puppeteer-ruby 0.0.2

data/lib/puppeteer/launcher.rb

@@ -0,0 +1,26 @@
+require_relative './launcher/base'
+require_relative './launcher/browser_options'
+require_relative './launcher/chrome'
+require_relative './launcher/chrome_arg_options'
+require_relative './launcher/launch_options'
+
+# https://github.com/puppeteer/puppeteer/blob/master/lib/Launcher.js
+module Puppeteer::Launcher
+
+  # @param {string} projectRoot
+  # @param {string} preferredRevision
+  # @param {boolean} isPuppeteerCore
+  # @param {string=} product
+  # @return {!Puppeteer.ProductLauncher}
+  module_function def new(project_root:, preferred_revision:, is_puppeteer_core:, product:)
+    if product == 'firefox'
+      raise NotImplementedError.new("FirefoxLauncher is not implemented yet.")
+    end
+
+    Chrome.new(
+      project_root: project_root,
+      preferred_revision: preferred_revision,
+      is_puppeteer_core: is_puppeteer_core,
+    )
+  end
+end

data/lib/puppeteer/launcher/base.rb

@@ -0,0 +1,48 @@
+module Puppeteer::Launcher
+  class Base
+    # @param {string} projectRoot
+    # @param {string} preferredRevision
+    def initialize(project_root:, preferred_revision:, is_puppeteer_core:)
+      @project_root = project_root
+      @preferred_revision = preferred_revision
+      @is_puppeteer_core = is_puppeteer_core
+    end
+
+    class ExecutablePathNotFound < StandardError ; end
+
+    # @returns [String] Chrome Executable file path.
+    # @raise [ExecutablePathNotFound]
+    def resolve_executable_path
+      if !@is_puppeteer_core
+        # puppeteer-core doesn't take into account PUPPETEER_* env variables.
+        executable_path = ENV['PUPPETEER_EXECUTABLE_PATH']
+        if FileTest.exist?(executable_path)
+          return executable_path
+        end
+        raise ExecutablePathNotFound.new(
+          "Tried to use PUPPETEER_EXECUTABLE_PATH env variable to launch browser but did not find any executable at: #{executablePath}",
+        )
+      end
+
+      # temporal logic.
+      if RUBY_PLATFORM.include?("darwin") # MacOS
+        '/Applications/Google Chrome.app/Contents/MacOS/Google Chrome'
+      else
+        '/usr/bin/google-chrome'
+      end
+
+      # const browserFetcher = new BrowserFetcher(launcher._projectRoot);
+      # if (!launcher._isPuppeteerCore) {
+      #   const revision = process.env['PUPPETEER_CHROMIUM_REVISION'];
+      #   if (revision) {
+      #     const revisionInfo = browserFetcher.revisionInfo(revision);
+      #     const missingText = !revisionInfo.local ? 'Tried to use PUPPETEER_CHROMIUM_REVISION env variable to launch browser but did not find executable at: ' + revisionInfo.executablePath : null;
+      #     return {executablePath: revisionInfo.executablePath, missingText};
+      #   }
+      # }
+      # const revisionInfo = browserFetcher.revisionInfo(launcher._preferredRevision);
+      # const missingText = !revisionInfo.local ? `Browser is not downloaded. Run "npm install" or "yarn install"` : null;
+      # return {executablePath: revisionInfo.executablePath, missingText};
+    end
+  end
+end

data/lib/puppeteer/launcher/browser_options.rb

@@ -0,0 +1,41 @@
+# const {
+#   ignoreDefaultArgs = false,
+#   args = [],
+#   dumpio = false,
+#   executablePath = null,
+#   pipe = false,
+#   env = process.env,
+#   handleSIGINT = true,
+#   handleSIGTERM = true,
+#   handleSIGHUP = true,
+#   ignoreHTTPSErrors = false,
+#   defaultViewport = {width: 800, height: 600},
+#   slowMo = 0,
+#   timeout = 30000
+# } = options;
+# const {
+#   devtools = false,
+#   headless = !devtools,
+#   args = [],
+#   userDataDir = null
+# } = options;
+
+
+module Puppeteer::Launcher
+  class BrowserOptions
+    # @property {boolean=} ignoreHTTPSErrors
+    # @property {(?Puppeteer.Viewport)=} defaultViewport
+    # @property {number=} slowMo
+    def initialize(options)
+      @ignore_https_errors = options[:ignore_https_errors] || false
+      @default_viewport = options[:default_viewport] || Puppeteer::Viewport.new(width: 800, height: 600)
+      @slow_mo = options[:slow_mo] || 0
+    end
+
+    attr_reader :default_viewport, :slow_mo
+
+    def ignore_https_errors?
+      @ignore_https_errors
+    end
+  end
+end

data/lib/puppeteer/launcher/chrome.rb

@@ -0,0 +1,165 @@
+require 'tmpdir'
+
+# https://github.com/puppeteer/puppeteer/blob/master/lib/Launcher.js
+module Puppeteer::Launcher
+  class Chrome < Base
+    # @param {!(Launcher.LaunchOptions & Launcher.ChromeArgOptions & Launcher.BrowserOptions)=} options
+    # @return {!Promise<!Browser>}
+    def launch(options = {})
+      @chrome_arg_options = ChromeArgOptions.new(options)
+      @launch_options = LaunchOptions.new(options)
+      @browser_options = BrowserOptions.new(options)
+
+      chrome_arguments =
+        if !@launch_options.ignore_default_args
+          default_args.to_a
+        elsif @launch_options.ignore_default_args.is_a?(Enumerable)
+          default_args.reject do |arg|
+            @launch_options.ignore_default_args.include?(arg)
+          end.to_a
+        else
+          @chrome_arg_options.args.dup
+        end
+
+      #
+      # let temporaryUserDataDir = null;
+
+      if chrome_arguments.none?{ |arg| arg.start_with?('--remote-debugging-') }
+        if @launch_options.pipe?
+          chrome_arguments << '--remote-debugging-pipe'
+        else
+          chrome_arguments << '--remote-debugging-port=0'
+        end
+      end
+
+      temporary_user_data_dir = nil
+      if chrome_arguments.none?{ |arg| arg.start_with?('--user-data-dir') }
+        temporary_user_data_dir = Dir.mktmpdir('puppeteer_dev_profile-')
+        chrome_arguments << "--user-data-dir=#{temporary_user_data_dir}"
+      end
+
+      chrome_executable = @launch_options.executable_path || resolve_executable_path
+      use_pipe = chrome_arguments.include?('--remote-debugging-pipe')
+      runner = Puppeteer::BrowserRunner.new(chrome_executable, chrome_arguments, temporary_user_data_dir)
+      runner.start(
+        handle_SIGHUP: @launch_options.handle_SIGHUP?,
+        handle_SIGTERM: @launch_options.handle_SIGTERM?,
+        handle_SIGINT: @launch_options.handle_SIGINT?,
+        dumpio: @launch_options.dumpio?,
+        env: @launch_options.env,
+        pipe: use_pipe,
+      );
+
+      begin
+        connection = runner.setup_connection(
+          use_pipe: use_pipe,
+          timeout: @launch_options.timeout,
+          slow_mo: @browser_options.slow_mo,
+          preferred_revision: @preferred_revision,
+        )
+
+        browser = Puppeteer::Browser.create(
+          connection: connection,
+          context_ids: [],
+          ignore_https_errors: @browser_options.ignore_https_errors?,
+          default_viewport: @browser_options.default_viewport,
+          process: runner.proc,
+          close_callback: ->{ runner.close },
+        )
+
+        browser.wait_for_target(predicate: ->(target) { target.type == 'page' })
+
+        browser
+      rescue
+        runner.kill
+        raise
+      end
+    end
+
+    class DefaultArgs
+      include Enumerable
+
+      # @param options [Launcher::ChromeArgOptions]
+      def initialize(chrome_arg_options)
+        chrome_arguments = [
+          '--disable-background-networking',
+          '--enable-features=NetworkService,NetworkServiceInProcess',
+          '--disable-background-timer-throttling',
+          '--disable-backgrounding-occluded-windows',
+          '--disable-breakpad',
+          '--disable-client-side-phishing-detection',
+          '--disable-component-extensions-with-background-pages',
+          '--disable-default-apps',
+          '--disable-dev-shm-usage',
+          '--disable-extensions',
+          '--disable-features=TranslateUI',
+          '--disable-hang-monitor',
+          '--disable-ipc-flooding-protection',
+          '--disable-popup-blocking',
+          '--disable-prompt-on-repost',
+          '--disable-renderer-backgrounding',
+          '--disable-sync',
+          '--force-color-profile=srgb',
+          '--metrics-recording-only',
+          '--no-first-run',
+          '--enable-automation',
+          '--password-store=basic',
+          '--use-mock-keychain',
+        ]
+
+        if chrome_arg_options.user_data_dir
+          chrome_arguments << "--user-data-dir=#{chrome_arg_options.user_data_dir}"
+        end
+
+        if chrome_arg_options.devtools?
+          chrome_arguments << '--auto-open-devtools-for-tabs'
+        end
+
+        if (chrome_arg_options.headless?)
+          chrome_arguments.concat([
+            '--headless',
+            '--hide-scrollbars',
+            '--mute-audio',
+          ])
+        end
+
+        if chrome_arg_options.args.all?{ |arg| arg.start_with?('-') }
+          chrome_arguments << 'about:blank'
+        end
+
+        chrome_arguments.concat(chrome_arg_options.args)
+
+        @chrome_arguments = chrome_arguments
+      end
+
+      def each(&block)
+        @chrome_arguments.each do |opt|
+          block.call(opt)
+        end
+      end
+    end
+
+    def default_args(options = nil)
+      if options.nil?
+        @default_args ||= DefaultArgs.new(@chrome_arg_options)
+      else
+        DefaultArgs.new(ChromeArgOptions.new(options))
+      end
+    end
+
+    # @param {!(Launcher.BrowserOptions & {browserWSEndpoint?: string, browserURL?: string, transport?: !Puppeteer.ConnectionTransport})} options
+    # @return {!Promise<!Browser>}
+    def connect(options)
+      raise NotImplementedError.new('Puppeteer.connect is not implemented yet')
+    end
+
+    # @return {string}
+    def executable_path
+      resolve_executable_path
+    end
+
+    private def product
+      'chrome'
+    end
+  end
+end

data/lib/puppeteer/launcher/chrome_arg_options.rb

@@ -0,0 +1,49 @@
+# const {
+#   ignoreDefaultArgs = false,
+#   args = [],
+#   dumpio = false,
+#   executablePath = null,
+#   pipe = false,
+#   env = process.env,
+#   handleSIGINT = true,
+#   handleSIGTERM = true,
+#   handleSIGHUP = true,
+#   ignoreHTTPSErrors = false,
+#   defaultViewport = {width: 800, height: 600},
+#   slowMo = 0,
+#   timeout = 30000
+# } = options;
+# const {
+#   devtools = false,
+#   headless = !devtools,
+#   args = [],
+#   userDataDir = null
+# } = options;
+
+module Puppeteer::Launcher
+  class ChromeArgOptions
+    # * @property {boolean=} headless
+    # * @property {Array<string>=} args
+    # * @property {string=} userDataDir
+    # * @property {boolean=} devtools
+    def initialize(options)
+      @args = options[:args] || []
+      @user_data_dir = options[:user_data_dir]
+      @devtools = options[:devtools] || false
+      @headless = options[:headless]
+      if @headless.nil?
+        @headless = !@devtools
+      end
+    end
+
+    attr_reader :args, :user_data_dir
+
+    def headless?
+      @headless
+    end
+
+    def devtools?
+      @devtools
+    end
+  end
+end

data/lib/puppeteer/launcher/launch_options.rb

@@ -0,0 +1,68 @@
+# const {
+#   ignoreDefaultArgs = false,
+#   args = [],
+#   dumpio = false,
+#   executablePath = null,
+#   pipe = false,
+#   env = process.env,
+#   handleSIGINT = true,
+#   handleSIGTERM = true,
+#   handleSIGHUP = true,
+#   ignoreHTTPSErrors = false,
+#   defaultViewport = {width: 800, height: 600},
+#   slowMo = 0,
+#   timeout = 30000
+# } = options;
+# const {
+#   devtools = false,
+#   headless = !devtools,
+#   args = [],
+#   userDataDir = null
+# } = options;
+
+module Puppeteer::Launcher
+  class LaunchOptions
+    # @property {string=} executablePath
+    # @property {boolean|Array<string>=} ignoreDefaultArgs
+    # @property {boolean=} handleSIGINT
+    # @property {boolean=} handleSIGTERM
+    # @property {boolean=} handleSIGHUP
+    # @property {number=} timeout
+    # @property {boolean=} dumpio
+    # @property {!Object<string, string | undefined>=} env
+    # @property {boolean=} pipe
+    def initialize(options)
+      @executable_path = options[:executable_path]
+      @ignore_default_args = options[:ignore_default_args] || false
+      @handle_SIGINT = options[:handle_SIGINT] || true
+      @handle_SIGTERM = options[:handle_SIGTERM] || true
+      @handle_SIGHUP = options[:handle_SIGHUP] || true
+      @timeout = options[:timeout] || 30000
+      @dumpio = options[:dumpio] || false
+      @env = options[:env] || ENV
+      @pipe = options[:pipe] || false
+    end
+
+    attr_reader :executable_path, :ignore_default_args, :timeout, :env
+
+    def handle_SIGINT?
+      @handle_SIGINT
+    end
+
+    def handle_SIGTERM?
+      @handle_SIGTERM
+    end
+
+    def handle_SIGHUP?
+      @handle_SIGHUP
+    end
+
+    def dumpio?
+      @dumpio
+    end
+
+    def pipe?
+      @pipe
+    end
+  end
+end

data/lib/puppeteer/lifecycle_watcher.rb

@@ -0,0 +1,168 @@
+require 'thread'
+
+# https://github.com/puppeteer/puppeteer/blob/master/lib/LifecycleWatcher.js
+class Puppeteer::LifecycleWatcher
+  include Puppeteer::IfPresent
+
+  class ExpectedLifecycle
+    PUPPETEER_TO_PROTOCOL_LIFECYCLE = {
+      'load' => 'load',
+      'domcontentloaded' => 'DOMContentLoaded',
+      'networkidle0' => 'networkIdle',
+      'networkidle2' => 'networkAlmostIdle',
+    }
+
+    def initialize(wait_until)
+      if wait_until.is_a?(Enumerable)
+        @wait_until = wait_until.map do |value|
+          unless PUPPETEER_TO_PROTOCOL_LIFECYCLE.has_key?(value.to_s)
+            raise ArgumentError.new("Unknown value for options.waitUntil: #{value}")
+          end
+          value.to_s
+        end
+      elsif wait_until.is_a?(String)
+        unless PUPPETEER_TO_PROTOCOL_LIFECYCLE.has_key?(wait_until)
+          raise ArgumentError.new("Unknown value for options.waitUntil: #{wait_until}")
+        end
+        @wait_until = [wait_until]
+      else
+        raise ArgumentError.new('wait_until should be a Array<String> or String')
+      end
+    end
+
+    private def expected_lifecycle
+      @expected_lifecycle ||= @wait_until.map do |value|
+        PUPPETEER_TO_PROTOCOL_LIFECYCLE[value]
+      end
+    end
+
+    # Check if navigation lifecycle has experienced the expected_lifecycle events.
+    #
+    # @param frame [Puppeteer::Frame]
+    def completed?(frame)
+      if expected_lifecycle.any? { |event| !frame.lifecycle_events.include?(event) }
+        return false
+      end
+      if frame.child_frames.any? { |child| !completed?(child) }
+        return false
+      end
+      true
+    end
+  end
+
+  class TerminatedError < StandardError ; end
+
+  #  * @param {!Puppeteer.FrameManager} frameManager
+  #  * @param {!Puppeteer.Frame} frame
+  #  * @param {string|!Array<string>} waitUntil
+  #  * @param {number} timeout
+  def initialize(frame_manager, frame, wait_until, timeout)
+    @expected_lifecycle = ExpectedLifecycle.new(wait_until)
+    @frame_manager = frame_manager
+    @frame = frame
+    @initial_loader_id = frame.loader_id
+    @timeout = timeout
+
+    @listener_ids = {}
+    @listener_ids['client'] = @frame_manager.client.add_event_listener('Events.CDPSession.Disconnected') do
+      terminate(TerminatedError.new('Navigation failed because browser has disconnected!'))
+    end
+    @listener_ids['frame_manager'] = [
+      @frame_manager.add_event_listener('Events.FrameManager.LifecycleEvent') do |frame|
+        check_lifecycle_complete
+      end,
+      @frame_manager.add_event_listener('Events.FrameManager.FrameNavigatedWithinDocument', &method(:navigated_within_document)),
+      @frame_manager.add_event_listener('Events.FrameManager.FrameDetached', &method(:handle_frame_detached)),
+    ]
+    @listener_ids['network_manager'] = @frame_manager.network_manager.add_event_listener('Events.NetworkManager.Request', &method(:handle_request))
+
+    @same_document_navigation_promise = resolvable_future
+    @lifecycle_promise = resolvable_future
+    @new_document_navigation_promise = resolvable_future
+    @termination_promise = resolvable_future
+    check_lifecycle_complete
+  end
+
+  # @param [Puppeteer::Request] request
+  def handle_request(request)
+    return if request.frame != @frame || !request.navigation_request?
+    @navigation_request = request
+  end
+
+  # @param frame [Puppeteer::Frame]
+  def handle_frame_detached(frame)
+    if @frame == frame
+      # this._terminationCallback.call(null, new Error('Navigating frame was detached'));
+      return
+    end
+    check_lifecycle_complete
+  end
+
+  # @return [Puppeteer::Response]
+  def navigation_response
+    if_present(@navigation_request) do |request|
+      request.response
+    end
+  end
+
+  # @param error [TerminatedError]
+  private def terminate(error)
+    @termination_promise.reject(error)
+  end
+
+  attr_reader(
+    :same_document_navigation_promise,
+    :new_document_navigation_promise,
+    :lifecycle_promise,
+  )
+
+  def timeout_or_termination_promise
+    if @timeout > 0
+      future {
+        begin
+          Timeout.timeout(@timeout / 1000.0) do
+            @termination_promise.value!
+          end
+        rescue Timeout::Error
+          raise Puppeteer::FrameManager::NavigationError.new("Navigation timeout of #{@timeout}ms exceeded")
+        end
+      }
+    else
+      @termination_promise
+    end
+  end
+
+  # @param frame [Puppeteer::Frame]
+  private def navigated_within_document(frame)
+    return if frame != @frame
+    @has_same_document_navigation = true
+    check_lifecycle_complete
+  end
+
+  private def check_lifecycle_complete
+    # We expect navigation to commit.
+    return unless @expected_lifecycle.completed?(@frame)
+    @lifecycle_promise.fulfill(true) if @lifecycle_promise.pending?
+    if @frame.loader_id == @initial_loader_id && !@has_same_document_navigation
+      return
+    end
+    if @has_same_document_navigation
+      @same_document_navigation_promise.fulfill(true)
+    end
+    if @frame.loader_id != @initial_loader_id
+      @new_document_navigation_promise.fulfill(true)
+    end
+  end
+
+  def dispose
+    if_present(@listener_ids['client']) do |id|
+      @frame_manager.client.remove_event_listener(id)
+    end
+    if_present(@listener_ids['frame_manager']) do |ids|
+      @frame_manager.remove_event_listener(*ids)
+    end
+    if_present(@listener_ids['network_manager']) do |id|
+      @frame_manager.network_manager.remove_event_listener(id)
+    end
+  end
+end