Browser Instances

Creating browser instances

To create a WebDriver browser instance:

from webdriverplus import WebDriver
browser = WebDriver('firefox')

Currently supported browsers are firefox, chrome, ie, htmlunit, remote and phantomjs.

The default browser is firefox. If called without any arguments, WebDriver() will create a firefox instance.

browser = WebDriver()


Because starting up a web browser instance on every test run can be a significant performance hit, WebDriverPlus provides an easy way to allows instances to be reused between test sessions.

Setting the reuse_browser flag ensures that when you call driver.quit() the browser will be returned to a browser pool, and reused when you create a new WebBrowserInstance:

browser = WebDriver('firefox', reuse_browser=True)

There are some important aspects to bear in mind about this behaviour:

  • WebDriver Plus currently has no way of clearing browser history or cache. Be aware that this may affect the behaviour of tests.
  • On quitting the browser and returning it to the pool, WebDriver Plus will clear any cookies for the browser for the current domain. It has no way of clearing all cookies for all domains. If you have test cases that access URLs from multiple domains, consider if you need to explicitly clear any cookies between sessions.
  • WebDriver Plus will only retain one instance of each browser type (Eg Firefox, Chrome etc...) in the pool. Instances will only be reused if the arguments to the WebDriver() constructor have not changed since the previous instance was created.


By default WebDriverPlus will ensure that when a python process quits it will attempt to quit any remaining open WebDriver instances.

If you do not want this behaviour set quit_on_exit to False:

browser = WebDriver('firefox', quit_on_exit=False)


By default WebDriverPlus will not wait for elements to become available. You can pass a wait argument to specify the number of seconds that find() should wait for elements to become available before timing out.

browser = WebDriver('firefox', wait=10)

This uses WebDriverWait under the covers, but is much less verbose. The idea behind setting a per-browser wait argument instead of forcing the programmer to use WebDriverWait around each piece of code that needs to wait for an element is to free the programmer from having to think about waiting, which we consider a low-level detail that the framework should deal with.


By default WebDriverPlus highlights elements it found. Setting highlight to falsy will disable this

Switching to an iframe

To switch to a specific iframe, call switch_to_frame()

Waiting for a specific condition

Right now webdriverplus supports expected condition style waiting with wait_for(selector, displayed=True. By default it will wait until the element with selector to be present AND visible. If displayed is set to False, it will only wait until element is present.


Quitting browser instances

To quit a WebDriver browser instance, call quit():



Setting the force flag causes the browser instance to quit and ignore the value of the reuse_browser flag. The instance will be terminated and will not be returned to the browser pool:


Supported browsers

  • Firefox - Should run out-of-the-box.
  • Chrome - Install the chrome driver first.
  • IE - Install the IE driver first.
  • HTMLUnit (headless browser) - should auto-install and run out-of-the-box.
  • PhantomJS - Install PhantomJS first.

Headless mode using Xvfb or Xvnc

Using pyvirtualdisplay, you can run real browser instances in a virtual X framebuffer or VNC session. This enables you to run Firefox or Chrome tests in headless mode, without having to install HTMLUnit.

$ pip install pyvirtualdisplay

You need to install either Xvfb or Xvnc as a backend for pyvirtualdisplay.

To run the headless tests, use the --headless argument:

$ python --headless
Running tests in headless mode.
Ran 57 tests in 7.715s