为什么我的单元测试可以在 Chrome 驱动程序上工作，但不能在 PhantomJS 上工作？

Question

我正在Selenium Pychon Bindings document 下编写我的第一个单元测试。

我使用 Chrome 驱动程序编写测试如下：

from selenium import webdriver
from selenium.webdriver.common.keys import Keys
import unittest

class PythonOrgSearch(unittest.TestCase):

    def setUp(self):
        self.driver = webdriver.Chrome(r"C:\chromedriver_win32\chromedriver")

    def test_search_in_python_org(self):
        driver = self.driver
        driver.get(r"http://www.python.org")
        self.assertIn("Python", driver.title)
        el = driver.find_element_by_name("q")
        el.send_keys("pycon")
        el = driver.find_element_by_id("submit")
        el.click()
        assert "No results found." not in driver.page_source

    def tearDown(self):
        self.driver.close()

if __name__ == "__main__":
    unittest.main()

在终端上运行它就可以了。终端返回：

.
----------------------------------------------------------------------
Ran 1 test in 7.873s

OK

但是如果我替换

self.driver = webdriver.Chrome(r"C:\chromedriver_win32\chromedriver")

与

self.driver = webdriver.PhantomJS(r"C:\phantomjs-2.1.1\bin\phantomjs")

代码如下

from selenium import webdriver
from selenium.webdriver.common.keys import Keys
import unittest


class PythonOrgSearch(unittest.TestCase):

    def setUp(self):
        self.driver = webdriver.PhantomJS(r"C:\phantomjs-2.1.1\bin\phantomjs")


    def test_search_in_python_org(self):
        driver = self.driver
        driver.get(r"http://www.python.org")
        self.assertIn("Python", driver.title)
        el = driver.find_element_by_name("q")
        el.send_keys("pycon")
        el = driver.find_element_by_id("submit")
        el.click()
        assert "No results found." not in driver.page_source

    def tearDown(self):
        self.driver.close()


if __name__ == "__main__":
    unittest.main()

我的终端给了我一个错误：

E
======================================================================
ERROR: test_search_in_python_org (__main__.PythonOrgSearch)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "test_python_org_search.py", line 24, in test_search_in_python_org
    el.click()
  File "C:\Program Files (x86)
    self._execute(Command.CLICK_ELEMENT)
  File "C:\Program Files (x86)
    return self._parent.execute(command, params)
  File "C:\Program Files (x86)
    self.error_handler.check_response(response)
  File "C:\Program Files (x86)    
    raise exception_class(message, screen, stacktrace)
selenium.common.exceptions.ElementNotVisibleException: Message: {"errorMessage":"Element is not   
"Content-Length":"81","Content-Type":"application/json;charset=UTF-8","Host":"127.0.0.1:50294",   
67e-433b649e227d\"}","url":"/click","urlParsed":{"anchor":"","query":"","file":"click",    
elative":"/click","port":"","host":"","password":"","user":"","userInfo":"","authority":"",    
","queryKey":{},"chunks":["click"]},    
46/click"}}
Screenshot: available via screen

----------------------------------------------------------------------
Ran 1 test in 6.214s

FAILED (errors=1)

有人能告诉我为什么我不能在 PhantomJS 中使用 click 吗？

Answer 1

使用 PhantomJS 时，屏幕截图是最有用的调试工具之一。我修改了您的脚本以在失败时记录屏幕截图：

from selenium import webdriver
from selenium.webdriver.common.keys import Keys
import unittest


class PythonOrgSearch(unittest.TestCase):

    def setUp(self):
        self.driver = webdriver.PhantomJS()


    def test_search_in_python_org(self):
        driver = self.driver
        driver.get(r"http://www.python.org")
        self.assertIn("Python", driver.title)
        el = driver.find_element_by_name("q")
        el.send_keys("pycon")
        el = driver.find_element_by_id("submit")
        try:
            el.click()
        except:
            screenshot = driver.get_screenshot_as_png()
            with open('screenshot.png', 'wb') as w:
                w.write(screenshot)
        assert "No results found." not in driver.page_source

    def tearDown(self):
        self.driver.close()


if __name__ == "__main__":
    unittest.main()

当我运行它时，我得到this screenshot。

您会注意到该页面的呈现方式就像在移动浏览器上一样。使用 phantomjs 时应设置浏览器窗口大小：

def setUp(self):
    self.driver = webdriver.PhantomJS()
    self.driver.set_window_size(800, 1000)

一旦你这样做了，然后比较 chrome 和 phantom 渲染，你会发现 phantom 没有正确加载到提交按钮中：

一种解决方法是发送 ENTER 键，而不是单击按钮：

def test_search_in_python_org(self):
    driver = self.driver
    driver.set_window_size(800, 1000)
    driver.get(r"http://www.python.org")
    self.assertIn("Python", driver.title)
    el = driver.find_element_by_name("q")
    el.send_keys("pycon")
    el.send_keys(Keys.ENTER)
    screenshot = driver.get_screenshot_as_png()
    with open('screenshot2.png', 'wb') as w:
        w.write(screenshot)

使用上面的代码，我现在看到了结果：

值得注意的是，这就是为什么包括我自己在内的许多人不使用 PhantomJS 进行测试的原因。我发现使用 Chrome 和/或 Firefox 会好得多，要么在它们自己的无头模式下运行它们，要么使用虚拟帧缓冲区（如 Xvfb）。

我个人在基于云的 linux 实例上进行测试，我使用 pyvirtualdisplay 和 Xvfb 来管理我的虚拟显示器。安装两者后，pyvirtualdisplay 将完全管理您的虚拟显示会话。要将它与您的上述脚本一起使用，我将执行以下操作：

from pyvirtualdisplay import Display

class PythonOrgSearch(unittest.TestCase):

    def setUp(self):
        self.display = Display(visible=False, size=(1200, 1500))
        self.display.start()
        self.driver = webdriver.Chrome()


    def test_search_in_python_org(self):
        # Do test

    def tearDown(self):
        self.driver.close()
        self.display.stop()

综上所述，我没有使用 Windows 的经验，但我相信该操作系统存在类似的配置