Implements peer connection stats measurements in WebRTC call test.

tools/perf/metrics/webrtc_stats.py

Index: tools/perf/metrics/webrtc_stats.py
diff --git a/tools/perf/metrics/webrtc_stats.py b/tools/perf/metrics/webrtc_stats.py
new file mode 100644
index 0000000000000000000000000000000000000000..01ead17a8c6981aa088ec8dd0b03ed27faaece4d
--- /dev/null
+++ b/tools/perf/metrics/webrtc_stats.py
@@ -0,0 +1,73 @@
+# Copyright 2014 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 json
+
+from metrics import Metric
+from telemetry.core import camel_case
+from telemetry.value import list_of_scalar_values
+
+INTERESTING_METRICS = {
+    'packetsReceived': {
+        'units': 'packets',
+        'description': 'Packets received by the peer connection',
+    },
+    'packetsSent': {
+        'units': 'packets',
+        'description': 'Packets sent by the peer connection',
+    },
+    'googDecodeMs': {
+        'units': 'ms',
+        'description': 'Time spent decoding.',
+    },
+    'googMaxDecodeMs': {
+        'units': 'ms',
+        'description': 'Maximum time spent decoding one frame.',
+    },
+    # TODO(phoglund): Add much more interesting metrics.
+}
+
+def SortStatsIntoTimeSeries(report_batches):
+  time_series = {}
+  for report_batch in report_batches:
+    for report in report_batch:
+      for stat_name, value in report.iteritems():
+        if stat_name not in INTERESTING_METRICS:
+          continue
+        time_series.setdefault(stat_name, []).append(float(value))
+
+  return time_series
+
+
+class WebRtcStatisticsMetric(Metric):
+  """Makes it possible to measure stats from peer connections."""
+
+  def __init__(self):
+    super(WebRtcStatisticsMetric, self).__init__()
+    self._all_reports = None
+
+  def Start(self, page, tab):
+    pass
+
+  def Stop(self, page, tab):
+    """Digs out stats from data populated by the javascript in webrtc_cases."""
+    self._all_reports = tab.EvaluateJavaScript(
+        'JSON.stringify(window.peerConnectionReports)')
+
+  def AddResults(self, tab, results):
+    if not self._all_reports:

[tonyg, 2014/09/29 16:45:21]

Are there cases when this is expected? If not, perhaps we want an assertion or
at least a logging.error.

[phoglund_chromium, 2014/09/30 13:37:48]

Yes, the list will be empty if a page didn't end up invoking the peer connection
constructor.

I can add a log statement here. If someone runs with this measurement on a page
that doesn't create peer connections, the measurement is meaningless.

Note that if the javascript from webrtc_cases doesn't load for whatever reason,
we'll fail on #56 already, so the code already makes that distinction at least.

+      return
+
+    reports = json.loads(self._all_reports)
+    num_peer_connections = len(reports)
+    for i in xrange(num_peer_connections):

[phoglund_chromium, 2014/09/12 08:26:07]

This is arguably complex enough to warrant a unit test; what do you think?

[tonyg, 2014/09/29 16:45:22]

Better written as:
for i, peer_connection in enumerate(reports):
  ...

[tonyg, 2014/09/29 16:45:22]

Definitely. The main value is that without a unittest, the stats could just
unexpectedly vanish and the perf dashboard won't let us know. With a unittest,
then any breakage to the stats themselves would fail at the CQ level.

I'd recommend keeping it very simple, just play the shortest webrtc stream and
verify the stats you want are all > 0.

[phoglund_chromium, 2014/09/30 13:37:48]

Hm. I don't think I should play an actual WebRTC stream, since the js code that
does the constructor overriding and stats recording is in webrtc_cases.py. I did
make a test which ensures this code does the right thing given valid JSON,
though. I think that test exercises this code well.

[phoglund_chromium, 2014/09/30 13:37:48]

Done.

+      time_series = SortStatsIntoTimeSeries(reports[i])
+
+      for stat_name, values in time_series.iteritems():
+        trace_name = 'pc_%d_%s' % (i, camel_case.ToUnderscore(stat_name))

[tonyg, 2014/09/29 16:45:21]

What is "pc_"? I recommend a more friendly name here.

[phoglund_chromium, 2014/09/30 13:37:48]

Done.

+        results.AddValue(list_of_scalar_values.ListOfScalarValues(
+            results.current_page, trace_name,
+            INTERESTING_METRICS[stat_name]['units'], values,
+            description=INTERESTING_METRICS[stat_name]['description'],
+            important=False))