Add HTML parsing benchmark.

tools/perf/benchmarks/html-parser.py

Index: tools/perf/benchmarks/html-parser.py
diff --git a/tools/perf/benchmarks/html-parser.py b/tools/perf/benchmarks/html-parser.py
new file mode 100644
index 0000000000000000000000000000000000000000..e90c4bd330df39da544add5e75e661f36cdf5fb8
--- /dev/null
+++ b/tools/perf/benchmarks/html-parser.py
@@ -0,0 +1,67 @@
+# 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 os
+
+from core import perf_benchmark
+from telemetry import page as page_module
+from telemetry import story
+from telemetry.page import legacy_page_test
+from telemetry.value import scalar
+
+
+class _HtmlParserMeasurement(legacy_page_test.LegacyPageTest):
+  def __init__(self):
+    super(_HtmlParserMeasurement, self).__init__()
+
+  def _GetResult(self, tab, index):
+    selector = 'document.querySelectorAll("section")[%d]' % index
+    tab.WaitForJavaScriptExpression(
+        '((%s) && (%s).textContent)' % (selector, selector), 100)
+    extract_time_js = '(/time: (\\d+)ms/.exec(%s.textContent))[1]' % selector
+    return float(tab.EvaluateJavaScript(extract_time_js))
+
+  def ValidateAndMeasurePage(self, page, tab, results):
+    innerHtmlTime = self._GetResult(tab, 0)
+    realBindingsTime = self._GetResult(tab, 1)
+    fakeBindingsTime = self._GetResult(tab, 2)
+    bindingsOverhead = max(0, realBindingsTime - fakeBindingsTime)
+    results.AddValue(scalar.ScalarValue(
+        results.current_page, 'innerHTML', 'ms', innerHtmlTime))
+    results.AddValue(scalar.ScalarValue(
+        results.current_page, 'real-bindings', 'ms', realBindingsTime))
+    results.AddValue(scalar.ScalarValue(
+        results.current_page, 'fake-bindings', 'ms', fakeBindingsTime))
+    results.AddValue(scalar.ScalarValue(
+        results.current_page, 'bindings-overhead', 'ms', bindingsOverhead))
+
+
+class HtmlParser(perf_benchmark.PerfBenchmark):
+  """The benchmark parses a saved page of wikipedia.
+
+  This is intended for real production use so authors could schedule HTML
+  parsing as needed, and to model data parsing into DOM.
+  It tests 3 things:
+
+  1. Parsing with the browser provided C++ innerHTML
+  2. Parsing with the JS based html parser, creating real DOM.
+  3. Parsing with the JS based html parser, creating *fake* DOM.
+  """
+  test = _HtmlParserMeasurement
+
+  @classmethod
+  def Name(cls):
+    return 'html-parser'
+
+  def CreateStorySet(self, options):
+    ps = story.StorySet(
+        archive_data_file='../page_sets/data/html-parser.json',
+        base_dir=os.path.dirname(os.path.abspath(__file__)),
+        cloud_storage_bucket=story.PARTNER_BUCKET)
+    ps.AddStory(page_module.Page(
+        'https://1ede9c8b4261401ce30e527057727688b46ce0d9.googledrive.com/'
+        'host/0B4QUVw-AB8wPVXpWVzZ5T1hObGc/dom_create.html',
+        ps, ps.base_dir, name='html-parser',
+        make_javascript_deterministic=False))
+    return ps