def testA_BasicTest(self):
"""
_testA_BasicTest_
Build a histogram from a set of uniformly
distributed pseudorandom numbers. Check
that the statistic properties
in the histogram are accurate to some degree,
that the histogram binning is done right and
that this can become a document an uploaded to couch
"""
inputData = self.buildRandomNumberList(1000)
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel',
'SomeoneElsesLabel')
# Populate the histogram
for point in inputData:
histogram.addPoint(point)
# Get the JSON
jsonHistogram = histogram.toJSON()
# Check the histogram core data
self.assertEqual(jsonHistogram["title"], "TestHisto")
self.assertEqual(jsonHistogram["xLabel"], "MyLabel")
self.assertAlmostEqual(jsonHistogram["average"], 0.0, places=0)
self.assertAlmostEqual(jsonHistogram["stdDev"], 1.0, places=0)
self.assertEqual(len(jsonHistogram["data"]), 16)
self.assertTrue(jsonHistogram["continuous"])
# Check the internal data
self.assertEqual(jsonHistogram["internalData"]["yLabel"],
"SomeoneElsesLabel")
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 1000)
# Try to commit it to couch
jsonHistogram["_id"] = jsonHistogram["title"]
self.histogramDB.commitOne(jsonHistogram)
storedJSON = self.histogramDB.document("TestHisto")
self.assertEqual(len(storedJSON["data"]), 16)
return
def testC_compactHistogram(self):
"""
_testC_compactHistogram_
Check that we can create smaller histograms objects
by chopping outliers and dropping the data all together
"""
# Input normally distributed data and chop anything above 1 stdev (32% of data)
histogram = ContinuousSummaryHistogram('TestHisto',
'MyLabel',
'SomeoneElsesLabel',
dropOutliers=True,
sigmaLimit=1)
inputData = self.buildRandomNumberList(1000)
for point in inputData:
histogram.addPoint(point)
jsonHistogram = histogram.toJSON()
self.assertAlmostEqual(jsonHistogram["average"], 0.0, places=0)
self.assertAlmostEqual(jsonHistogram["stdDev"], 1.0, places=0)
self.assertEqual(len(jsonHistogram["data"]), 16)
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 1000)
pointsInHistogram = sum([x for x in jsonHistogram["data"].values()])
# With high probability we must have chopped at least one point
self.assertTrue(pointsInHistogram < 1000)
self.assertAlmostEqual(pointsInHistogram / 1000.0, 0.68, places=1)
# Create a histogram without histogram data
histogram = ContinuousSummaryHistogram('TestHisto',
'MyLabel',
'SomeoneElsesLabel',
storeHistogram=False)
inputData = self.buildRandomNumberList(1000)
for point in inputData:
histogram.addPoint(point)
jsonHistogram = histogram.toJSON()
self.assertAlmostEqual(jsonHistogram["average"], 0.0, places=0)
self.assertAlmostEqual(jsonHistogram["stdDev"], 1.0, places=0)
self.assertEqual(len(jsonHistogram["data"]), 0)
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 1000)
return
def testA_BasicTest(self):
"""
_testA_BasicTest_
Build a histogram from a set of uniformly
distributed pseudorandom numbers. Check
that the statistic properties
in the histogram are accurate to some degree,
that the histogram binning is done right and
that this can become a document an uploaded to couch
"""
inputData = self.buildRandomNumberList(1000)
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel', 'SomeoneElsesLabel')
# Populate the histogram
for point in inputData:
histogram.addPoint(point)
# Get the JSON
jsonHistogram = histogram.toJSON()
# Check the histogram core data
self.assertEqual(jsonHistogram["title"], "TestHisto")
self.assertEqual(jsonHistogram["xLabel"], "MyLabel")
self.assertAlmostEqual(jsonHistogram["average"], 0.0, places = 0)
self.assertAlmostEqual(jsonHistogram["stdDev"], 1.0, places = 0)
self.assertEqual(len(jsonHistogram["data"]), 16)
self.assertTrue(jsonHistogram["continuous"])
# Check the internal data
self.assertEqual(jsonHistogram["internalData"]["yLabel"], "SomeoneElsesLabel")
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 1000)
# Try to commit it to couch
jsonHistogram["_id"] = jsonHistogram["title"]
self.histogramDB.commitOne(jsonHistogram)
storedJSON = self.histogramDB.document("TestHisto")
self.assertEqual(len(storedJSON["data"]), 16)
return
def testC_compactHistogram(self):
"""
_testC_compactHistogram_
Check that we can create smaller histograms objects
by chopping outliers and dropping the data all together
"""
# Input normally distributed data and chop anything above 1 stdev (32% of data)
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel', 'SomeoneElsesLabel',
dropOutliers = True, sigmaLimit = 1)
inputData = self.buildRandomNumberList(1000)
for point in inputData:
histogram.addPoint(point)
jsonHistogram = histogram.toJSON()
self.assertAlmostEqual(jsonHistogram["average"], 0.0, places = 0)
self.assertAlmostEqual(jsonHistogram["stdDev"], 1.0, places = 0)
self.assertEqual(len(jsonHistogram["data"]), 16)
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 1000)
pointsInHistogram = sum([x for x in jsonHistogram["data"].values()])
# With high probability we must have chopped at least one point
self.assertTrue(pointsInHistogram < 1000)
self.assertAlmostEqual(pointsInHistogram / 1000.0, 0.68, places = 1)
# Create a histogram without histogram data
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel', 'SomeoneElsesLabel',
storeHistogram = False)
inputData = self.buildRandomNumberList(1000)
for point in inputData:
histogram.addPoint(point)
jsonHistogram = histogram.toJSON()
self.assertAlmostEqual(jsonHistogram["average"], 0.0, places = 0)
self.assertAlmostEqual(jsonHistogram["stdDev"], 1.0, places = 0)
self.assertEqual(len(jsonHistogram["data"]), 0)
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 1000)
return
def testB_extremeData(self):
"""
_testB_extremeData_
Put extreme points in the data and try to build a histogram.
Check that it can process all this correctly
"""
# First no data
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel', 'SomeoneElsesLabel')
jsonHistogram = histogram.toJSON()
self.assertEqual(jsonHistogram["title"], "TestHisto")
self.assertEqual(jsonHistogram["xLabel"], "MyLabel")
self.assertEqual(jsonHistogram["average"], 0.0)
self.assertEqual(jsonHistogram["stdDev"], 0.0)
self.assertEqual(len(jsonHistogram["data"]), 0)
# Data with NaNs and Infs
inputData = self.buildRandomNumberList(100)
inputData.append(float('NaN'))
inputData.append(float('Inf'))
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel', 'SomeoneElsesLabel')
for point in inputData:
histogram.addPoint(point)
jsonHistogram = histogram.toJSON()
self.assertAlmostEqual(jsonHistogram["average"], 0.0, places = 0)
self.assertAlmostEqual(jsonHistogram["stdDev"], 1.0, places = 0)
self.assertEqual(len(jsonHistogram["data"]), 7)
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 100)
# One single point, P5
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel', 'SomeoneElsesLabel')
histogram.addPoint(5)
jsonHistogram = histogram.toJSON()
self.assertEqual(jsonHistogram["average"], 5.0)
self.assertEqual(jsonHistogram["stdDev"], 0.0)
self.assertEqual(len(jsonHistogram["data"]), 1)
self.assertEqual(jsonHistogram["data"]["5.0,5.0"], 1)
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 1)
# Test that toJSON is idempotent
inputData = self.buildRandomNumberList(100)
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel', 'SomeoneElsesLabel')
for point in inputData:
histogram.addPoint(point)
jsonHistogram = histogram.toJSON()
oldData = jsonHistogram["data"]
jsonHistogram = histogram.toJSON()
self.assertAlmostEqual(jsonHistogram["average"], 0.0, places = 0)
self.assertAlmostEqual(jsonHistogram["stdDev"], 1.0, places = 0)
self.assertEqual(len(jsonHistogram["data"]), 7)
self.assertEqual(jsonHistogram["data"], oldData)
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 100)
return
def testB_extremeData(self):
"""
_testB_extremeData_
Put extreme points in the data and try to build a histogram.
Check that it can process all this correctly
"""
# First no data
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel', 'SomeoneElsesLabel')
jsonHistogram = histogram.toJSON()
self.assertEqual(jsonHistogram["title"], "TestHisto")
self.assertEqual(jsonHistogram["xLabel"], "MyLabel")
self.assertEqual(jsonHistogram["average"], 0.0)
self.assertEqual(jsonHistogram["stdDev"], 0.0)
self.assertEqual(len(jsonHistogram["data"]), 0)
# Data with NaNs and Infs
inputData = self.buildRandomNumberList(100)
inputData.append(float('NaN'))
inputData.append(float('Inf'))
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel', 'SomeoneElsesLabel')
for point in inputData:
histogram.addPoint(point)
jsonHistogram = histogram.toJSON()
self.assertAlmostEqual(jsonHistogram["average"], 0.0, places = 0)
self.assertAlmostEqual(jsonHistogram["stdDev"], 1.0, places = 0)
self.assertEqual(len(jsonHistogram["data"]), 7)
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 100)
# One single point, P5
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel', 'SomeoneElsesLabel')
histogram.addPoint(5)
jsonHistogram = histogram.toJSON()
self.assertEqual(jsonHistogram["average"], 5.0)
self.assertEqual(jsonHistogram["stdDev"], 0.0)
self.assertEqual(len(jsonHistogram["data"]), 1)
self.assertEqual(jsonHistogram["data"]["5.0,5.0"], 1)
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 1)
# Test that toJSON is idempotent
inputData = self.buildRandomNumberList(100)
histogram = ContinuousSummaryHistogram('TestHisto', 'MyLabel', 'SomeoneElsesLabel')
for point in inputData:
histogram.addPoint(point)
jsonHistogram = histogram.toJSON()
oldData = jsonHistogram["data"]
jsonHistogram = histogram.toJSON()
self.assertAlmostEqual(jsonHistogram["average"], 0.0, places = 0)
self.assertAlmostEqual(jsonHistogram["stdDev"], 1.0, places = 0)
self.assertEqual(len(jsonHistogram["data"]), 7)
self.assertEqual(jsonHistogram["data"], oldData)
self.assertEqual(jsonHistogram["internalData"]["nPoints"], 100)
return
