Added support for improvement_direction to relevant values, which is propogated to chartjson.

tools/perf/metrics/timeline.py

Index: tools/perf/metrics/timeline.py
diff --git a/tools/perf/metrics/timeline.py b/tools/perf/metrics/timeline.py
index 54ac3951b2b899844c56c4852eca1e0454d7fcb3..390ef1d887f82651a2cbaecdfc60b14c433389b0 100644
--- a/tools/perf/metrics/timeline.py
+++ b/tools/perf/metrics/timeline.py
@@ -47,11 +47,14 @@ class LoadTimesTimelineMetric(timeline_based_metric.TimelineBasedMetric):
 
         full_name = thread_name + '|' + sanitized_event_name
         results.AddValue(scalar.ScalarValue(
-            results.current_page, full_name, 'ms', total))
+            results.current_page, full_name, 'ms', total,
+            higher_is_better=False))
         results.AddValue(scalar.ScalarValue(
-            results.current_page, full_name + '_max', 'ms', biggest_jank))
+            results.current_page, full_name + '_max', 'ms', biggest_jank,
+            higher_is_better=False))
         results.AddValue(scalar.ScalarValue(
-            results.current_page, full_name + '_avg', 'ms', total / len(times)))
+            results.current_page, full_name + '_avg', 'ms', total / len(times),
+            higher_is_better=False))
 
     for counter_name, counter in renderer_process.counters.iteritems():
       total = sum(counter.totals)
@@ -60,10 +63,11 @@ class LoadTimesTimelineMetric(timeline_based_metric.TimelineBasedMetric):
       sanitized_counter_name = counter_name.replace('.', '_')
 
       results.AddValue(scalar.ScalarValue(
-          results.current_page, sanitized_counter_name, 'count', total))
+          results.current_page, sanitized_counter_name, 'count', total,
+          higher_is_better=False))
       results.AddValue(scalar.ScalarValue(
           results.current_page, sanitized_counter_name + '_avg', 'count',
-          total / float(len(counter.totals))))
+          total / float(len(counter.totals)), higher_is_better=False))
 
 # We want to generate a consistant picture of our thread usage, despite
 # having several process configurations (in-proc-gpu/single-proc).
@@ -195,10 +199,10 @@ class ResultsForThread(object):
     tasks_per_frame = Rate(len(self.toplevel_slices), num_frames)
     results.AddValue(scalar.ScalarValue(
         results.current_page, ThreadCpuTimeResultName(self.name),
-        'ms', cpu_per_frame))
+        'ms', cpu_per_frame, higher_is_better=False))
     results.AddValue(scalar.ScalarValue(
         results.current_page, ThreadTasksResultName(self.name),
-        'tasks', tasks_per_frame))
+        'tasks', tasks_per_frame, higher_is_better=False))
     # Report mean frame time if this is the thread we are using for normalizing
     # other results. We could report other frame rates (eg. renderer_main) but
     # this might get confusing.
@@ -207,7 +211,7 @@ class ResultsForThread(object):
       mean_frame_time = Rate(self.all_action_time, num_frames)
       results.AddValue(scalar.ScalarValue(
           results.current_page, ThreadMeanFrameTimeResultName(self.name),
-          'ms', mean_frame_time))
+          'ms', mean_frame_time, higher_is_better=False))
 
   def AddDetailedResults(self, num_frames, results):
     slices_by_category = collections.defaultdict(list)
@@ -220,13 +224,13 @@ class ResultsForThread(object):
       self_time_result = (float(self_time) / num_frames) if num_frames else 0
       results.AddValue(scalar.ScalarValue(
           results.current_page, ThreadDetailResultName(self.name, category),
-          'ms', self_time_result))
+          'ms', self_time_result, higher_is_better=False))
     all_measured_time = sum(all_self_times)
     idle_time = max(0, self.all_action_time - all_measured_time)
     idle_time_result = (float(idle_time) / num_frames) if num_frames else 0
     results.AddValue(scalar.ScalarValue(
         results.current_page, ThreadDetailResultName(self.name, "idle"),
-        'ms', idle_time_result))
+        'ms', idle_time_result, higher_is_better=False))
 
   def CountTracesWithName(self, substring):
     count = 0