如何在Shadow Root中点击按钮？Selenium报错解决方案

问题场景 #

尝试用Selenium点击网页Shadow Root中的同意广告按钮，代码如下：

from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from selenium.webdriver.chrome.service import Service
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.common.by import By
import time


print(f"Checking Browser driver...")
options = Options()
# options.add_argument('--headless=new')  
options.add_argument("start-maximized")
options.add_argument('--log-level=3')  
options.add_experimental_option("prefs", {"profile.default_content_setting_values.notifications": 1})    
options.add_experimental_option("excludeSwitches", ["enable-automation"])
options.add_experimental_option('excludeSwitches', ['enable-logging'])
options.add_experimental_option('useAutomationExtension', False)
options.add_argument('--disable-blink-features=AutomationControlled') 
srv=Service()
driver = webdriver.Chrome (service=srv, options=options)    
waitWD = WebDriverWait (driver, 10)         

link = "https://www.firmy.cz/kraj-liberecky/liberec/1818-liberec?q="
driver.get (link)     
time.sleep(5)
shadowHost = driver.find_element(By.XPATH,'//div[@class="szn-cmp-dialog-container"]')
shadowRoot = shadowHost.shadow_root
time.sleep(5)
shadowRoot.find_element(By.XPATH, '//button[@data-testid="cw-button-agree-with-ads"]').click()      
input("Press!")

运行后抛出错误：

(selenium) C:\DEVNEU\Fiverr2025\TRY\readingmadness>python test.py
Checking Browser driver...
Traceback (most recent call last):
  File "C:\DEVNEU\Fiverr2025\TRY\readingmadness\test.py", line 29, in <module>
    shadowRoot.find_element(By.XPATH, '//button[@data-testid="cw-button-agree-with-ads"]').click()
    ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\DEVNEU\.venv\selenium\Lib\site-packages\selenium\webdriver\remote\shadowroot.py", line 53, in find_element
    return self._execute(Command.FIND_ELEMENT_FROM_SHADOW_ROOT, {"using": by, "value": value})["value"]
           ~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\DEVNEU\.venv\selenium\Lib\site-packages\selenium\webdriver\remote\shadowroot.py", line 82, in _execute
    return self.session.execute(command, params)
           ~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
  File "C:\DEVNEU\.venv\selenium\Lib\site-packages\selenium\webdriver\remote\webdriver.py", line 384, in execute
    self.error_handler.check_response(response)
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^
  File "C:\DEVNEU\.venv\selenium\Lib\site-packages\selenium\webdriver\remote\errorhandler.py", line 232, in check_response
    raise exception_class(message, screen, stacktrace)
selenium.common.exceptions.InvalidArgumentException: Message: invalid argument: invalid locator
  (Session info: chrome=133.0.6943.60)
Stacktrace:
        GetHandleVerifier [0x00007FF64EE46F15+28773]
        (No symbol) [0x00007FF64EDB2600]
        (No symbol) [0x00007FF64EC48FAA]
        (No symbol) [0x00007FF64EC9F05A]
        (No symbol) [0x00007FF64EC9F4FC]
        (No symbol) [0x00007FF64EC91CAC]
        (No symbol) [0x00007FF64ECC728F]
        (No symbol) [0x00007FF64EC91B36]
        (No symbol) [0x00007FF64ECC7460]
        (No symbol) [0x00007FF64ECEF6F3]
        (No symbol) [0x00007FF64ECC7023]
        (No symbol) [0x00007FF64EC8FF5E]
        (No symbol) [0x00007FF64EC911E3]
        GetHandleVerifier [0x00007FF64F19425D+3490733]
        GetHandleVerifier [0x00007FF64F1ABA43+3586963]
        GetHandleVerifier [0x00007FF64F1A147D+3544525]
        GetHandleVerifier [0x00007FF64EF0C9DA+838442]
        (No symbol) [0x00007FF64EDBD04F]
        (No symbol) [0x00007FF64EDB9614]
        (No symbol) [0x00007FF64EDB97B6]
        (No symbol) [0x00007FF64EDA8CE9]
        BaseThreadInitThunk [0x00007FFC4B4D259D+29]
    RtlUserThreadStart [0x00007FFC4C7AAF38+40]

错误原因 #

Selenium的ShadowRoot对象对XPath定位的支持存在兼容性限制，直接使用XPath会触发InvalidArgumentException。此外，硬编码的time.sleep可靠性差，容易因页面加载速度波动导致元素未找到。

修正方案 #

方案1：使用CSS选择器定位（推荐） #

将Shadow Root内的元素定位方式改为CSS选择器，同时用显式等待替代time.sleep，确保元素加载完成后再操作：

from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from selenium.webdriver.chrome.service import Service
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.by import By


print(f"Checking Browser driver...")
options = Options()
# options.add_argument('--headless=new')  
options.add_argument("start-maximized")
options.add_argument('--log-level=3')  
options.add_experimental_option("prefs", {"profile.default_content_setting_values.notifications": 1})    
options.add_experimental_option("excludeSwitches", ["enable-automation"])
options.add_experimental_option('excludeSwitches', ['enable-logging'])
options.add_experimental_option('useAutomationExtension', False)
options.add_argument('--disable-blink-features=AutomationControlled') 
srv=Service()
driver = webdriver.Chrome(service=srv, options=options)    
waitWD = WebDriverWait(driver, 10)         

link = "https://www.firmy.cz/kraj-liberecky/liberec/1818-liberec?q="
driver.get(link)     

# 显式等待Shadow Host加载完成
shadowHost = waitWD.until(EC.presence_of_element_located((By.CSS_SELECTOR, 'div.szn-cmp-dialog-container')))
shadowRoot = shadowHost.shadow_root

# 显式等待Shadow Root内的按钮加载完成，并用CSS选择器定位
agree_button = waitWD.until(lambda d: shadowRoot.find_element(By.CSS_SELECTOR, 'button[data-testid="cw-button-agree-with-ads"]'))
agree_button.click()      

input("Press!")
driver.quit()

方案2：通过JavaScript执行点击 #

如果CSS选择器仍存在问题，可以直接用JavaScript穿透Shadow Root定位元素并点击，这种方式不受Selenium Shadow Root API限制：

from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from selenium.webdriver.chrome.service import Service
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.by import By


print(f"Checking Browser driver...")
options = Options()
# options.add_argument('--headless=new')  
options.add_argument("start-maximized")
options.add_argument('--log-level=3')  
options.add_experimental_option("prefs", {"profile.default_content_setting_values.notifications": 1})    
options.add_experimental_option("excludeSwitches", ["enable-automation"])
options.add_experimental_option('excludeSwitches', ['enable-logging'])
options.add_experimental_option('useAutomationExtension', False)
options.add_argument('--disable-blink-features=AutomationControlled') 
srv=Service()
driver = webdriver.Chrome(service=srv, options=options)    
waitWD = WebDriverWait(driver, 10)         

link = "https://www.firmy.cz/kraj-liberecky/liberec/1818-liberec?q="
driver.get(link)     

# 等待Shadow Host出现后，用JavaScript找到按钮并点击
waitWD.until(EC.presence_of_element_located((By.CSS_SELECTOR, 'div.szn-cmp-dialog-container')))
driver.execute_script("""
    const shadowHost = document.querySelector('div.szn-cmp-dialog-container');
    const shadowRoot = shadowHost.shadowRoot;
    const agreeBtn = shadowRoot.querySelector('button[data-testid="cw-button-agree-with-ads"]');
    if (agreeBtn) agreeBtn.click();
""")

input("Press!")
driver.quit()

关键说明 #

• 替换time.sleep为显式等待：避免因页面加载延迟导致的元素未找到问题，提升代码稳定性。
• 优先使用CSS选择器：Selenium对Shadow Root内的CSS选择器支持更稳定，XPath容易触发定位错误。
• JavaScript执行兜底：当Shadow DOM结构复杂时，直接用JS操作是最可靠的方式，无需依赖Selenium的Shadow Root API。

内容的提问来源于stack exchange，提问作者Rapid1898