝试打开任何页面时,我收到错误提示 检查您的代理设置或联系您的网络管理员.

Then, when I try to open any page, I get error stating Check your proxy settings or contact your network administrator.

在常规代理上使用相同的代码示例时,它工作得很好.当我通过 9050 端口发出粗体请求时——它工作得很好.

While using same code sample on regular proxy – it works just fine. When I do bold request through 9050 port – it works just fine.

最后,当我使用与 webdriver.Chrome 而不是 webdriver.Remote 相同的代码示例时,它工作正常!

And finally, when I use same code sample with webdriver.Chrome instead of webdriver.Remote it works fine!

如果有任何建议可以使事情通过远程 webdriver 工作,我将不胜感激.

I would appreciate any suggestions to make things work through Remote webdriver.

更新:我正在使用 selenium==3.14.0 并且 RemoteDriver 正在获取 docker 图像 selenium/node-chrome-debug:3.141.59-radium.

UPDATE: I'm using selenium==3.14.0 and RemoteDriver is getting docker image selenium/node-chrome-debug:3.141.59-radium.


解决方案

对于 macOS 和 Windows,您可以使用 host.docker.internal 从容器访问本地主机:

For macOS and Windows you can use host.docker.internal to access local host from container:

from selenium import webdriver

opts = webdriver.ChromeOptions()
opts.add_argument("--no-sandbox")
opts.add_argument("--disable-dev-shm-usage")
opts.add_argument("--proxy-server=socks5://host.docker.internal:9050")
driver = webdriver.Remote(command_executor="http://127.0.0.1:4444/wd/hub", desired_capabilities=opts.to_capabilities())

driver.get("http://jsonip.com/")
print(driver.find_element_by_css_selector("html").text)
driver.quit()

这里是 selenium hub 如何与 tor 代理.您可以在 docker 中创建一个网络,将容器附加到它,然后使用容器名称作为代理主机:

Here is how selenium hub works with tor proxy. You can create a network in docker, attach containers to it and then use container name as a proxy host:

docker network create mynetwork
docker run -it -p 8118:8118 -p 9050:9050 --name tor-proxy -d dperson/torproxy
docker run -d -p 4444:4444 --name selenium-hub -v /dev/shm:/dev/shm selenium/standalone-chrome:3.141.59-yttrium
docker network connect mynetwork hub
docker network connect mynetwork tor-proxy

这是相同的示例,但使用了 docker-compose:

Here is the same example but with docker-compose:

version: '3.5'

services:

  tor-proxy:
    image: dperson/torproxy
    container_name: tor-proxy
    ports:
      - "8118:8118"
      - "9050:9050"
    networks:
      - mynetwork
  
  selenium-hub:
    image: selenium/standalone-chrome:3.141.59-yttrium
    container_name: selenium-hub
    ports:
      - "4444:4444"
    networks:
      - mynetwork

networks:
  mynetwork:
    name: mynetwork
    driver: bridge

Python 代码:

from selenium import webdriver

opts = webdriver.ChromeOptions()
opts.add_argument("--no-sandbox")
opts.add_argument("--disable-dev-shm-usage")
opts.add_argument("--proxy-server=socks5://tor-proxy:9050")
driver = webdriver.Remote(command_executor="http://127.0.0.1:4444/wd/hub", desired_capabilities=opts.to_capabilities())

driver.get("http://jsonip.com/")
print(driver.find_element_by_css_selector("html").text)
driver.quit()

结果:

{ip":18.27.197.252",关于":https://jsonip.com/about",Pro!":http://getjsonip.com","获取通知":https://jsonip.com/notify"}

{"ip":"18.27.197.252","about":"https://jsonip.com/about","Pro!":"http://getjsonip.com","Get Notifications": "https://jsonip.com/notify"}

进程以退出代码 0 结束

Process finished with exit code 0

使用更改的 IP 再次运行:

Run again with changed IP:

{ip":178.165.72.177",about":https://jsonip.com/about",Pro!":http://getjsonip.com","获取通知":https://jsonip.com/notify"}

{"ip":"178.165.72.177","about":"https://jsonip.com/about","Pro!":"http://getjsonip.com","Get Notifications": "https://jsonip.com/notify"}

进程以退出代码 0 结束

Process finished with exit code 0

相关文章