Waterfall components of regression range finder.

appengine/findit/waterfall/trigger_flake_swarming_task_pipeline.py

+from google.appengine.ext import ndb
+
+from waterfall.trigger_base_swarming_task_pipeline import (
+    TriggerBaseSwarmingTaskPipeline as TBSTP)
+from model.flake.flake_swarming_task import FlakeSwarmingTask
+
+class TriggerFlakeSwarmingTaskPipeline(TBSTP):
+  """A pipeline to check if selected tests of a step are flaky.
+
+  This pipeline only supports test steps that run on Swarming and support the
+  gtest filter.
+  """
+
+  @ndb.transactional

[stgao, 2016/07/09 00:04:34]

Why this is needed?

[caiw, 2016/07/14 00:59:45]

Done.

+  def _GetSwarmingTask(self,master_name, builder_name, step_name, build_number):
+    # Get the appropriate kind of Swarming Task (Flake).
+    swarming_task = FlakeSwarmingTask.Get(
+        master_name, builder_name, step_name, build_number)
+    return swarming_task
+
+  @ndb.transactional

[stgao, 2016/07/09 00:04:34]

Why this is needed too?

[caiw, 2016/07/14 00:59:45]

Done.

+  def _CreateSwarmingTask(self, master_name, builder_name,
+                          step_name, build_number):
+    # Create the appropriate kind of Swarming Task (Flake).
+    swarming_task = FlakeSwarmingTask.Create(
+        master_name, builder_name,  step_name, build_number)
+    return swarming_task
+
+  def _GetIterationsToRerun(self):
+    # How many times we want to run the swarming rerun.
+    # Placeholder:
+    return 10

[stgao, 2016/07/09 00:04:34]

How about adding a TODO here to move this into the config?

[caiw, 2016/07/14 00:59:45]

Done.