[Findit] Use step level analysis to exclude flaky test failures.

appengine/findit/waterfall/buildbot.py

 # Copyright 2014 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 datetime import datetime
 import json
 import re
 import urllib
+import gzip

[stgao, 2015/05/21 00:29:56]

not in order.

+
+import cloudstorage as gcs
 
 from waterfall.build_info import BuildInfo
 
 _MASTER_URL_PATTERN = re.compile(r'^https?://build\.chromium\.org/p/([^/]+)'
                                   '(/.*)?$')
 
 _BUILD_URL_PATTERN = re.compile(r'^https?://build\.chromium\.org/p/([^/]+)/'
                                  'builders/([^/]+)/builds/([\d]+)(/.*)?$')
 
 _STEP_URL_PATTERN = re.compile(r'^https?://build\.chromium\.org/p/([^/]+)/'
@@ skipping 71 matching lines @@
 
 
 def CreateStdioLogUrl(master_name, builder_name, build_number, step_name):
   builder_name = urllib.quote(builder_name)
   step_name = urllib.quote(step_name)
   return ('https://build.chromium.org/p/%s/builders/%s/builds/%s/'
           'steps/%s/logs/stdio/text') % (
               master_name, builder_name, build_number, step_name)
 
 
-def GetBuildDataFromBuildMaster(master_name, 
+def CreateStepLogPath(master_name, builder_name, build_number, step_name):

[stgao, 2015/05/21 00:29:56]

"GtestResult" might be more clear than "StepLog". Same for others below.

+  return ('/chrome-gtest-results/buildbot/%s/%s/%s/%s.json.gz') % (
+              master_name, builder_name, build_number, step_name)
+
+
+def GetBuildDataFromBuildMaster(master_name,
                                 builder_name, build_number, http_client):
   """Returns the json-format data of the build from buildbot json API."""
   status_code, data = http_client.Get(
       CreateBuildUrl(master_name, builder_name, build_number, json_api=True))
   if status_code != 200:
     return None
   else:
     return data
 
 
-def GetBuildDataFromArchive(master_name, 
+def GetBuildDataFromArchive(master_name,
                             builder_name, build_number, http_client):
   """Returns the json-format data of the build from build archive."""
   status_code, data = http_client.Get(
       CreateArchivedBuildUrl(master_name, builder_name, build_number))
   if status_code != 200:
     return None
   else:
     return data
 
 
 def GetStepStdio(master_name, builder_name, build_number,
                  step_name, http_client):
   """Returns the raw string of stdio of the specified step."""
   status_code, data = http_client.Get(
       CreateStdioLogUrl(master_name, builder_name, build_number, step_name))
   if status_code != 200:
     return None
   else:
     return data
 
 
+def GetGsStepLog(master_name,
+                 builder_name, build_number, step_name):  # pragma: no cover
+  """Returns the archived Step Log file."""
+  try:
+    step_log_file = gcs.open(CreateStepLogPath(

[stgao, 2015/05/21 00:29:56]

Let us use "with" for the two opens here. It is cleaner.

Please refer to contextlib.closing in
https://chromium.googlesource.com/infra/infra/+/master/appengine/sheriff_o_ma...

+        master_name, builder_name, build_number, step_name))
+    unzipped_step_log_file = gzip.open(step_log_file )
+    log_file_content = step_log_file.read()
+    unzipped_step_log_file.close()
+    step_log_file.close()
+    return log_file_content
+  except gcs.NotFoundError:
+    return None
+
+
 def GetStepResult(step_data_json):
   """Returns the result of a step."""
   result = step_data_json.get('results')
   if result is None and step_data_json.get('isFinished'):
     # Without parameter filter=0 in the http request to the buildbot json api,
     # the value of the result of a passed step won't be present.
     return SUCCESS
 
   while isinstance(result, list):
     result = result[0]
@@ skipping 53 matching lines @@
       # Skip steps that haven't started yet or are still running.
       continue
 
     step_result = GetStepResult(step_data)
     if step_result in (SUCCESS, WARNINGS):
       build_info.passed_steps.append(step_name)
     elif step_result == FAILURE:
       build_info.failed_steps.append(step_name)
 
   return build_info