def testRaisesCustomExceptionOnMissingMandatoryFields(self):
with self.assertRaises(ValueError):
# Missing name.
histogram_proto_converter.ConvertHistogram({})
with self.assertRaises(ValueError):
# Missing unit.
histogram_proto_converter.ConvertHistogram({'name': 'eh'})
def testComplexBinBounds(self):
hist_dict = histogram_proto_converter.ConvertHistogram(
_Hist(
bin_boundaries={
'firstBinBoundary':
17,
'binSpecs': [{
'binBoundary': 18
}, {
'binSpec': {
'boundaryType': 'EXPONENTIAL',
'maximumBinBoundary': 19,
'numBinBoundaries': 20
},
}, {
'binSpec': {
'boundaryType': 'LINEAR',
'maximumBinBoundary': 21,
'numBinBoundaries': 22
}
}]
}))
# Don't parse this into a histogram for this case. The format for bins is
# relatively easily understood, whereas how bins are generated is very
# complex. See the histogram spec in docs/histogram-set-json-format.md.
bins = hist_dict['binBoundaries']
self.assertEqual(bins, [17, 18, [1, 19, 20], [0, 21, 22]])
def testMinimalBinBounds(self):
hist_dict = histogram_proto_converter.ConvertHistogram(
_Hist(bin_boundaries={
'firstBinBoundary': 1,
}))
bins = hist_dict['binBoundaries']
self.assertEqual(bins, [1])
def testUnitWithImprovementSmallerIsBetter(self):
hist_dict = histogram_proto_converter.ConvertHistogram(
_Hist(unit={
'unit': 'TS_MS',
'improvement_direction': 'SMALLER_IS_BETTER'
}))
hist = histogram.Histogram.FromDict(hist_dict)
self.assertEqual(hist.unit, 'tsMs_smallerIsBetter')
def testUnitWithImprovementBiggerIsBetter(self):
hist_dict = histogram_proto_converter.ConvertHistogram(
_Hist(unit={
'unit': 'SIGMA',
'improvement_direction': 'BIGGER_IS_BETTER'
}))
hist = histogram.Histogram.FromDict(hist_dict)
self.assertEqual(hist.unit, 'sigma_biggerIsBetter')
def testUnsupportedDiagnostics(self):
with self.assertRaises(ValueError):
histogram_proto_converter.ConvertHistogram(
_Hist(diagnostics={
'diagnosticMap': {
'myDiagnostic': {
'breakdown': {}
},
}
}))
def testSharedDiagnosticsAreMappedIntoHistogram(self):
guid1 = 'f7f17394-fa4a-481e-86bd-a82cd55935a7'
guid2 = '88ea36c7-6dcb-4ba8-ba56-1979de05e16f'
hist_dict = histogram_proto_converter.ConvertHistogram(
_Hist(
diagnostics={
'diagnosticMap': {
'bots': {
'sharedDiagnosticGuid': guid1,
},
'pointId': {
'sharedDiagnosticGuid': guid2,
},
}
}))
shared_diagnostics = histogram_proto_converter.ConvertSharedDiagnostics(
{
guid1: {
'genericSet': {
'values': [
'"webrtc_perf_tests"',
]
},
},
guid2: {
'genericSet': {
'values': [
'123456',
]
},
},
'some other guid': {
'genericSet': {
'values': ['2']
}
}
})
# Import into a HistogramSet since it does the resolving of shared
# diagnostics.
hist_set = histogram_set.HistogramSet()
for shared in shared_diagnostics:
hist_set.ImportLegacyDict(shared)
hist_set.ImportLegacyDict(hist_dict)
hist = hist_set.GetFirstHistogram()
self.assertIsNotNone(hist)
self.assertEqual(len(hist.diagnostics), 2)
self.assertEqual(hist.diagnostics['pointId'],
generic_set.GenericSet(values=[123456]))
self.assertEqual(hist.diagnostics['bots'],
generic_set.GenericSet(values=['webrtc_perf_tests']))
def testUnitWithImprovementDontCare(self):
# Don't care should not really be set - protobuf doesn't write it out in the
# JSON because it's the base enum.
hist_dict = histogram_proto_converter.ConvertHistogram(
_Hist(unit={
'unit': 'HERTZ',
'improvement_direction': 'NOT_SPECIFIED'
}))
hist = histogram.Histogram.FromDict(hist_dict)
self.assertEqual(hist.unit, 'Hz')
def ImportProtoDict(self, dicts):
shared_diagnostics = dicts.get('sharedDiagnostics')
if shared_diagnostics:
convert = histogram_proto_converter.ConvertSharedDiagnostics
converted = convert(shared_diagnostics)
for d in converted:
self.ImportLegacyDict(d)
for histogram_dict in dicts["histograms"]:
hist = histogram_proto_converter.ConvertHistogram(histogram_dict)
self.ImportLegacyDict(hist)
def testMinimalValidHistogram(self):
hist_dict = histogram_proto_converter.ConvertHistogram({
'name': 'name!',
'unit': {
'unit': 'UNITLESS'
},
})
hist = histogram.Histogram.FromDict(hist_dict)
self.assertEqual(hist.name, 'name!')
self.assertEqual(hist.unit, 'unitless')
def testMinimalStats(self):
hist_dict = histogram_proto_converter.ConvertHistogram(
_Hist(
running={
# The proto will not write ints that are 0 on the wire.
'count': 1
}))
hist = histogram.Histogram.FromDict(hist_dict)
self.assertEqual(hist.running.count, 1)
self.assertEqual(hist.running.mean, 0)
self.assertEqual(hist.running.variance, 0)
def testGenericSetWithInvalidJson(self):
with self.assertRaises(TypeError):
histogram_proto_converter.ConvertHistogram(
_Hist(
diagnostics={
'diagnosticMap': {
'myDiagnostic': {
'genericSet': {
'values':
['this_is_an_undefined_json_indentifier']
}
},
}
}))
def testBasicDiagnostics(self):
hist_dict = histogram_proto_converter.ConvertHistogram(
_Hist(
diagnostics={
'diagnosticMap': {
'myDiagnostic': {
'genericSet': {
'values': ["\"some value\""]
}
},
}
}))
hist = histogram.Histogram.FromDict(hist_dict)
expected = generic_set.GenericSet(values=['some value'])
self.assertEqual(hist.diagnostics['myDiagnostic'], expected)
def testSummaryOptions(self):
hist_dict = histogram_proto_converter.ConvertHistogram(
_Hist(
summary_options={
'nans': False,
'geometricMean': False,
'percentile': [0.90, 0.95, 0.99]
}))
hist = histogram.Histogram.FromDict(hist_dict)
# See the histogram spec in docs/histogram-set-json-format.md.
self.assertEqual(
hist.statistics_names,
set([
'std', 'count', 'pct_090', 'pct_095', 'max', 'sum', 'min',
'pct_099', 'avg'
]))
def testSimpleFields(self):
hist_dict = histogram_proto_converter.ConvertHistogram({
'name':
'whatever',
'unit': {
'unit': 'MS'
},
'description':
'description!',
'sampleValues': [21, 22, 23],
'maxNumSampleValues':
3,
'numNans':
1,
})
hist = histogram.Histogram.FromDict(hist_dict)
self.assertEqual(hist.description, 'description!')
self.assertEqual(hist.sample_values, [21, 22, 23])
self.assertEqual(hist.max_num_sample_values, 3)
self.assertEqual(hist.num_nans, 1)
def testAllBins(self):
hist_dict = histogram_proto_converter.ConvertHistogram(
_Hist(
all_bins={
'0': {
'binCount':
24,
'diagnosticMaps': [{
'diagnosticMap': {
'some bin diagnostic': {
'genericSet': {
'values': ["\"some value\""]
}
}
}
}, {
'diagnosticMap': {
'other bin diagnostic': {
'genericSet': {
'values': ["\"some other value\""]
}
}
}
}]
}
}))
hist = histogram.Histogram.FromDict(hist_dict)
self.assertEqual(len(hist.bins), 1)
self.assertEqual(len(hist.bins[0].diagnostic_maps), 2)
self.assertEqual(len(hist.bins[0].diagnostic_maps[0]), 1)
self.assertEqual(len(hist.bins[0].diagnostic_maps[1]), 1)
self.assertEqual(
hist.bins[0].diagnostic_maps[0]['some bin diagnostic'],
generic_set.GenericSet(values=['some value']))
self.assertEqual(
hist.bins[0].diagnostic_maps[1]['other bin diagnostic'],
generic_set.GenericSet(values=['some other value']))
def testRunningStatistics(self):
hist_dict = histogram_proto_converter.ConvertHistogram(
_Hist(
running={
'count': 4,
'max': 23,
'meanlogs': 1,
'mean': 22,
'min': 21,
'sum': 66,
'variance': 1
}))
hist = histogram.Histogram.FromDict(hist_dict)
# We get at meanlogs through geometric_mean. Variance is after Bessel's
# correction has been applied.
self.assertEqual(hist.running.count, 4)
self.assertEqual(hist.running.max, 23)
self.assertEqual(hist.running.geometric_mean, math.exp(1))
self.assertEqual(hist.running.mean, 22)
self.assertEqual(hist.running.min, 21)
self.assertEqual(hist.running.sum, 66)
self.assertAlmostEqual(hist.running.variance, 0.3333333333)
