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

tools/telemetry/telemetry/value/scalar.py

Index: tools/telemetry/telemetry/value/scalar.py
diff --git a/tools/telemetry/telemetry/value/scalar.py b/tools/telemetry/telemetry/value/scalar.py
index d83753af414e27ab98aa576ee823413b400d728f..e838f2d81316cf81c2c47e1fd4520e26a2b03e35 100644
--- a/tools/telemetry/telemetry/value/scalar.py
+++ b/tools/telemetry/telemetry/value/scalar.py
@@ -7,18 +7,21 @@ import numbers
 from telemetry import value as value_module
 from telemetry.value import list_of_scalar_values
 from telemetry.value import none_values
+from telemetry.value import summarizable
 
 
-class ScalarValue(value_module.Value):
+class ScalarValue(summarizable.SummarizableValue):
   def __init__(self, page, name, units, value, important=True,
-               description=None, none_value_reason=None):
+               description=None, none_value_reason=None,
+               improvement_direction=None):
     """A single value (float or integer) result from a test.
 
     A test that counts the number of DOM elements in a page might produce a
     scalar value:
        ScalarValue(page, 'num_dom_elements', 'count', num_elements)
     """
-    super(ScalarValue, self).__init__(page, name, units, important, description)
+    super(ScalarValue, self).__init__(
+        page, name, units, important, description, improvement_direction)
     assert value is None or isinstance(value, numbers.Number)
     none_values.ValidateNoneValueReason(value, none_value_reason)
     self.value = value
@@ -29,13 +32,15 @@ class ScalarValue(value_module.Value):
       page_name = self.page.url
     else:
       page_name = None
-    return 'ScalarValue(%s, %s, %s, %s, important=%s, description=%s)' % (
+    return ('ScalarValue(%s, %s, %s, %s, important=%s, description=%s, '
+            'improvement_direction=%s)') % (
       page_name,
       self.name,
       self.units,
       self.value,
       self.important,
-      self.description)
+      self.description,
+      self.improvement_direction)
 
   def GetBuildbotDataType(self, output_context):
     if self._IsImportantGivenOutputIntent(output_context):
@@ -70,6 +75,7 @@ class ScalarValue(value_module.Value):
   def FromDict(value_dict, page_dict):
     kwargs = value_module.Value.GetConstructorKwArgs(value_dict, page_dict)
     kwargs['value'] = value_dict['value']
+    kwargs['improvement_direction'] = value_dict['improvement_direction']
 
     if 'none_value_reason' in value_dict:
       kwargs['none_value_reason'] = value_dict['none_value_reason']
@@ -100,4 +106,5 @@ class ScalarValue(value_module.Value):
       none_value_reason = none_values.MERGE_FAILURE_REASON
     return list_of_scalar_values.ListOfScalarValues(
         page, name, v0.units, merged_value, important=v0.important,
-        none_value_reason=none_value_reason)
+        none_value_reason=none_value_reason,
+        improvement_direction=v0.improvement_direction)

[nednguyen, 2015/01/08 18:07:41]

This is the ugliness that I mentioned. How do we guarantee that all values to be
merge have the same improvement_direction? :-(

It's fine if you want to punt this issue.

[qyearsley, 2015/01/27 00:50:46]

What does "like values" mean in this context? Maybe based on what's meant here,
it does make sense to assume that they'll all always have the same improvement
direction (e.g. if they all represent the same type of measurement?) If so, then
a comment could be added to explain this.