Ensure fixed amount of scrolling in infinite scroll stories.

tools/perf/page_sets/infinite_scroll_cases.py

-# Copyright 2015 The Chromium Authors. All rights reserved.
+# Copyright 2017 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
-from telemetry.page import page as page_module
+
+import sys
+
+from page_sets.login_helpers import facebook_login
+from page_sets.system_health import platforms
+from telemetry.core import discover
+from telemetry.page import page
 from telemetry.page import shared_page_state
 from telemetry import story
 
-TIME_TO_WAIT_BEFORE_STARTING_IN_SECONDS = 5
-SCROLL_TIMEOUT_IN_SECONDS = 120
+class _InfiniteScrollStory(page.Page):
+  """ Base class for infinite scroll stories."""
 
-# TODO(ulan): Remove this once crbug.com/541508 is fixed.
-STARTUP_SCRIPT = '''
-    window.WebSocket = undefined;
-    window.Worker = undefined;
-    window.performance = undefined;'''
+  NAME = NotImplemented
+  URL = NotImplemented
+  SUPPORTED_PLATFORMS = platforms.ALL_PLATFORMS
 
+  SCROLL_DISTANCE = 25000
+  SCROLL_STEP = 1000
+  MAX_SCROLL_RETRIES = 3
+  TIME_BEFORE_SCROLL_RETRY_IN_SECONDS = 1
+  TIME_TO_WAIT_BEFORE_STARTING_IN_SECONDS = 5
 
-class InfiniteScrollPage(page_module.Page):
-  def __init__(self, url, page_set, name, scroll_amount, delay, repeat,
-               credentials=None):
-    super(InfiniteScrollPage, self).__init__(
-        url=url, page_set=page_set, name=name,
+  def __init__(self, story_set):
+    super(_InfiniteScrollStory, self).__init__(
+        page_set=story_set, url=self.URL, name=self.NAME,
         shared_page_state_class=shared_page_state.SharedPageState,
-       credentials_path='data/credentials.json')
-    self.credentials = credentials
-    self.script_to_evaluate_on_commit = STARTUP_SCRIPT
-    self.scroll_amount = scroll_amount
-    self.delay = delay
-    self.repeat = repeat
+        credentials_path='data/credentials.json')
+    # TODO(ulan): Remove this once crbug.com/541508 is fixed.
+    self.script_to_evaluate_on_commit = '''
+        window.WebSocket = undefined;
+        window.Worker = undefined;
+        window.performance = undefined;'''

[nednguyen, 2017/03/30 13:09:46]

Does this make any of the page's JS fail? Seems a bit scary to me, tbh

[ulan, 2017/03/30 14:35:09]

This part did not change with this CL. At the time of recording there were no
errors.

 
   def RunPageInteractions(self, action_runner):
-    self._WaitAction(action_runner)
-    self._ScrollAction(action_runner, self.scroll_amount, self.delay,
-                       self.repeat)
-
-  def _ScrollAction(self, action_runner, scroll_amount, delay, repeat):
-    with action_runner.CreateInteraction('Begin'):
-      action_runner.tab.browser.DumpMemory()
-    with action_runner.CreateInteraction('Scrolling'):
-      action_runner.RepeatableBrowserDrivenScroll(
-        y_scroll_distance_ratio=scroll_amount,
-        repeat_delay_ms=delay,
-        repeat_count=repeat,
-        timeout=SCROLL_TIMEOUT_IN_SECONDS)
-    with action_runner.CreateInteraction('End'):
-      action_runner.tab.browser.DumpMemory()
-
-  def _WaitAction(self, action_runner):
     with action_runner.CreateInteraction('Load'):
       action_runner.WaitForJavaScriptCondition(
         'document.body != null && '
         'document.body.scrollHeight > window.innerHeight && '
         '!document.body.addEventListener("touchstart", function() {})')
     with action_runner.CreateInteraction('Wait'):
-      action_runner.Wait(TIME_TO_WAIT_BEFORE_STARTING_IN_SECONDS)
+      action_runner.Wait(self.TIME_TO_WAIT_BEFORE_STARTING_IN_SECONDS)
     with action_runner.CreateInteraction('GC'):
       action_runner.ForceGarbageCollection()
+    with action_runner.CreateInteraction('Begin'):

[nednguyen, 2017/03/30 13:09:46]

not in this CL but these interactions are only needed for the legacy TBM
benchmark. I think we can remove them to make the code cleaner now?

[ulan, 2017/03/30 14:35:08]

Yep, will do in separete CL.

+      action_runner.tab.browser.DumpMemory()
+    with action_runner.CreateInteraction('Scrolling'):
+      self._Scroll(action_runner, self.SCROLL_DISTANCE, self.SCROLL_STEP)
+    with action_runner.CreateInteraction('End'):
+      action_runner.tab.browser.DumpMemory()
 
+  def _Scroll(self, action_runner, distance, step_size):

[nednguyen, 2017/03/30 13:09:46]

this method seems useful & common enough to be an action_runner API, like
action_runner.ScrollUntilExactDistanceReached(..). But feel free to do this in a
different CL

[ulan, 2017/03/30 14:35:09]

OK. I would wait until it is needed by another page set.

+    """ This function scrolls the webpage by the given scroll distance in
+    multiple steps, where each step (except the last one) has the given size.
 
-class InfiniteScrollPageSet(story.StorySet):
-  """ Top pages that can be scrolled for many pages. """
+    If scrolling gets stuck, the functions retries scrolling MAX_SCROLL_RETRIES
+    times waiting TIME_BEFORE_SCROLL_RETRY_IN_SECONDS seconds between retries.
+    """
+    remaining = distance - action_runner.EvaluateJavaScript('window.scrollY')
+    retry_count = 0
+    # Scroll until the window.scrollY is within 1 pixel of the target distance.
+    while remaining > 1:
+      action_runner.ScrollPage(distance=min(remaining, step_size) + 1)
+      new_remaining = (distance -
+          action_runner.EvaluateJavaScript('window.scrollY'))
+      if remaining == new_remaining:
+        # Scrolling is stuck. This can happen if the page is loading
+        # resources. Give the page some time and retry scrolling.
+        if retry_count == self.MAX_SCROLL_RETRIES:
+          raise Exception('Scrolling stuck at %d' % remaining)
+        retry_count += 1
+        action_runner.Wait(self.TIME_BEFORE_SCROLL_RETRY_IN_SECONDS)
+      else:
+        retry_count = 0
+        remaining = new_remaining
+
+class DiscourseDesktopStory(_InfiniteScrollStory):
+  NAME = 'discourse'
+  URL = ('https://meta.discourse.org/t/the-official-discourse-tags-plugin' +
+     '-discourse-tagging/26482')
+  SUPPORTED_PLATFORMS = platforms.DESKTOP_ONLY
+
+class DiscourseMobileStory(_InfiniteScrollStory):
+  NAME = 'discourse'
+  URL = ('https://meta.discourse.org/t/the-official-discourse-tags-plugin' +
+     '-discourse-tagging/26482')
+  SUPPORTED_PLATFORMS = platforms.MOBILE_ONLY
+  SCROLL_DISTANCE = 15000
+
+class FacebookDesktopStory(_InfiniteScrollStory):
+  NAME = 'facebook'
+  URL = 'https://www.facebook.com/shakira'
+  SUPPORTED_PLATFORMS = platforms.DESKTOP_ONLY
+
+class FacebookMobileStory(_InfiniteScrollStory):
+  NAME = 'facebook'
+  URL = 'https://m.facebook.com/shakira'
+  SUPPORTED_PLATFORMS = platforms.MOBILE_ONLY
+  def RunNavigateSteps(self, action_runner):
+    facebook_login.LoginWithMobileSite(
+        action_runner, 'facebook3', self.credentials_path)
+    super(FacebookMobileStory, self).RunNavigateSteps(action_runner)
+
+class FlickrDesktopStory(_InfiniteScrollStory):
+  NAME = 'flickr'
+  URL = 'https://www.flickr.com/explore'
+  SUPPORTED_PLATFORMS = platforms.DESKTOP_ONLY
+
+class FlickrMobileStory(_InfiniteScrollStory):
+  NAME = 'flickr'
+  URL = 'https://www.flickr.com/explore'
+  SUPPORTED_PLATFORMS = platforms.MOBILE_ONLY
+  SCROLL_DISTANCE = 10000
+
+class PinterestMobileStory(_InfiniteScrollStory):
+  NAME = 'pinterest'
+  URL = 'https://www.pinterest.com/all'
+  SUPPORTED_PLATFORMS = platforms.MOBILE_ONLY
+
+class TumblrStory(_InfiniteScrollStory):
+  NAME = 'tumblr'
+  URL = 'http://techcrunch.tumblr.com/'
+
+class TwitterDesktopStory(_InfiniteScrollStory):
+  NAME = 'twitter'
+  URL = 'https://twitter.com/taylorswift13'
+  SUPPORTED_PLATFORMS = platforms.DESKTOP_ONLY
+
+class InfiniteScrollStorySet(story.StorySet):
+  """ Desktop story set. """
   def __init__(self):
-    super(InfiniteScrollPageSet, self).__init__(
+    super(InfiniteScrollStorySet, self).__init__(
         archive_data_file='data/infinite_scroll.json',
         cloud_storage_bucket=story.PARTNER_BUCKET)
-    # The scroll distance is chosen such that the page can be scrolled
-    # continuously through the test without hitting the end of the page.
-    SCROLL_FAR = 60
-    SCROLL_PAGE = 1
-    pages = [
-        ('https://www.facebook.com/shakira', 'facebook', SCROLL_FAR, 0, 0),
-        ('https://twitter.com/taylorswift13', 'twitter', SCROLL_PAGE, 10, 30),
-        ('http://techcrunch.tumblr.com/', 'tumblr', SCROLL_FAR, 0, 0),
-        ('https://www.flickr.com/explore', 'flickr', SCROLL_FAR, 0, 0),
-        ('https://meta.discourse.org/t/the-official-discourse-tags-plugin-discourse-tagging/26482',
-         'discourse', SCROLL_PAGE, 10, 30)
-    ]
-    for (url, name, scroll_amount, delay, repeat) in pages:
-      self.AddStory(
-        InfiniteScrollPage(url, self, name, scroll_amount, delay, repeat))
+    for story_class in _FindInfiniteScrollStoryClasses(platforms.DESKTOP):
+      self.AddStory(story_class(self))
+
+class MobileInfiniteScrollStorySet(story.StorySet):
+  """ Mobile story set. """
+  def __init__(self):
+    super(MobileInfiniteScrollStorySet, self).__init__(
+        archive_data_file='data/mobile_infinite_scroll.json',
+        cloud_storage_bucket=story.PARTNER_BUCKET)
+    for story_class in _FindInfiniteScrollStoryClasses(platforms.MOBILE):
+      self.AddStory(story_class(self))
+
+def _FindInfiniteScrollStoryClasses(platform):
+  # Sort the classes by their names so that their order is stable and
+  # deterministic.
+  for unused_cls_name, cls in sorted(discover.DiscoverClassesInModule(
+      module=sys.modules[__name__], base_class=_InfiniteScrollStory,
+      index_by_class_name=True).iteritems()):
+    if platform in cls.SUPPORTED_PLATFORMS:
+      yield cls