def setUp(self):
self.eventLevelSlidingWindowStrategy = IBMPaperStrategy('TestData', subWindowIntervalDelta=timedelta(hours=5))
self.eventLevelSlidingWindowStrategy6 = IBMPaperStrategy('TestData', numberOfSubWindows=6, subWindowIntervalDelta=timedelta(hours=1))
# Parsed training data for five sub-window example
self.fiveSubWindowTrainingData = [
(
# Event level counts for sub-windows
(2, 0, 1, 0), (1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0),
# Total event level counts
3, 3, 2, 0,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(2, 0, 1, 0), (1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0)
),
# Event level count means
(0.75, 0.75, 0.50, 0.00),
# Event level count standard deviations
(0.83, 1.30, 0.50, 0.00),
# Number of intervals since last fatal event
INT_MAX,
True
),
(
# Event level counts for sub-windows
(1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1),
# Total event level counts
2, 3, 1, 1,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1)
),
# Event level count means
(0.50, 0.75, 0.25, 0.25),
# Event level count standard deviations
(0.50, 1.30, 0.43, 0.43),
# Number of intervals since last fatal event
1,
False
),
(
# Event level counts for sub-windows
(0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1), (0, 0, 0, 0),
# Total event level counts
1, 3, 0, 1,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1), (0, 0, 0, 0)
),
# Event level count means
(0.25, 0.75, 0.00, 0.25),
# Event level count standard deviations
(0.43, 1.30, 0.00, 0.43),
# Number of intervals since last fatal event
2,
False
)
]
# Parsed training data for six sub-window example
self.sixSubWindowTrainingData = [
(
# Event level counts for sub-windows
(2, 0, 1, 0), (0, 0, 0, 0), (1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0),
# Total event level counts
3, 3, 2, 0,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(1,0,0,0), (0,0,1,0), (1,0,0,0), (0,0,0,0), # Hours 1-4
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 5-8
(0,0,0,0), (0,0,0,0), (0,0,1,0), (1,0,0,0), # Hours 9-12
(0,3,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 13-16
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 17-20
),
# Event level count means
(0.15, 0.15, 0.10, 0.00),
# Event level count standard deviations
(0.36, 0.65, 0.30, 0.00),
# Number of intervals since last fatal event
INT_MAX,
True
),
(
# Event level counts for sub-windows
(0, 0, 0, 0), (1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1),
# Total event level counts
2, 3, 1, 1,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 5-8
(0,0,0,0), (0,0,0,0), (0,0,1,0), (1,0,0,0), # Hours 9-12
(0,3,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 13-16
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 17-20
(0,0,0,0), (0,0,0,1), (0,0,0,0), (1,0,0,0), # Hours 21-0 (next day)
),
# Event level count means
(0.10, 0.15, 0.05, 0.05),
# Event level count standard deviations
(0.30, 0.65, 0.22, 0.22),
# Number of intervals since last fatal event
1,
False
),
(
# Event level counts for sub-windows
(1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1), (0, 0, 0, 0),
# Total event level counts
2, 3, 1, 1,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(0,0,0,0), (0,0,0,0), (0,0,1,0), (1,0,0,0), # Hours 9-12
(0,3,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 13-16
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 17-20
(0,0,0,0), (0,0,0,1), (0,0,0,0), (1,0,0,0), # Hours 21-0 (next day)
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 1-4 (next day)
),
# Event level count means
(0.10, 0.15, 0.05, 0.05),
# Event level count standard deviations
(0.30, 0.65, 0.22, 0.22),
# Number of intervals since last fatal event
2,
False
)
]
self.projectRoot = os.environ['PROJECT_ROOT']
class IBMPaperStrategyTest(unittest.TestCase):
"""
Unit tests for the SlidingWindow class
"""
def setUp(self):
self.eventLevelSlidingWindowStrategy = IBMPaperStrategy('TestData', subWindowIntervalDelta=timedelta(hours=5))
self.eventLevelSlidingWindowStrategy6 = IBMPaperStrategy('TestData', numberOfSubWindows=6, subWindowIntervalDelta=timedelta(hours=1))
# Parsed training data for five sub-window example
self.fiveSubWindowTrainingData = [
(
# Event level counts for sub-windows
(2, 0, 1, 0), (1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0),
# Total event level counts
3, 3, 2, 0,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(2, 0, 1, 0), (1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0)
),
# Event level count means
(0.75, 0.75, 0.50, 0.00),
# Event level count standard deviations
(0.83, 1.30, 0.50, 0.00),
# Number of intervals since last fatal event
INT_MAX,
True
),
(
# Event level counts for sub-windows
(1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1),
# Total event level counts
2, 3, 1, 1,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1)
),
# Event level count means
(0.50, 0.75, 0.25, 0.25),
# Event level count standard deviations
(0.50, 1.30, 0.43, 0.43),
# Number of intervals since last fatal event
1,
False
),
(
# Event level counts for sub-windows
(0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1), (0, 0, 0, 0),
# Total event level counts
1, 3, 0, 1,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1), (0, 0, 0, 0)
),
# Event level count means
(0.25, 0.75, 0.00, 0.25),
# Event level count standard deviations
(0.43, 1.30, 0.00, 0.43),
# Number of intervals since last fatal event
2,
False
)
]
# Parsed training data for six sub-window example
self.sixSubWindowTrainingData = [
(
# Event level counts for sub-windows
(2, 0, 1, 0), (0, 0, 0, 0), (1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0),
# Total event level counts
3, 3, 2, 0,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(1,0,0,0), (0,0,1,0), (1,0,0,0), (0,0,0,0), # Hours 1-4
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 5-8
(0,0,0,0), (0,0,0,0), (0,0,1,0), (1,0,0,0), # Hours 9-12
(0,3,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 13-16
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 17-20
),
# Event level count means
(0.15, 0.15, 0.10, 0.00),
# Event level count standard deviations
(0.36, 0.65, 0.30, 0.00),
# Number of intervals since last fatal event
INT_MAX,
True
),
(
# Event level counts for sub-windows
(0, 0, 0, 0), (1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1),
# Total event level counts
2, 3, 1, 1,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 5-8
(0,0,0,0), (0,0,0,0), (0,0,1,0), (1,0,0,0), # Hours 9-12
(0,3,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 13-16
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 17-20
(0,0,0,0), (0,0,0,1), (0,0,0,0), (1,0,0,0), # Hours 21-0 (next day)
),
# Event level count means
(0.10, 0.15, 0.05, 0.05),
# Event level count standard deviations
(0.30, 0.65, 0.22, 0.22),
# Number of intervals since last fatal event
1,
False
),
(
# Event level counts for sub-windows
(1, 0, 1, 0), (0, 3, 0, 0), (0, 0, 0, 0), (1, 0, 0, 1), (0, 0, 0, 0),
# Total event level counts
2, 3, 1, 1,
# Event level counts for sub-window intervals (each sub-window is 1 hour interval
(
(0,0,0,0), (0,0,0,0), (0,0,1,0), (1,0,0,0), # Hours 9-12
(0,3,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 13-16
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 17-20
(0,0,0,0), (0,0,0,1), (0,0,0,0), (1,0,0,0), # Hours 21-0 (next day)
(0,0,0,0), (0,0,0,0), (0,0,0,0), (0,0,0,0), # Hours 1-4 (next day)
),
# Event level count means
(0.10, 0.15, 0.05, 0.05),
# Event level count standard deviations
(0.30, 0.65, 0.22, 0.22),
# Number of intervals since last fatal event
2,
False
)
]
self.projectRoot = os.environ['PROJECT_ROOT']
def testParseTrainingDataEmpty(self):
"""
Tests that no training data is returned when passed an empty log
"""
actualTrainingData = self.eventLevelSlidingWindowStrategy.parseWindowedLogData([], [])
self.assertEqual([], actualTrainingData)
def testParseTrainingDataNone(self):
"""
Tests that no training data is returned when passed a null log
"""
actualTrainingData = self.eventLevelSlidingWindowStrategy.parseWindowedLogData(None, None)
self.assertEqual([], actualTrainingData)
def testParseTrainingDataUnevenSplit(self):
"""
Tests that the training data fails to parse if the sub-window intervals do not divide evenly into the sub-windows
"""
try:
invalidStrategy = IBMPaperStrategy('TestData', subWindowIntervalDelta=timedelta(hours=4))
self.fail('Trying to construct an invalid strategy should have thrown an exception!')
except StrategyError, error:
self.assertEqual('Error parsing windowed log data, cannot divide sub-windows into smaller intervals evenly!', error.message)
