[Findit] Adding schedule_flake_try_job_pipeline.py to trigger flake try jobs

appengine/findit/waterfall/flake/schedule_flake_try_job_pipeline.py

Index: appengine/findit/waterfall/flake/schedule_flake_try_job_pipeline.py
diff --git a/appengine/findit/waterfall/flake/schedule_flake_try_job_pipeline.py b/appengine/findit/waterfall/flake/schedule_flake_try_job_pipeline.py
new file mode 100644
index 0000000000000000000000000000000000000000..4bea3cd5bce99f73e214023bfff1b751e5ea494c
--- /dev/null
+++ b/appengine/findit/waterfall/flake/schedule_flake_try_job_pipeline.py
@@ -0,0 +1,63 @@
+# Copyright 2016 The Chromium Authors. All rights reserved.

[stgao, 2017/01/10 05:56:02]

nit: 2017

[lijeffrey, 2017/01/10 09:32:07]

Done.

+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+from model.flake.flake_try_job import FlakeTryJob
+from model.flake.flake_try_job_data import FlakeTryJobData
+from waterfall import waterfall_config
+from waterfall.schedule_try_job_pipeline import ScheduleTryJobPipeline
+
+
+class ScheduleFlakeTryJobPipeline(ScheduleTryJobPipeline):
+  """A pipeline for scheduling a new flake try job for a flaky test."""
+
+  # Arguments number differs from overridden method - pylint: disable=W0221
+  def _GetBuildProperties(
+      self, master_name, builder_name, step_name, test_name, git_hash):
+    iterations = waterfall_config.GetCheckFlakeSettings().get(
+        'iterations_to_rerun')
+
+    return {
+        'recipe': 'findit/chromium/flake',
+        'target_mastername': master_name,
+        'target_testername': builder_name,
+        'test_revision': git_hash,
+        'test_repeat_count': iterations,
+        'tests': {
+            step_name: [test_name]
+        }
+    }
+
+  def _CreateTryJobData(self, build_id, try_job_key):
+    try_job_data = FlakeTryJobData.Create(build_id)
+    try_job_data.try_job_key = try_job_key
+    try_job_data.put()
+
+  # Arguments number differs from overridden method - pylint: disable=W0221
+  def run(self, master_name, builder_name, step_name, test_name, git_hash):
+    """
+    Args:
+      master_name (str): The master name of a flaky test.
+      builder_name (str): the builder name of a flaky test.

[stgao, 2017/01/10 05:56:02]

nit: the -> The

[lijeffrey, 2017/01/10 09:32:07]

Done.

+      step_name (str): The name of the step the flaky test occurred on.

[stgao, 2017/01/10 05:56:02]

Should it be test_target or step_name?

@chanli: what's the one in the try-job for reliable test failure?

+      test_name (str): The name of the flaky test.
+      git_hash (str): The git hash of the revision to run the try job against.
+
+    Returns:
+      build_id (str): Id of the triggered try job.
+    """
+    properties = self._GetBuildProperties(
+        master_name, builder_name, step_name, test_name, git_hash)
+    build_id = self._TriggerTryJob(master_name, builder_name, properties, {})
+
+    try_job = FlakeTryJob.Get(
+        master_name, builder_name, step_name, test_name, git_hash)
+    try_job.flake_results.append({'try_job_id': build_id})
+    try_job.try_job_ids.append(build_id)
+    try_job.put()
+
+    # Create a corresponding Flake entity to capture as much metadata as early
+    # as possible.
+    self._CreateTryJobData(build_id, try_job.key)
+
+    return build_id