These are the database objects used while finding the regression range. 1/3

appengine/findit/model/base_flake_model.py

Index: appengine/findit/model/base_flake_model.py
diff --git a/appengine/findit/model/base_build_model.py b/appengine/findit/model/base_flake_model.py
similarity index 69%
copy from appengine/findit/model/base_build_model.py
copy to appengine/findit/model/base_flake_model.py
index 73a9f84dd1741ac77df461882f852899f943be6f..afba53ff2d7fe46dbb95b941821b73b0435b5105 100644
--- a/appengine/findit/model/base_build_model.py
+++ b/appengine/findit/model/base_flake_model.py
@@ -5,7 +5,7 @@
 from google.appengine.ext import ndb
 
 
-class BaseBuildModel(ndb.Model):  # pragma: no cover
+class BaseFlakeModel(ndb.Model):  # pragma: no cover
   """A base class to provide computed properties from the key.
 
   The computed properties are master name, builder name, and build number.
@@ -16,8 +16,8 @@ class BaseBuildModel(ndb.Model):  # pragma: no cover
   """
 
   @staticmethod
-  def CreateBuildId(master_name, builder_name, build_number):
-    return '%s/%s/%s' % (master_name, builder_name, build_number)
+  def CreateFlakeId(master_name, builder_name, step_name, build_number):
+    return '%s/%s/%s/%s' % (master_name, builder_name, step_name, build_number)

[stgao, 2016/07/14 18:01:33]

Is there a special consideration of putting step name before build number?

If they could be swapped, we could reuse code in base_build_model.py

 
   @ndb.ComputedProperty
   def master_name(self):
@@ -28,5 +28,10 @@ class BaseBuildModel(ndb.Model):  # pragma: no cover
     return self.key.pairs()[0][1].split('/')[1]
 
   @ndb.ComputedProperty
+  def step_name(self):
+    return self.key.pairs()[0][1].split('/')[2]
+
+  @ndb.ComputedProperty
   def build_number(self):
-    return int(self.key.pairs()[0][1].split('/')[2])
+    return int(self.key.pairs()[0][1].split('/')[3]
+)