JRuby script that logs you into Github and scrapes the "source" of your Wiki pages (i.e. what you typed) and saves it to a file.

scrape-github-wiki.rb

require 'rubygems'
require 'celerity'
require 'hpricot'
require 'htmlentities'

# You obviously need all of the above gems installed before proceeding

user, password, project = ARGV # 'tobi', 'my_password', 'liquid'
raise(ArgumentError, "jruby scrape-github-wiki <username> <password> <projectname>") unless user and project and password

# Constants
WIKI_PAGES_URL = "https://github.com/#{user}/#{project}/wiki/_pages"
BASE_URL = "https://github.com"

# Start browser
puts "Starting headless browser and scraping wiki pages..."
@b = Celerity::Browser.new
@b.goto "https://github.com/login"

# For decoding
@ent = HTMLEntities.new

begin
	@b.text_field(:name => 'login').set user
	@b.text_field(:name => 'password').set password
	@b.button(:value => "Log in").click

	@b.goto WIKI_PAGES_URL
	toc_links = @b.elements_by_xpath("//*[@id='guides']/div/div[contains(@class, 'wikistyle')]/ul/li/strong/a")
	wiki_links = []
	# We have to get the href's up front, because celerity
	# won't find the elements in its cache once we navigate away
	toc_links.each do |l|
		wiki_links << BASE_URL + l.href
	end
	
	wiki_text = ''
	wiki_links.each do |l|
		@b.goto l
		@b.link(:class => /btn-edit/).click
		
		wiki_page_title = @b.text_field(:id => "wiki_name").text
		puts "Scraping wiki page with title: #{wiki_page_title}"
		
		wiki_text << "Wiki Page Title: #{wiki_page_title}\n"
		wiki_text << @ent.decode(Hpricot(@b.html).at("#wiki_body").inner_html)
		wiki_text << "\n\n" + ("#" * 80) + "\n\n"
	end
rescue StandardError => e
	puts "An error occurred: " + e
ensure
	@b.close
end

puts "Saving wiki page source to 'github_wiki_pages.txt'..."
File.open('github_wiki_pages.txt', 'w') { |f| f.write(wiki_text)}
puts "\nDone\n"

You can also clone your wiki like a normal github repo, it's a lot simpler. :)

check out my fork

Absolutely true :) There I go forgetting that they're repo's. Though it represents a fine "hello world" of using the Watir/Celerity API and some super-simple Hpricot.