require 'rubygems' require 'activesupport' require 'taza/settings' module Taza # An abstraction of a website, but more really a container for a sites pages. # # You can generate a site by performing the following command: # $ ./script/generate site google # # This will generate a site file for google, a flows folder, and a pages folder in lib # # Example: # # require 'taza' # # class Google < Taza::Site # # end class Site # These methods are available for user contexts, including Site itself module Methods attr_accessor :browser, :site # Return an instance of the specified page. The name # Given should be the human-form of the page, without the # "page" suffix. # If a block is given, it yields to the page. def page(page_name, &block) method_name = page_name.computerize + '_page' send method_name, &block end # Send the browser to a url, relative to the site origin. def goto relative_url destination = File.join(@site.origin, relative_url) @browser.goto destination end def pages @pages || @site.pages end end @@before_browser_closes = Proc.new() {} # Use this to do something with the browser before it closes, but note that it is a class method which # means that this will get called for any instance of a site. # # Here's an example of how you might use it to print the DOM output of a browser before it closes: # # Taza::Site.before_browser_closes do |browser| # puts browser.html # end def self.before_browser_closes(&block) @@before_browser_closes = block end # Site attr_accessor :methods_module # A site can be called a few different ways # # The following example creates a new browser object and closes it: # Google.new do # google.search.set "taza" # google.submit.click # end # # This example will create a browser object but not close it: # Google.new.search.set "taza" # # Sites can take a couple of parameters in the constructor: # :browser => a browser object to act on instead of creating one automatically # (not sure if this is a useful feature or not) def initialize(params={}, &block) @site = self @module_name = self.class.parent.to_s @class_name = self.class.to_s.split("::").last define_flows page_loader = PageLoader.new(@module_name, pages_path) @pages = page_loader.page_names @methods_module = page_loader.page_methods @methods_module.send(:include, Methods) self.extend(@methods_module) @browser = params[:browser] initialize_browser execute_block_and_close_browser(&block) if block_given? end # Initializes browser based on configuration settings. def initialize_browser if @browser @leave_open = true else @browser = Browser.create(config) end @browser.goto origin bring_to_front_if_appropriate @leave_open ||= config[:leave_open] end def config Settings.config(@class_name) end # The base url of the site. This is configured in environments.yml. def origin config[:url] end def bring_to_front_if_appropriate return unless config[:bring_to_front] return unless config[:browser] == 'ie' @browser.bring_to_front end def execute_block_and_close_browser begin yield self rescue => site_block_exception ensure begin close rescue => close_exception "" end exception = site_block_exception || close_exception raise exception if exception end end private :execute_block_and_close_browser def close begin @@before_browser_closes.call(@browser) rescue => before_browser_closes_block_exception "" # so basically rcov has a bug where it would insist this block is uncovered when empty end begin @browser.close unless @leave_open ensure raise before_browser_closes_block_exception if before_browser_closes_block_exception end end # Return a context that supports the "watircraft" commands. # Currently used by the script\console def execution_context initialize_context!(Object.new) end def initialize_context!(context) context.extend @methods_module context.site = @site context.browser = @site.browser context end private def pages_path # :nodoc: File.join(path,'pages','**','*.rb') # does this need to include partials? end def define_flows # :nodoc: Dir.glob(flows_path) do |file| require file end end # This is used to call a flow belonging to the site # # Example: # Google.new do |google| # google.flow(:perform_search, :query => "taza") # end # # Where the flow would be defined under lib/sites/google/flows/perform_search.rb and look like: # class PerformSearch < Taza::Flow # alias :google :site # # def run(params={}) # google.search.set params[:query] # google.submit.click # end # end # # methods that neither depend on or modify state # private def flows_path # :nodoc: File.join(path,'flows','*.rb') end def path # :nodoc: File.join(base_path,'lib') end def base_path # :nodoc: APP_ROOT end class PageLoader attr_reader :page_methods, :page_names def initialize site_module, pages_path @site_module = site_module @pages_path = pages_path @page_methods = Module.new @page_names = [] define_site_pages end private def define_site_pages # :nodoc: Dir.glob(@pages_path) do |file| require file page_name = File.basename(file,'.rb') @page_names << page_name page_class = "#{@site_module}::#{page_name.camelize}" @page_methods.module_eval <<-EOS def #{page_name} page = #{page_class}.new page.browser = @browser page.site = @site yield page if block_given? page end EOS end end end end end