Refactored auto_bisect bot:

 - Extracted BisectPrinter, which contains everything related to printing
 - Extracted BisectState and RevisionState, which represent bisect-in-progress
 - Rewrote BisectResults - in particular split up GetRevisionDict, which is now its constructor
 - Added tests for BisectResults, fixed bugs in _FindOtherRegressions
 - Added tests for BisectState

R=qyearsley@chromium.org

Review URL: https://codereview.chromium.org/664753002

Cr-Commit-Position: refs/heads/master@{#300677}

1 files changed, 242 insertions, 24 deletions

diff --git a/tools/auto_bisect/bisect_results_test.py b/tools/auto_bisect/bisect_results_test.py
index 361f5f1..4b77806 100644
--- a/tools/auto_bisect/bisect_results_test.py
+++ b/tools/auto_bisect/bisect_results_test.py
@@ -2,38 +2,256 @@
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
+import os
 import unittest
 
-import bisect_results
-import ttest
+from bisect_results import BisectResults
+import source_control
 
 
-class ConfidenceScoreTest(unittest.TestCase):
+class MockDepotRegistry(object):
+  def ChangeToDepotDir(self, depot):
+    pass
+
+
+class MockRevisionState(object):
+  def __init__(self, revision, index, depot='chromium', value=None,
+               perf_time=0, build_time=0, passed='?', external=None):
+    self.depot = depot
+    self.revision = revision
+    self.index = index
+    self.value = value
+    self.perf_time = perf_time
+    self.build_time = build_time
+    self.passed = passed
+    self.external = external
+
+
+class MockBisectState(object):
+
+  def __init__(self):
+    self.mock_revision_states = []
+
+    mock_bad_val = {'values': [100, 105, 95]}
+    for i, rev in enumerate(['a', 'b']):
+      mock_rev_state = MockRevisionState(rev, i, value=mock_bad_val, passed=0)
+      self.mock_revision_states.append(mock_rev_state)
+
+    mock_good_val = {'values': [1, 2, 3]}
+    for i, rev in enumerate(['c', 'd', 'e'], start=2):
+      mock_rev_state = MockRevisionState(rev, i, value=mock_good_val, passed=1)
+      self.mock_revision_states.append(mock_rev_state)
+
+  def GetRevisionStates(self):
+    return self.mock_revision_states
+
+
+class MockBisectOptions(object):
+
+  def __init__(self):
+    self.repeat_test_count = 3
+
+
+class BisectResultsTest(unittest.TestCase):
+
+  def setUp(self):
+    self.mock_bisect_state = MockBisectState()
+    self.mock_depot_registry = MockDepotRegistry()
+    self.mock_opts = MockBisectOptions()
+    self.mock_warnings = []
+
+    self.original_getcwd = os.getcwd
+    self.original_chdir = os.chdir
+    self.original_query_revision_info = source_control.QueryRevisionInfo
+
+    os.getcwd = lambda: '/path'
+    os.chdir = lambda _: None
+
+    revision_infos = {'b': {'test': 'b'}, 'c': {'test': 'c'}}
+    source_control.QueryRevisionInfo = lambda rev: revision_infos[rev]
+
+  def tearDown(self):
+    os.getcwd = self.original_getcwd
+    os.chdir = self.original_chdir
+    source_control.QueryRevisionInfo = self.original_query_revision_info
+
+  def _AssertConfidence(self, score, bad_values, good_values):
+    """Checks whether the given sets of values have a given confidence score.
+
+    The score represents our confidence that the two sets of values wouldn't
+    be as different as they are just by chance; that is, that some real change
+    occurred between the two sets of values.
+
+    Args:
+      score: Expected confidence score.
+      bad_values: First list of numbers.
+      good_values: Second list of numbers.
+    """
+    # ConfidenceScore takes a list of lists but these lists are flattened
+    # inside the function.
+    confidence = BisectResults.ConfidenceScore(
+        [[v] for v in bad_values],
+        [[v] for v in good_values])
+    self.assertEqual(score, confidence)
 
   def testConfidenceScoreIsZeroOnTooFewLists(self):
-    self.assertEqual(bisect_results.ConfidenceScore([], [[1], [2]]), 0.0)
-    self.assertEqual(bisect_results.ConfidenceScore([[1], [2]], []), 0.0)
-    self.assertEqual(bisect_results.ConfidenceScore([[1]], [[1], [2]]), 0.0)
-    self.assertEqual(bisect_results.ConfidenceScore([[1], [2]], [[1]]), 0.0)
+    self._AssertConfidence(0.0, [], [[1], [2]])
+    self._AssertConfidence(0.0, [[1], [2]], [])
+    self._AssertConfidence(0.0, [[1]], [[1], [2]])
+    self._AssertConfidence(0.0, [[1], [2]], [[1]])
 
   def testConfidenceScoreIsZeroOnEmptyLists(self):
-    self.assertEqual(bisect_results.ConfidenceScore([[], []], [[1], [2]]), 0.0)
-    self.assertEqual(bisect_results.ConfidenceScore([[1], [2]], [[], []]), 0.0)
-
-  def testConfidenceScoreIsUsingTTestWelchsTTest(self):
-    original_WelchsTTest = ttest.WelchsTTest
-    try:
-      ttest.WelchsTTest = lambda _sample1, _sample2: (0, 0, 0.42)
-      self.assertAlmostEqual(
-        bisect_results.ConfidenceScore([[1], [1]], [[2], [2]]), 58.0)
-    finally:
-      ttest.WelchsTTest = original_WelchsTTest
-
-
-class BisectResulstsTest(unittest.TestCase):
-  # TODO(sergiyb): Write tests for GetResultDicts when it is broken into smaller
-  # pieces.
-  pass
+    self.assertEqual(BisectResults.ConfidenceScore([[], []], [[1], [2]]), 0.0)
+    self.assertEqual(BisectResults.ConfidenceScore([[1], [2]], [[], []]), 0.0)
+
+  def testConfidenceScore_ZeroConfidence(self):
+    # The good and bad sets contain the same values, so the confidence that
+    # they're different should be zero.
+    self._AssertConfidence(0.0, [4, 5, 7, 6, 8, 7], [8, 7, 6, 7, 5, 4])
+
+  def testConfidenceScore_MediumConfidence(self):
+    self._AssertConfidence(80.0, [0, 1, 1, 1, 2, 2], [1, 1, 1, 3, 3, 4])
+
+  def testConfidenceScore_HighConfidence(self):
+    self._AssertConfidence(95.0, [0, 1, 1, 1, 2, 2], [1, 2, 2, 3, 3, 4])
+
+  def testConfidenceScore_VeryHighConfidence(self):
+    # Confidence is high if the two sets of values have no internal variance.
+    self._AssertConfidence(99.9, [1, 1, 1, 1], [1.2, 1.2, 1.2, 1.2])
+    self._AssertConfidence(99.9, [1, 1, 1, 1], [1.01, 1.01, 1.01, 1.01])
+
+  def testConfidenceScore_UnbalancedSampleSize(self):
+    # The second set of numbers only contains one number, so confidence is 0.
+    self._AssertConfidence(0.0, [1.1, 1.2, 1.1, 1.2, 1.0, 1.3, 1.2], [1.4])
+
+  def testConfidenceScore_EmptySample(self):
+    # Confidence is zero if either or both samples are empty.
+    self._AssertConfidence(0.0, [], [])
+    self._AssertConfidence(0.0, [], [1.1, 1.2, 1.1, 1.2, 1.0, 1.3, 1.2, 1.3])
+    self._AssertConfidence(0.0, [1.1, 1.2, 1.1, 1.2, 1.0, 1.3, 1.2, 1.3], [])
+
+  def testConfidenceScore_FunctionalTestResults(self):
+    self._AssertConfidence(80.0, [1, 1, 0, 1, 1, 1, 0, 1], [0, 0, 1, 0, 1, 0])
+    self._AssertConfidence(99.9, [1, 1, 1, 1, 1, 1, 1, 1], [0, 0, 0, 0, 0, 0])
+
+  def testConfidenceScore_RealWorldCases(self):
+    """This method contains a set of data from actual bisect results.
+
+    The confidence scores asserted below were all copied from the actual
+    results, so the purpose of this test method is mainly to show what the
+    results for real cases are, and compare when we change the confidence
+    score function in the future.
+    """
+    self._AssertConfidence(80, [133, 130, 132, 132, 130, 129], [129, 129, 125])
+    self._AssertConfidence(99.5, [668, 667], [498, 498, 499])
+    self._AssertConfidence(80, [67, 68], [65, 65, 67])
+    self._AssertConfidence(0, [514], [514])
+    self._AssertConfidence(90, [616, 613, 607, 615], [617, 619, 619, 617])
+    self._AssertConfidence(0, [3.5, 5.8, 4.7, 3.5, 3.6], [2.8])
+    self._AssertConfidence(90, [3, 3, 3], [2, 2, 2, 3])
+    self._AssertConfidence(0, [1999004, 1999627], [223355])
+    self._AssertConfidence(90, [1040, 934, 961], [876, 875, 789])
+    self._AssertConfidence(90, [309, 305, 304], [302, 302, 299, 303, 298])
+
+  def testCorrectlyFindsBreakingRange(self):
+    revision_states = self.mock_bisect_state.mock_revision_states
+    revision_states[0].passed = 0
+    revision_states[1].passed = 0
+    revision_states[2].passed = 1
+    revision_states[3].passed = 1
+    revision_states[4].passed = 1
+
+    results = BisectResults(self.mock_bisect_state, self.mock_depot_registry,
+                            self.mock_opts, self.mock_warnings)
+    self.assertEqual(revision_states[2], results.first_working_revision)
+    self.assertEqual(revision_states[1], results.last_broken_revision)
+
+  def testCorrectlyComputesRegressionStatistics(self):
+    revision_states = self.mock_bisect_state.mock_revision_states
+    revision_states[0].passed = 0
+    revision_states[0].value = {'values': [1000, 999, 998]}
+    revision_states[1].passed = 0
+    revision_states[1].value = {'values': [980, 1000, 999]}
+    revision_states[2].passed = 1
+    revision_states[2].value = {'values': [50, 45, 55]}
+    revision_states[3].passed = 1
+    revision_states[3].value = {'values': [45, 56, 45]}
+    revision_states[4].passed = 1
+    revision_states[4].value = {'values': [51, 41, 58]}
+
+    results = BisectResults(self.mock_bisect_state, self.mock_depot_registry,
+                            self.mock_opts, self.mock_warnings)
+    self.assertAlmostEqual(99.9, results.confidence)
+    self.assertAlmostEqual(1909.86547085, results.regression_size)
+    self.assertAlmostEqual(7.16625904, results.regression_std_err)
+
+  def testFindsCulpritRevisions(self):
+    revision_states = self.mock_bisect_state.mock_revision_states
+    revision_states[1].depot = 'chromium'
+    revision_states[2].depot = 'webkit'
+
+    results = BisectResults(self.mock_bisect_state, self.mock_depot_registry,
+                          self.mock_opts, self.mock_warnings)
+
+    self.assertEqual(1, len(results.culprit_revisions))
+    self.assertEqual(('b', {'test': 'b'}, 'chromium'),
+                     results.culprit_revisions[0])
+
+  def testFindsOtherRegressions(self):
+    revision_states = self.mock_bisect_state.mock_revision_states
+    revision_states[0].passed = 0
+    revision_states[0].value = {'values': [100, 100, 100]}
+    revision_states[1].passed = 0
+    revision_states[1].value = {'values': [100, 100, 100]}
+    revision_states[2].passed = 1
+    revision_states[2].value = {'values': [10, 10, 10]}
+    revision_states[3].passed = 1
+    revision_states[3].value = {'values': [100, 100, 100]}
+    revision_states[4].passed = 1
+    revision_states[4].value = {'values': [60, 60, 60]}
+
+    results = BisectResults(self.mock_bisect_state, self.mock_depot_registry,
+                            self.mock_opts, self.mock_warnings)
+    expected_regressions = [[revision_states[2], revision_states[1], 99.9],
+                            [revision_states[4], revision_states[3], 80.0]]
+    self.assertEqual(expected_regressions, results.other_regressions)
+
+  def testNoResultBasedWarningsForNormalState(self):
+    results = BisectResults(self.mock_bisect_state, self.mock_depot_registry,
+                            self.mock_opts, self.mock_warnings)
+    self.assertEqual(0, len(results.warnings))
+
+  def testWarningForMultipleCulpritRevisions(self):
+    self.mock_bisect_state.mock_revision_states[2].passed = 'Skipped'
+    results = BisectResults(self.mock_bisect_state, self.mock_depot_registry,
+                            self.mock_opts, self.mock_warnings)
+    self.assertEqual(1, len(results.warnings))
+
+  def testWarningForTooLowRetryLimit(self):
+    self.mock_opts.repeat_test_count = 1
+    results = BisectResults(self.mock_bisect_state, self.mock_depot_registry,
+                            self.mock_opts, self.mock_warnings)
+    self.assertEqual(1, len(results.warnings))
+
+  def testWarningForTooLowConfidence(self):
+    revision_states = self.mock_bisect_state.mock_revision_states
+    revision_states[2].value = {'values': [95, 100, 90]}
+    revision_states[3].value = {'values': [95, 100, 90]}
+    revision_states[4].value = {'values': [95, 100, 90]}
+    results = BisectResults(self.mock_bisect_state, self.mock_depot_registry,
+                            self.mock_opts, self.mock_warnings)
+    self.assertGreater(results.confidence, 0)
+    self.assertEqual(1, len(results.warnings))
+
+  def testWarningForZeroConfidence(self):
+    revision_states = self.mock_bisect_state.mock_revision_states
+    revision_states[2].value = {'values': [100, 105, 95]}
+    revision_states[3].value = {'values': [100, 105, 95]}
+    revision_states[4].value = {'values': [100, 105, 95]}
+    results = BisectResults(self.mock_bisect_state, self.mock_depot_registry,
+                            self.mock_opts, self.mock_warnings)
+    self.assertEqual(0, results.confidence)
+    self.assertEqual(1, len(results.warnings))
 
 
 if __name__ == '__main__':