Add benchmark that imitates news reading.

tools/perf/page_sets/news_stories.py

Index: tools/perf/page_sets/news_stories.py
diff --git a/tools/perf/page_sets/news_stories.py b/tools/perf/page_sets/news_stories.py
new file mode 100644
index 0000000000000000000000000000000000000000..553b6a9561b57dd14d159d994e1f77cd9fc94639
--- /dev/null
+++ b/tools/perf/page_sets/news_stories.py
@@ -0,0 +1,213 @@
+# Copyright 2016 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.
+
+import sys
+
+from telemetry.page import page
+
+from telemetry.core import discover
+from telemetry import story
+
+_ALL_PLATFORMS = frozenset({'desktop', 'mobile'})
+_DESKTOP_ONLY = frozenset({'desktop'})
+_MOBILE_ONLY = frozenset({'mobile'})
+_NO_PLATFORMS = frozenset()
+_NEWS_ITEM_READ_TIME_IN_SECONDS = 3
+
+class _NewsStory(page.Page):

[petrcermak, 2016/07/07 14:55:56]

I can see that this class is modelled very closely after SinglePageStory. My
suggestion is that you:

1) Move this file to tools/perf/page_sets/system_health/browsing_stories.py.
2) Do the same for https://codereview.chromium.org/2124823002/ (merge the
stories into browsing_stories).
3) Change single_page_stories.py to export the SinglePageStory subclass (remove
underscore from name) so that you can re-use the code.
4) Re-name the stories to 'browse:news:XYZ' and 'browse:media:XYZ' to be in line
with the remaining SH stories.

Please do NOT add the stories to the SystemHealthStorySet just yet because your
stories rely on periodic dumps while SH stories use explicit dumps, so we need
to figure out how to make this work for everyone.

How does that sound?

[ulan, 2016/07/07 17:34:38]

Sounds good. I will do this change tomorrow.

+  """Abstract base class for news user stories.
+
+  A news story imitates browsing a news website:
+  1. Load the main page.
+  2. Open and scroll the first news item.
+  3. Go back to the main page and scroll it.
+  4. Open and scroll the second news item.
+  5. Go back to the main page and scroll it.
+  6. etc.
+  """
+
+  NAME = NotImplemented
+  URL = NotImplemented
+  NEWS_ITEM_SELECTOR = NotImplemented
+  NEWS_ITEMS_TO_VISIT = 4
+  SUPPORTED_PLATFORMS = _ALL_PLATFORMS
+  IS_SINGLE_PAGE_APP = False
+
+  def __init__(self, story_set):
+    super(_NewsStory, self).__init__(
+        page_set=story_set, name=self.NAME, url=self.URL)
+
+  def RunPageInteractions(self, action_runner):
+    action_runner.tab.WaitForDocumentReadyStateToBeComplete()
+    self._DidLoadDocument(action_runner)
+    for i in xrange(self.NEWS_ITEMS_TO_VISIT):
+      self._ClickNewsItem(action_runner, i)
+      self._ReadNewsItem(action_runner)
+      self._NavigateBack(action_runner)
+      self._ScrollMainPage(action_runner)
+
+  def _NewsItem(self, index):
+    return 'document.querySelectorAll(\'%s\')[%d]' % (
+        self.NEWS_ITEM_SELECTOR, index)
+
+  def _DidLoadDocument(self, action_runner):
+    pass
+
+  def _WaitForNavigation(self, action_runner):
+    if not self.IS_SINGLE_PAGE_APP:
+      action_runner.WaitForNavigate()
+
+  def _ClickNewsItem(self, action_runner, index):
+    action_runner.WaitForElement(element_function=self._NewsItem(index))
+    action_runner.ClickElement(element_function=self._NewsItem(index))
+    self._WaitForNavigation(action_runner)
+
+  def _ReadNewsItem(self, action_runner):
+    action_runner.tab.WaitForDocumentReadyStateToBeComplete()
+    action_runner.Wait(_NEWS_ITEM_READ_TIME_IN_SECONDS)
+    action_runner.RepeatableBrowserDrivenScroll(repeat_count=2)
+
+  def _NavigateBack(self, action_runner):
+    action_runner.ExecuteJavaScript('window.history.back()')
+    self._WaitForNavigation(action_runner)
+
+  def _ScrollMainPage(self, action_runner):
+    action_runner.tab.WaitForDocumentReadyStateToBeComplete()
+    action_runner.RepeatableBrowserDrivenScroll(repeat_count=0)
+
+
+def IterAllStoryClasses():
+  # 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=_NewsStory,
+      index_by_class_name=True).iteritems()):
+    yield cls
+
+
+##############################################################################
+# News stories.
+##############################################################################
+
+
+class RedditDesktopStory(_NewsStory):
+  """The top website in http://www.alexa.com/topsites/category/News"""
+  NAME = 'reddit'
+  URL = 'https://www.reddit.com/top/?sort=top&t=week'
+  SUPPORTED_PLATFORMS = _DESKTOP_ONLY
+  NEWS_ITEM_SELECTOR = '.thing .title > a'
+
+
+class RedditMobileStory(_NewsStory):
+  """The top website in http://www.alexa.com/topsites/category/News"""
+  NAME = 'reddit'
+  URL = 'https://m.reddit.com/?sort=top&time=week'
+  SUPPORTED_PLATFORMS = _MOBILE_ONLY
+  NEWS_ITEM_SELECTOR = '.PostHeader__post-title-line'
+  IS_SINGLE_PAGE_APP = True
+
+
+class CnnStory(_NewsStory):
+  """The second top website in http://www.alexa.com/topsites/category/News"""
+  NAME = 'cnn'
+  URL = 'http://edition.cnn.com/'
+  NEWS_ITEM_SELECTOR = '.cd__content > h3 > a'
+  # TODO(ulan): Enable this story on mobile once it uses less memory and
+  # does not crash with OOM.
+  SUPPORTED_PLATFORMS = _DESKTOP_ONLY
+
+
+class NytimesMobileStory(_NewsStory):
+  """The third top website in http://www.alexa.com/topsites/category/News"""
+  NAME = 'nytimes'
+  URL = 'http://mobile.nytimes.com'
+  SUPPORTED_PLATFORMS = _MOBILE_ONLY
+  NEWS_ITEM_SELECTOR = '.sfgAsset-link'
+  # Visiting more items causes OOM.
+  NEWS_ITEMS_TO_VISIT = 2
+
+
+class NytimesDesktopStory(_NewsStory):
+  """The third top website in http://www.alexa.com/topsites/category/News"""
+  NAME = 'nytimes'
+  URL = 'http://www.nytimes.com'
+  SUPPORTED_PLATFORMS = _DESKTOP_ONLY
+  NEWS_ITEM_SELECTOR = '.story-heading > a'
+
+
+class HackerNewsStory(_NewsStory):
+  NAME = 'hackernews'
+  URL = 'https://news.ycombinator.com'
+  NEWS_ITEM_SELECTOR = '.athing .title > a'
+
+
+# Facebook on desktop is not interesting because it embeds post comments
+# directly in the main timeline.
+class FacebookMobileStory(_NewsStory):
+  NAME = 'facebook'
+  URL = 'https://facebook.com/natgeo'
+  NEWS_ITEM_SELECTOR = 'article ._5msj'
+  SUPPORTED_PLATFORMS = _MOBILE_ONLY
+
+
+class TwitterMobileStory(_NewsStory):
+  NAME = 'twitter'
+  URL = 'https://mobile.twitter.com/cnnbrk'
+  NEWS_ITEM_SELECTOR = '.Tweet-text'
+  SUPPORTED_PLATFORMS = _MOBILE_ONLY
+
+
+class TwitterDesktopStory(_NewsStory):
+  NAME = 'twitter'
+  URL = 'https://twitter.com/cnnbrk'
+  NEWS_ITEM_SELECTOR = '.tweet-text'
+  SUPPORTED_PLATFORMS = _DESKTOP_ONLY
+  IS_SINGLE_PAGE_APP = True
+
+
+class WashingtonPostMobileStory(_NewsStory):
+  """Progressive website"""
+  NAME = 'washingtonpost'
+  URL = 'https://www.washingtonpost.com/pwa'
+  SUPPORTED_PLATFORMS = _MOBILE_ONLY
+  NEWS_ITEM_SELECTOR = '.hed > a'
+  IS_SINGLE_PAGE_APP = True
+
+  def _DidLoadDocument(self, action_runner):
+    # Close the popup window.
+    action_runner.ClickElement(selector='.close')
+
+
+class WashingtonPostDesktopStory(_NewsStory):
+  NAME = 'washingtonpost'
+  URL = 'https://www.washingtonpost.com'
+  SUPPORTED_PLATFORMS = _DESKTOP_ONLY
+  NEWS_ITEM_SELECTOR = '.headline > a'
+
+
+##############################################################################
+# News story sets.
+##############################################################################
+
+
+class _NewsStorySet(story.StorySet):
+  PLATFORM = NotImplemented
+
+  def __init__(self):
+    super(_NewsStorySet, self).__init__(
+        archive_data_file=('../data/news_%s.json' % self.PLATFORM),
+        cloud_storage_bucket=story.PARTNER_BUCKET)
+    for story_class in IterAllStoryClasses():
+      if self.PLATFORM not in story_class.SUPPORTED_PLATFORMS:
+        continue
+      self.AddStory(story_class(self))
+
+
+class DesktopNewsStorySet(_NewsStorySet):
+  PLATFORM = 'desktop'
+
+
+class MobileNewsStorySet(_NewsStorySet):
+  PLATFORM = 'mobile'