Added skeleton code for the Detection of regression range for flaky

appengine/findit/handlers/flake/check_flaky.py

Index: appengine/findit/handlers/flake/check_flaky.py
diff --git a/appengine/findit/handlers/flake/check_flaky.py b/appengine/findit/handlers/flake/check_flaky.py
new file mode 100644
index 0000000000000000000000000000000000000000..f6c10420b38774cfaa67922b8643aebce809c8e7
--- /dev/null
+++ b/appengine/findit/handlers/flake/check_flaky.py
@@ -0,0 +1,32 @@
+# Copyright 2016 The Chromium Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+from common.base_handler import BaseHandler
+from common.base_handler import Permission
+from waterfall import swarming_util
+
+class CheckFlaky(BaseHandler):

[stgao, 2016/06/08 17:24:12]

style nit: one more line above.

[caiw, 2016/06/13 22:43:56]

Acknowledged.

+  PERMISSION_LEVEL = Permission.CORP_USER
+
+  def HandleGet(self):
+    #get input parameters

[lijeffrey, 2016/06/08 18:05:56]

Style nit:

comments should be in the form

# Get input parameters. <-- space after #, capital first letter, and end with
some punctuation.

Before submitting your code for review you can run

gpylint <path to .py file>

And it'll find most style nits for you

[caiw, 2016/06/13 22:43:55]

Acknowledged.

+    master_name = self.request.get('master_name').strip()
+    builder_name = self.request.get('builder_name').strip()
+    build_number = self.request.get('build_number').strip()
+    test_target_name = self.request.get('test_target_name').strip()
+    testcase_list = self.request.get('testcase_list').strip()

[stgao, 2016/06/08 17:24:12]

To make it simpler, how about starting with a single testcase?

If we do multiple testcases, the Swarming reruns will be not that 
straightforward. Because different flaky tests might have different regression
ranges.

[caiw, 2016/06/13 22:43:56]

Acknowledged.

[stgao, 2016/06/13 22:56:11]

It seems that this comment was not addressed in the latest patch.
What's your thought?

+    print(master_name, builder_name, build_number,

[lijeffrey, 2016/06/08 18:05:56]

This print looks like it's for debugging. Be sure to remove it before committing
the code.

If it's actual logging you want you can use the logging module and look at some
of the other code in Findit for how to use it

[caiw, 2016/06/13 22:43:55]

Acknowledged.

[caiw, 2016/06/13 22:43:56]

Acknowledged.

+          test_target_name, testcase_list)
+
+    # get task id

[chanli, 2016/06/08 20:46:00]

Style nit like ln 13

[caiw, 2016/06/13 22:43:55]

Acknowledged.

[caiw, 2016/06/13 22:43:55]

Acknowledged.

+    tasks = swarming_util.ListSwarmingTasksDataByTags(

[stgao, 2016/06/08 17:24:12]

As we only need the task id for the specific test step (implied by the
test_target_name), we could filter by master_name, builder_name, build_number,
and step_name or name(test_target_name).

[stgao, 2016/06/08 17:24:12]

This will send one or more http request to the Swarming server, and it might
take a while. In order to avoid the 1-minute deadline for the frontend http
request, it seems better to move this to the App Engine task or App Engine
pipeline which will run on the module waterfall-backend.

[caiw, 2016/06/13 22:43:55]

Acknowledged.

[caiw, 2016/06/13 22:43:56]

Acknowledged.

[caiw, 2016/06/13 22:43:56]

Acknowledged.

[caiw, 2016/06/14 19:48:05]

Acknowledged.

+        master_name=master_name, builder_name=builder_name,
+        build_number=build_number,http_client='')
+    print(tasks)
+    # TODO: trigger swarming reruns

[lijeffrey, 2016/06/08 18:05:57]

Nit: for TODOs, if you plan to do them yourself, you should put your ldap after
or a bug number.

i.e.

# TODO(caiw): Trigger swarming reruns.

[caiw, 2016/06/13 22:43:56]

Acknowledged.

+    # TODO: get the results from swarming reruns
+    # TODO: this should return some kind of template
+
+  def HandlePost(self):  # pragma: no cover
+    return self.HandleGet()