Generating a webpage screenshot in ruby

As I have spent the last hour and a half figuring out, generating a screenshot of a website in Ruby is harder than it seems.

So first off, we’ll need a browser’s rendering engine. Preferably one that is headless, so we’ll go with PhantomJS. Make sure it’s installed.

Then, we need a way to control PhantomJS, and I’m a fan of Watir so we’ll use that. To connect to PhantomJS, you also need the selenium webdriver gem. So your gemfile would look like:

gem 'watir'
gem 'watir-webdriver'

Then let’s write the code to actually take the screenshot:

browser = Watir::Browser.new :phantomjs
browser.goto("https://google.com")
browser.screenshot.save 'google.com.png'

However, if you try to run it, the result isn’t what you’d expect from Google. It has the old black navigation bar at the top. As it turns out, Google does UA sniffing, so let’s pretend to be Firefox instead:

capabilities = Selenium::WebDriver::Remote::Capabilities.phantomjs('phantomjs.page.settings.userAgent' => 'Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Firefox/33.0')

browser = Watir::Browser.new :phantomjs, :desired_capabilities => capabilities

So that should work for everything and every body now, right? Nope!

As it also turns out, PhantomJS includes it’s own certificate store, and if it doesn’t recognize a site’s SSL certificate, it’ll error out, returning a transparent image as the screenshot. (Cause hey, about:blank is actually blank!) You can set the command line parameter --ignore-ssl-errors=true to fix that. You can pass CLI parameters in via a capability. So modifying the previous code…

capabilities = Selenium::WebDriver::Remote::Capabilities.phantomjs('phantomjs.page.settings.userAgent' => 'Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Firefox/33.0',
                                                                   'phantomjs.cli.args' => ['--ignore-ssl-errors=true'])

Now personally, I think the default resolution (400x300) of PhantomJS is a little small. I personally like the screenshot to be a bit bigger, so let’s expand it:

dimensions = Selenium::WebDriver::Dimension.new(1024, 768)
browser.driver.manage.window.size = dimensions

There you go! A side note, the dimensions just set the size of the viewport, the website can of course scroll beyond that size. Taking a screenshot will capture the entire page, not just what is visible in the viewport.

The complete code:

require 'watir'
require 'watir-webdriver'

capabilities = Selenium::WebDriver::Remote::Capabilities.phantomjs('phantomjs.page.settings.userAgent' => 'Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Firefox/33.0',
                                                                   'phantomjs.cli.args' => ['--ignore-ssl-errors=true'])

browser = Watir::Browser.new :phantomjs, :desired_capabilities => capabilities

dimensions = Selenium::WebDriver::Dimension.new(1024, 768)
browser.driver.manage.window.size = dimensions

browser.goto("https://google.com")

browser.screenshot.save 'google.com.png'

Thanks to these other blogs:

  • http://blog.vbalazs.me/2014/08/my-adventures-with-tracking-down-a-database-leak.html
  • http://stackoverflow.com/a/17687452
  • http://yizeng.me/2013/08/10/set-user-agent-using-selenium-webdriver-c-and-ruby/#phantomjs
|