[Findit] Reduce redundant ndb reads by querying necessary entities ahead of time and share them amo…

appengine/findit/waterfall/suspected_cl_util.py

Index: appengine/findit/waterfall/suspected_cl_util.py
diff --git a/appengine/findit/waterfall/suspected_cl_util.py b/appengine/findit/waterfall/suspected_cl_util.py
index 28b2786d10d19910c27d430667cd58f96c8e5864..d191b567dca7e625e6de8f0210e6d907a19fc98a 100644
--- a/appengine/findit/waterfall/suspected_cl_util.py
+++ b/appengine/findit/waterfall/suspected_cl_util.py
@@ -60,18 +60,18 @@ def UpdateSuspectedCL(
 
   suspected_cl.put()
 
+
 def _RoundConfidentToInteger(confidence):
-  return round(confidence * 100)
+  return int(round(confidence * 100))
 
 
 def GetSuspectedCLConfidenceScore(confidences, cl_build):
-
   if not confidences or not cl_build:
     return None
 
   if cl_build['failure_type'] == failure_type.COMPILE:
-    if cl_build['approaches'] == [
-      analysis_approach_type.HEURISTIC, analysis_approach_type.TRY_JOB]:
+    if sorted(cl_build['approaches']) == sorted([
+      analysis_approach_type.HEURISTIC, analysis_approach_type.TRY_JOB]):
       return _RoundConfidentToInteger(
           confidences.compile_heuristic_try_job.confidence)
     elif cl_build['approaches'] == [analysis_approach_type.TRY_JOB]:
@@ -84,8 +84,8 @@ def GetSuspectedCLConfidenceScore(confidences, cl_build):
           return _RoundConfidentToInteger(confidences_info.confidence)
     return None
   else:
-    if cl_build['approaches'] == [
-      analysis_approach_type.HEURISTIC, analysis_approach_type.TRY_JOB]:
+    if sorted(cl_build['approaches']) == sorted([
+      analysis_approach_type.HEURISTIC, analysis_approach_type.TRY_JOB]):
       return _RoundConfidentToInteger(
           confidences.test_heuristic_try_job.confidence)
     elif cl_build['approaches'] == [analysis_approach_type.TRY_JOB]:
@@ -96,3 +96,39 @@ def GetSuspectedCLConfidenceScore(confidences, cl_build):
         if confidences_info.score == cl_build['top_score']:
           return _RoundConfidentToInteger(confidences_info.confidence)
     return None
+
+
+def _HasNewFailures(current_failures, new_failures):
+  """Checks if there are any new failures in the current build."""
+  if current_failures == new_failures:
+    return False
+
+  for step, tests in current_failures.iteritems():
+    if not new_failures.get(step):  # New step.
+      return True
+
+    for test in tests:
+      if not test in new_failures[step]:  # New test.
+        return True
+
+  return False
+
+
+def GetSuspectedCLConfidenceScoreAndApproach(
+    confidences, cl_build, cl_first_build):

[lijeffrey, 2016/10/28 02:32:32]

nit: rename these cl_from_analyzed_build and  cl_from_first_failed_build for
clarity

[chanli, 2016/10/28 02:41:46]

Done.

+  if not confidences or not cl_build:
+    return None, None
+
+  if (cl_first_build and not _HasNewFailures(
+      cl_build.get('failures'), cl_first_build.get('failures'))):
+      # For non-first-time failures, the try job result is not recorded.
+      # If there is no new failures in current build, use first failed build to
+      # make sure the confidence score is correct.
+      cl_build = cl_first_build
+
+  confidence = GetSuspectedCLConfidenceScore(confidences, cl_build)
+  approach = (
+      analysis_approach_type.TRY_JOB if analysis_approach_type.TRY_JOB in
+      cl_build['approaches'] else analysis_approach_type.HEURISTIC)
+
+  return confidence, approach