10

I'm learning to use Selenium for basic things like taking a screenshot, scraping, and testing and would like to use it with headless Chrome, which is now stable as of Chrome 59.

I have been able to take a screenshot using the 'selenium-webdriver' gem and chromedriver, but not headless.

Here is the ruby script that I am running which hangs after starting to initialize the driver

require 'rubygems' require 'selenium-webdriver' Selenium::WebDriver.logger.level = :debug p 'initializing driver' driver = Selenium::WebDriver.for :chrome, switches: %w[--headless --disable-gpu --screenshot --hide-scrollbars] p 'navigating to Google' driver.navigate.to "http://google.com" driver.save_screenshot("./screen.png") driver.quit

and the output from the logs:

:> ruby rubytest.rb "initializing driver" 2017-06-07 15:55:43 DEBUG Selenium Executing Process ["/Users/name/Documents/scrapings/python/env/bin/chromedriver", "--port=9515"] 2017-06-07 15:55:43 DEBUG Selenium polling for socket on ["127.0.0.1", 9515] Starting ChromeDriver 2.29.461585 (0be2cd95f834e9ee7c46bcc7cf405b483f5ae83b) on port 9515 Only local connections are allowed. 2017-06-07 15:55:43 INFO Selenium -> POST session 2017-06-07 15:55:43 INFO Selenium >>> http://127.0.0.1:9515/session | {"desiredCapabilities":{"browserName":"chrome","version":"","platform":"ANY","javascriptEnabled":true,"cssSelectorsEnabled":true,"takesScreenshot":false,"nativeEvents":false,"rotatable":false,"chromeOptions":{"args":["--headless","--disable-gpu","--screenshot","--hide-scrollbars"]}}} 2017-06-07 15:55:43 DEBUG Selenium > {"Accept"=>"application/json", "Content-Type"=>"application/json; charset=utf-8", "Content-Length"=>"284"} [RUBY BACKTRACE TO DRIVER INITIALIZATION]

I have tried using the JavaScript and Python drivers with similar code and nothing works. When I try this with Python, the error message is 

WebDriverException: Message: unknown error: Chrome failed to start: exited abnormally (Driver info: chromedriver=2.29.461585 (0be2cd95f834e9ee7c46bcc7cf405b483f5ae83b),platform=Mac OS X 10.12.5 x86_64)
Improve this question
4
  • 1
    Is this of use? medium.com/@CukeHub/… Commented Jun 8, 2017 at 0:37
  • Yes, it was of some use, thank you. Passing the --headless argument through desired_capabilities allowed me to navigate and select elements headlessly; however, this method still does not give a screenshot like normal 'non-headless' Chrome. Commented Jun 8, 2017 at 2:11
  • I just found a nice little link that deals with the same/similar problem. Commented Jun 8, 2017 at 11:15
  • I'm having the same issue with Ruby. Browser launches in headless mode OK, but resulting screenshot is empty. This works fine for me in Python and JavaScript. Commented Jun 8, 2017 at 11:22

3 Answers 3

Reset to default
13

I found this blog post useful for setting up headless chrome with selenium in ruby

require "selenium-webdriver" # configure the driver to run in headless mode options = Selenium::WebDriver::Chrome::Options.new options.add_argument('--headless') driver = Selenium::WebDriver.for :chrome, options: options driver.navigate.to "https://www.google.com" # resize the window and take a screenshot driver.manage.window.resize_to(800, 800) driver.save_screenshot "screenshot.png"
Improve this answer
Sign up to request clarification or add additional context in comments.

Comments

2

Managed to work through this in the end via various docs, blog posts and gists.

caps = Selenium::WebDriver::Remote::Capabilities.chrome("desiredCapabilities" => {"takesScreenshot" => true}, "chromeOptions" => {"binary" => "/Applications/Google Chrome Canary.app/Contents/MacOS/Google Chrome Canary"}) driver = Selenium::WebDriver.for :chrome, desired_capabilities: caps, switches: %w[--headless --no-sandbox --disable-gpu --remote-debugin-port=9222 --screen-size=1200x800]

You need to use a very recent version of Chrome (I'm using Canary) and tell Selenium the path to the binary. You also need to set the desired capabilities for 'takesScreenshot' to true.

Improve this answer

4 Comments

Note for anyone running things inside a Docker container, you'll need to replace element.send_keys() with this util function: def send_keys(d, el, str) d.execute_script('arguments[0].setAttribute("value", arguments[1])', el, str) end Use it like so: send_keys(driver, searchElement, "WebDriver")
I'm on Chrome 59 which is the most recent version, andI'm still getting a blank screenshot. Also, is there a reason that you have desiredCapabilities nested twice?
Good question, and in fact you're right we can clean this up like so: caps = Selenium::WebDriver::Remote::Capabilities.chrome("desiredCapabilities" => {"takesScreenshot" => true}, "chromeOptions" => {"binary" => "/Applications/Google Chrome Canary.app/Contents/MacOS/Google Chrome Canary", "args" => ["--headless", "--no-sandbox", "--disable-gpu", "--remote-debugin-port=9222", "--screen-size=1200x800"]}) driver = Selenium::WebDriver.for :chrome, desired_capabilities: caps
Not sure why screenshots don't work with v59, but you can download Canary (v61) here: google.com/chrome/browser/canary.html Hopefully whatever fixes are in v61 will filter into the stable release sooner rather than later.
2

I've written a blog post about how to do it. Summing up:

1) Make sure you have Chrome version 57+ on Linux, 59+ on macOS or 60+ on Windows;

2) Add/update the gem selenium-webdriver;

3) Make sure you’re using ChromeDriver version 2.30 or higher;

4) Add the following driver to your spec_helper.rb or rails_helper.rb:

Capybara.register_driver :chrome do |app| Capybara::Selenium::Driver.new app, browser: :chrome, options: Selenium::WebDriver::Chrome::Options.new(args: %w[headless disable-gpu]) end Capybara.javascript_driver = :chrome
Improve this answer

Comments

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.