splinter: how to add chrome options? - selenium

Question

I'm using splinter(v0.7.3) for web testing under linux, while on chrome, the default sample code can not running:
from splinter import Browser
from pyvirtualdisplay import Display
d = Display(visible=0, size=(800, 600))
d.start()
b = Browser('chrome')
b.visit('http://www.google.com')
b.quit()
d.stop()
While running, I got the exception like this:
selenium.common.exceptions.WebDriverException: Message: chrome not reachable
And I test the same function in selenium with some chrome option added:
from selenium import web driver
from selenium.webdriver.chrome.options import Options
from pyvirtualdisplay import Display
d = Display(visible=0, size=(800, 600))
d.start()
opt = Options()
opt.add_argument('--disable-setuid-sandbox')
b = webdriver.Chrome(chrome_options=opt)
b.get('http://www.google.com')
b.quit()
d.stop()
This works ok, the difference is the --disable-setuid-sandbox option added to chrome driver, if the option not add, there would be a zombie chrome-sandbox process under chromium-browser.
The problem here is, I don't know how to pass a chrome.options.Option instance to splinter.Browser(), I browsed the implementation under splinter/driver/webdriver/chrome.py, it seems that there is no entry to pass such a instance to splinter.Browser(). Is there some other way to pass options to chrome driver?

Answer 1

Create a new instance of BaseWebDriver and set .driver with an instance of the Chrome driver. This example starts Chrome maximized:
from selenium.webdriver import Chrome
from selenium.webdriver.chrome.options import Options
from splinter.driver.webdriver import BaseWebDriver, WebDriverElement
options = Options()
options.add_argument('--start-maximized')
browser = BaseWebDriver()
browser.driver = Chrome(chrome_options=options)
browser.visit('https://www.google.com')

Answer 2

The only way I could ever do this was by using the add_argument method with selenium.webdriver.ChromeOptions like so:
from selenium.webdriver import ChromeOptions
from splinter import Browser
chrome_options = ChromeOptions()
chrome_options.add_argument(your_argument)
b=Browser("chrome", options=chrome_options)
b.visit('http://www.google.com')
b.quit()
so in your code would be:
from splinter import Browser
from selenium.webdriver import ChromeOptions
from pyvirtualdisplay import Display #I'm not certain what this is...
d = Display(visible=0, size=(800, 600))
d.start()
chrome_options = ChromeOptions()
chrome_options.add_argument('disable-setuid-sandbox')
b = Browser('chrome')
b.visit('http://www.google.com')
b.quit()
d.stop()
Note: I was unable to test this with your argument specifically because I recently broke my GRUB so I am stuck in windows, and the disable-setuid-sandbox option is linux-only. However, I have been using this method with the headless argument for a while.

Answer 3

I am not 100% sure that this will work but I just looked at the docs for splinter and it says.
You can also pass additional arguments that correspond to Selenium DesiredCapabilities arguments.
Looking into the sourcecode of Splinter calling Browser can take some arguments. These arguments will then be passed to create an Instance of the Chrome WebDriver. So I went to the selenium sourcecode and saw the constructor looks like this:
def __init__(self, executable_path="chromedriver", port=0,
chrome_options=None, service_args=None,
desired_capabilities=None, service_log_path=None):
There is a parameter for chrome_options so it should be possible to pass it using this parameter. So if I'm correct this should work fine:
opt = Options()
opt.add_argument('--disable-setuid-sandbox')
b = Browser(browser='chrome', chrome_options=opt)
Edit
Alternatively you could pass the options as desired capabilities aswell:
opt = Options()
opt.add_argument('--disable-setuid-sandbox')
dc = opt.to_capabilities()
b = Browser(browser='chrome', desired_capabilities=dc)

Answer 4

I've been working on a fork of splinter for the past couple weeks, you can check out my dev branch if you want. I have added this and other features.
Options can be passed as a list to the chrome_options parameter
from splinter import Browser
options = ['--start-maximized', '--disable-setuid-sandbox']
with Browser('chrome', chrome_options=options) as browser:
browser.visit('http://www.google.com')
Edit:
So it turns out this was possible with splinter all along just using **kwargs which passes the various available options to the selenium driver(s). For example:
from splinter import Browser
options = {'chrome_options':['--start-maximized', '--disable-setuid-sandbox']}
with Browser('chrome', **options) as browser:
browser.visit('http://www.google.com')