def test_profiled_quantiles(self):
"""
Checks the quantiles of profiled numerical columns.
:return:
"""
# this data has 4 bins, range of 3
# with equal bin size, each bin has the width of 0.75
data = ["1.0", "2.0", "3.0", "4.0"]
df = pd.Series(data)
profiler = FloatColumn(df.name)
profiler.update(df)
profile = profiler.profile
est_quantiles = profile['quantiles']
est_Q1 = est_quantiles[249]
est_Q2 = est_quantiles[499]
est_Q3 = est_quantiles[749]
self.assertEqual(999, len(est_quantiles))
self.assertAlmostEqual(1.000012, est_quantiles[0])
self.assertEqual(est_Q1, 1.003)
self.assertEqual(est_Q2, 2.5)
self.assertEqual(est_Q3, 3.001)
self.assertAlmostEqual(3.999988, est_quantiles[-1])
def test_profiled_quantiles(self):
"""
Checks the quantiles of profiled numerical columns.
:return:
"""
# this data has 4 bins, range of 3
# with equal bin size, each bin has the width of 0.75
data = ["1.0", "2.0", "3.0", "4.0"]
df = pd.Series(data)
profiler = FloatColumn(df.name)
profiler.update(df)
profile = profiler.profile
est_quartiles = profile['quantiles']
est_Q1 = est_quartiles[249]
est_Q2 = est_quartiles[499]
est_Q3 = est_quartiles[749]
data_to_num = [float(item) for item in data]
exact_Q1 = np.percentile(data_to_num, 25)
exact_Q2 = np.percentile(data_to_num, 50)
exact_Q3 = np.percentile(data_to_num, 75)
self.assertEqual(est_Q1, exact_Q1)
self.assertEqual(est_Q2, exact_Q2)
self.assertEqual(est_Q3, exact_Q3)
def test_profile_histogram_w_updates(self):
"""
Checks if histogram properly resets the _profiled histogram after
merge or update.
:return:
"""
list_data_test = []
# this data has 4 bins, range of 3
# with equal bin size, each bin has the width of 0.75
df1 = pd.Series(["1.0", "2.0", "3.0", "4.0"])
expected_histogram1 = {
'bin_counts': np.array([1, 1, 1, 1]),
'bin_edges': np.array([1.0, 1.75, 2.5, 3.25, 4.0]),
}
list_data_test.append([df1, expected_histogram1])
# this data will be the second update of the profile.
# this results in the combination of the previous data and this data.
# the range should update to 12 from 3.
df2 = pd.Series(["1.0", "5.0", "8.0", "13.0"])
expected_histogram2 = {
'bin_counts': np.array([4, 1, 1, 1, 0, 1]),
'bin_edges': np.array([1.0, 3.0, 5.0, 7.0, 9.0, 11.0, 13.0]),
}
list_data_test.append([df2, expected_histogram2])
profiler = FloatColumn("test")
for i, (df, expected_histogram) in enumerate(list_data_test):
profiler.update(df)
self.assertIsNone(profiler.histogram_selection)
profile = profiler.profile
self.assertIsNotNone(profiler.histogram_selection)
histogram = profile['histogram']
self.assertEqual(expected_histogram['bin_counts'].tolist(),
histogram['bin_counts'].tolist())
self.assertCountEqual(
np.round(expected_histogram['bin_edges'], 12),
np.round(histogram['bin_edges'], 12))
# apply test to merging profiles
expected_histogram = {
'bin_edges':
np.array([1., 19 / 7, 31 / 7, 43 / 7, 55 / 7, 67 / 7, 79 / 7,
13.]),
'bin_counts':
np.array([6, 4, 2, 0, 2, 0, 2])
}
merged_profiler = profiler + profiler
self.assertIsNone(merged_profiler.histogram_selection)
profile = merged_profiler.profile
self.assertIsNotNone(merged_profiler.histogram_selection)
histogram = profile['histogram']
self.assertEqual(expected_histogram['bin_counts'].tolist(),
histogram['bin_counts'].tolist())
self.assertCountEqual(np.round(expected_histogram['bin_edges'], 12),
np.round(histogram['bin_edges'], 12))
def test_custom_bin_count_merge(self):
options = FloatOptions()
options.histogram_and_quantiles.bin_count_or_method = 10
data = [2.0, 'not a float', 6.0, 'not a float']
df = pd.Series(data).apply(str)
profiler1 = FloatColumn("Float", options)
profiler1.update(df)
data2 = [10.0, 'not a float', 15.0, 'not a float']
df2 = pd.Series(data2).apply(str)
profiler2 = FloatColumn("Float", options)
profiler2.update(df2)
# no warning should occur
with warnings.catch_warnings(record=True) as w:
merge_profile = profiler1 + profiler2
self.assertListEqual([], w)
self.assertEqual(10, merge_profile.user_set_histogram_bin)
# make bin counts different and get warning
profiler2.user_set_histogram_bin = 120
with self.assertWarnsRegex(
UserWarning, 'User set histogram bin counts did not '
'match. Choosing the larger bin count.'):
merged_profile = profiler1 + profiler2
self.assertEqual(120, merged_profile.user_set_histogram_bin)
def test_base_case(self):
data = pd.Series([], dtype=object)
profiler = FloatColumn(data.name)
profiler.update(data)
self.assertEqual(profiler.match_count, 0)
self.assertEqual(profiler.min, None)
self.assertEqual(profiler.max, None)
self.assertEqual(profiler.mean, 0)
self.assertEqual(profiler.variance, 0)
self.assertTrue(profiler.stddev is np.nan)
self.assertIsNone(profiler.histogram_selection)
self.assertEqual(len(profiler.quantiles), 1000)
self.assertIsNone(profiler.data_type_ratio)
def test_null_values_for_histogram(self):
data = pd.Series(['-inf', 'inf'])
profiler = FloatColumn(data.name)
profiler.update(data)
profile = profiler.profile
histogram = profile['histogram']
self.assertEqual(histogram['bin_counts'], None)
self.assertEqual(histogram['bin_edges'], None)
data = pd.Series(['-2', '-1', '1', '2', '-inf', 'inf'])
profiler = FloatColumn(data.name)
profiler.update(data)
profile = profiler.profile
histogram = profile['histogram']
expected_histogram = {
'bin_counts': np.array([1, 1, 0, 2]),
'bin_edges': np.array([-2., -1., 0., 1., 2.]),
}
self.assertCountEqual(histogram['bin_counts'],
expected_histogram['bin_counts'])
self.assertCountEqual(histogram['bin_edges'],
expected_histogram['bin_edges'])
def test_histogram_loss(self):
# run time is small
diff_var, avg_diffvar, total_var, avg_totalvar, run_time, avg_runtime =\
0.3, 0.2, 0.1, 0.05, 0.0014, 0.0022
expected_loss = 0.1 / 0.2 + 0.05 / 0.05
est_loss = FloatColumn._histogram_loss(diff_var, avg_diffvar,
total_var, avg_totalvar,
run_time, avg_runtime)
self.assertEqual(expected_loss, est_loss)
# run time is big
diff_var, avg_diffvar, total_var, avg_totalvar, run_time, avg_runtime =\
0.3, 0.2, 0.1, 0.05, 22, 14
expected_loss = 0.1 / 0.2 + 0.05 / 0.05 + 8 / 14
est_loss = FloatColumn._histogram_loss(diff_var, avg_diffvar,
total_var, avg_totalvar,
run_time, avg_runtime)
self.assertEqual(expected_loss, est_loss)
def test_profile_merge_no_bin_overlap(self):
data = [2.0, 'not a float', 6.0, 'not a float']
df = pd.Series(data).apply(str)
profiler1 = FloatColumn("Float")
profiler1.update(df)
data2 = [10.0, 'not a float', 15.0, 'not a float']
df2 = pd.Series(data2).apply(str)
profiler2 = FloatColumn("Float")
profiler2.update(df2)
# set bin names so no overlap
profiler1.histogram_bin_method_names = ['No overlap 1']
profiler2.histogram_bin_method_names = ['No overlap 2']
with self.assertRaisesRegex(
ValueError, 'Profiles have no overlapping bin methods '
'and therefore cannot be added together.'):
profiler1 + profiler2
def test_option_timing(self):
data = [2.0, 12.5, 'not a float', 6.0, 'not a float']
df = pd.Series(data).apply(str)
options = FloatOptions()
options.set({"min.is_enabled": False})
profiler = FloatColumn(df.name, options=options)
time_array = [float(i) for i in range(100, 0, -1)]
with mock.patch('time.time', side_effect=lambda: time_array.pop()):
# Validate that the times dictionary is empty
self.assertEqual(defaultdict(float), profiler.profile['times'])
profiler.update(df)
# Validate the time in the datetime class has the expected time.
profile = profiler.profile
expected = defaultdict(float, {'max': 1.0, 'sum': 1.0,\
'variance': 1.0, 'precision': 1.0,\
'histogram_and_quantiles': 15.0})
self.assertEqual(expected, profile['times'])
# Validate time in datetime class has expected time after second update
profiler.update(df)
expected = defaultdict(float, {'max': 2.0, 'sum': 2.0,\
'variance': 2.0, 'precision': 2.0,\
'histogram_and_quantiles': 30.0})
self.assertEqual(expected, profiler.profile['times'])
def test_merge_histogram(self):
data = pd.Series([], dtype=object)
profiler = FloatColumn(data.name)
profiler.update(data)
profiler._stored_histogram['histogram']['bin_counts'] = np.array(
[3, 2])
profiler._stored_histogram['histogram']['bin_edges'] = \
np.array([1.0, 3.0, 5.0])
input_array = [0.5, 1.0, 2.0, 5.0]
profiler._merge_histogram(input_array)
merged_hist = profiler._histogram_for_profile('sqrt')[0]
expected_bin_counts, expected_bin_edges = \
[5, 2, 2], [0.5, 2.0, 3.5, 5.0]
self.assertEqual(expected_bin_counts,
merged_hist['bin_counts'].tolist())
self.assertCountEqual(expected_bin_edges, merged_hist['bin_edges'])
def test_histogram_to_array(self):
data = pd.Series([], dtype=object)
profiler = FloatColumn(data.name)
profiler.update(data)
profiler._stored_histogram['histogram']['bin_counts'] = \
np.array([3, 2, 1])
profiler._stored_histogram['histogram']['bin_edges'] = \
np.array([1.0, 3.0, 5.0, 7.0])
array_from_histogram = profiler._histogram_to_array()
expected_array = [1.0, 1.0, 1.0, 3.0, 3.0, 7.0]
self.assertEqual(expected_array, array_from_histogram.tolist())
def test_histogram_to_array(self):
data = pd.Series([], dtype=object)
profiler = FloatColumn(data.name)
profiler.update(data)
profiler.histogram_methods['auto']['histogram']['bin_counts'] = \
np.array([3, 2, 1])
profiler.histogram_methods['auto']['histogram']['bin_edges'] = \
np.array([1.0, 3.0, 5.0, 7.0])
array_from_histogram = profiler._histogram_to_array('auto')
expected_array = [1.0, 1.0, 1.0, 3.0, 3.0, 7.0]
self.assertCountEqual(array_from_histogram, expected_array)
def test_total_histogram_bin_variance(self):
data = pd.Series([], dtype=object)
profiler = FloatColumn(data.name)
profiler.update(data)
profiler._stored_histogram['histogram']['bin_counts'] = \
np.array([3, 2, 1])
profiler._stored_histogram['histogram']['bin_edges'] = \
np.array([1.0, 3.0, 5.0, 7.0])
input_array = np.array([1.1, 1.5, 2.3, 3.5, 4.0, 6.5])
expected_total_var = np.array([1.1, 1.5, 2.3]).var() \
+ np.array([3.5, 4.0]).var() \
+ np.array([6.5]).var()
est_total_var = profiler._total_histogram_bin_variance(input_array)
self.assertEqual(expected_total_var, est_total_var)
def test_data_type_ratio(self):
data = np.linspace(-5, 5, 4)
df = pd.Series(data).apply(str)
profiler = FloatColumn(df.name)
profiler.update(df)
self.assertEqual(profiler.data_type_ratio, 1.0)
df = pd.Series(['not a float'])
profiler.update(df)
self.assertEqual(profiler.data_type_ratio, 0.8)
def test_estimate_stats_from_histogram(self):
data = pd.Series([], dtype=object)
profiler = FloatColumn(data.name)
profiler.update(data)
profiler._stored_histogram['histogram']['bin_counts'] = \
np.array([1, 2, 1])
profiler._stored_histogram['histogram']['bin_edges'] = \
np.array([1.0, 3.0, 5.0, 7.0])
expected_mean = (2.0 * 1 + 4.0 * 2 + 6.0 * 1) / 4
expected_var = (1 * (2.0 - expected_mean)**2 + 2 *
(4.0 - expected_mean)**2 + 1 *
(6.0 - expected_mean)**2) / 4
expected_std = np.sqrt(expected_var)
est_var = profiler._estimate_stats_from_histogram()
self.assertEqual(expected_var, est_var)
def test_merge_histogram(self):
data = pd.Series([], dtype=object)
profiler = FloatColumn(data.name)
profiler.update(data)
profiler.histogram_methods['sqrt']['histogram']['bin_counts'] = \
np.array([3, 2])
profiler.histogram_methods['sqrt']['histogram']['bin_edges'] = \
np.array([1.0, 3.0, 5.0])
input_array = [0.5, 1.0, 2.0, 5.0]
profiler._merge_histogram(input_array, 'sqrt')
merged_bin_counts = \
profiler.histogram_methods['sqrt']['histogram']['bin_counts']
merged_bin_edges = \
profiler.histogram_methods['sqrt']['histogram']['bin_edges']
expected_bin_counts, expected_bin_edges = \
[5, 2, 2], [0.5, 2.0, 3.5, 5.0]
self.assertCountEqual(merged_bin_counts, expected_bin_counts)
self.assertCountEqual(merged_bin_edges, expected_bin_edges)
def test_profile_merge_with_different_options(self):
# Creating first profiler with default options
options = FloatOptions()
options.max.is_enabled = False
options.min.is_enabled = False
data = [2, 4, 6, 8]
df = pd.Series(data).apply(str)
profiler1 = FloatColumn("Float", options=options)
profiler1.update(df)
# Creating second profiler with separate options
options = FloatOptions()
options.min.is_enabled = False
options.precision.is_enabled = False
data2 = [10, 15]
df2 = pd.Series(data2).apply(str)
profiler2 = FloatColumn("Float", options=options)
profiler2.update(df2)
# Asserting warning when adding 2 profilers with different options
with warnings.catch_warnings(record=True) as w:
profiler3 = profiler1 + profiler2
list_of_warning_messages = []
for warning in w:
list_of_warning_messages.append(str(warning.message))
warning1 = "precision is disabled because it is not enabled in both" \
" profiles."
warning2 = "max is disabled because it is not enabled in both " \
"profiles."
self.assertIn(warning1, list_of_warning_messages)
self.assertIn(warning2, list_of_warning_messages)
# Assert that these features are still merged
self.assertEqual("rice", profiler3.histogram_selection)
self.assertEqual(21.5, profiler3.variance)
self.assertEqual(45.0, profiler3.sum)
# Assert that these features are not calculated
self.assertIsNone(profiler3.max)
self.assertIsNone(profiler3.min)
self.assertEqual(0, profiler3.precision)
def test_profile_merge(self):
data = [2.0, 'not a float', 6.0, 'not a float']
df = pd.Series(data).apply(str)
profiler1 = FloatColumn("Float")
profiler1.update(df)
data2 = [10.0, 'not a float', 15.0, 'not a float']
df2 = pd.Series(data2).apply(str)
profiler2 = FloatColumn("Float")
profiler2.update(df2)
expected_profile = dict(
min=2.0,
max=15.0,
mean=8.25,
variance=30.916666666666668,
stddev=np.sqrt(30.916),
histogram={
'bin_counts': np.array([1, 1, 1, 1]),
'bin_edges': np.array([2., 5.25, 8.5, 11.75, 15.])
},
)
profiler3 = profiler1 + profiler2
expected_histogram = expected_profile.pop('histogram')
profile3 = profiler3.profile
histogram = profile3.pop('histogram')
self.assertAlmostEqual(profiler3.stddev,
expected_profile.pop('stddev'),
places=3)
self.assertAlmostEqual(profiler3.variance,
expected_profile.pop('variance'),
places=3)
self.assertEqual(profiler3.mean, expected_profile.pop('mean'))
self.assertEqual(profiler3.histogram_selection, 'rice')
self.assertEqual(profiler3.min, expected_profile.pop('min'))
self.assertEqual(profiler3.max, expected_profile.pop('max'))
self.assertCountEqual(histogram['bin_counts'],
expected_histogram['bin_counts'])
self.assertCountEqual(histogram['bin_edges'],
expected_histogram['bin_edges'])
def test_single_data_variance_case(self):
data = pd.Series([1.5]).apply(str)
profiler = FloatColumn(data.name)
profiler.update(data)
self.assertEqual(profiler.match_count, 1.0)
self.assertEqual(profiler.mean, 1.5)
self.assertEqual(profiler.variance, 0.0)
data = pd.Series([2.5]).apply(str)
profiler.update(data)
self.assertEqual(profiler.match_count, 2)
self.assertEqual(profiler.mean, 2.0)
self.assertEqual(profiler.variance, 0.5)
def test_select_method_for_histogram(self):
data = pd.Series([], dtype=object)
profiler = FloatColumn(data.name)
profiler.update(data)
list_method = [
'auto', 'fd', 'doane', 'scott', 'rice', 'sturges', 'sqrt'
]
current_exact_var = 0
# sqrt has the least current loss
current_est_var = np.array([0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.005])
current_total_var = np.array(
[0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01])
current_run_time = np.array([0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01])
# all methods have the same total loss
list_total_loss = np.array([0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1])
for i, method in enumerate(list_method):
profiler.histogram_methods[method]['total_loss'] = \
list_total_loss[i]
selected_method = profiler._select_method_for_histogram(
current_exact_var, current_est_var, current_total_var,
current_run_time)
self.assertEqual(selected_method, 'sqrt')
# another test
current_exact_var = 0
# sqrt has the least current loss
current_est_var = np.array([0.03, 0.03, 0.03, 0.03, 0.03, 0.03, 0.029])
current_total_var = np.array(
[0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01])
current_run_time = np.array([0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01])
# but sturges has the least total loss
list_total_loss = np.array([0.1, 0.1, 0.1, 0.1, 0.1, 0.05, 0.1])
for i, method in enumerate(list_method):
profiler.histogram_methods[method]['total_loss'] = \
list_total_loss[i]
selected_method = profiler._select_method_for_histogram(
current_exact_var, current_est_var, current_total_var,
current_run_time)
self.assertEqual(selected_method, 'sturges')
def test_profiled_min(self):
# test with multiple values
data = np.linspace(-5, 5, 11)
df = pd.Series(data).apply(str)
profiler = FloatColumn(df.name)
profiler.update(df[1:])
self.assertEqual(profiler.min, -4)
profiler.update(df)
self.assertEqual(profiler.min, -5)
profiler.update(pd.Series(['-4']))
self.assertEqual(profiler.min, -5)
# empty data
data = pd.Series([], dtype=object)
profiler = FloatColumn(data.name)
profiler.update(data)
self.assertEqual(profiler.min, None)
# data with None value
df = pd.Series([2.0, 3.0, None, np.nan]).apply(str)
profiler = FloatColumn(df.name)
profiler.update(df)
self.assertEqual(profiler.min, 2.0)
# data with one value
df = pd.Series([2.0]).apply(str)
profiler = FloatColumn(df.name)
profiler.update(df)
self.assertEqual(profiler.min, 2.0)
# data with unique value
df = pd.Series([2.0, 2.0, 2.0, 2.0, 2.0]).apply(str)
profiler = FloatColumn(df.name)
profiler.update(df)
self.assertEqual(profiler.min, 2.0)
# data with unique value as zero
df = pd.Series([0.0, 0.0, 0.0, 0.0, 0.0]).apply(str)
profiler = FloatColumn(df.name)
profiler.update(df)
self.assertEqual(profiler.min, 0.0)
def test_profile_merge_edge_case(self):
data = [2.0, 'not a float', 6.0, 'not a float']
df = pd.Series(data).apply(str)
profiler1 = FloatColumn("Float")
profiler1.update(df)
profiler1.match_count = 0
data2 = [10.0, 'not a float', 15.0, 'not a float']
df2 = pd.Series(data2).apply(str)
profiler2 = FloatColumn("Float")
profiler2.update(df2)
profiler3 = profiler1 + profiler2
self.assertEqual(profiler3.stddev, profiler2.stddev)
# test merge with empty data
df1 = pd.Series([], dtype=object)
profiler1 = FloatColumn("Float")
profiler1.update(df1)
df2 = pd.Series([], dtype=object)
profiler2 = FloatColumn("Float")
profiler2.update(df2)
profiler = profiler1 + profiler2
self.assertEqual(profiler.min, None)
self.assertEqual(profiler.max, None)
df3 = pd.Series([2.0, 3.0]).apply(str)
profiler3 = FloatColumn("Float")
profiler3.update(df3)
profiler = profiler1 + profiler3
self.assertEqual(profiler.min, 2.0)
self.assertEqual(profiler.max, 3.0)
df4 = pd.Series([4.0, 5.0]).apply(str)
profiler4 = FloatColumn("Float")
profiler4.update(df4)
profiler = profiler3 + profiler4
self.assertEqual(profiler.min, 2.0)
self.assertEqual(profiler.max, 5.0)
def test_option_precision(self):
data = [1.1, 2.2, 3.3, 4.4]
df = pd.Series(data).apply(str)
# Turn off precision
options = FloatOptions()
options.set({"precision.is_enabled": False})
profiler = FloatColumn(df.name, options=options)
profiler.update(df)
self.assertEqual(None, profiler.precision['sample_size'])
# Turn on precision, check sample_size
options = FloatOptions()
options.set({"precision.is_enabled": True})
profiler = FloatColumn(df.name, options=options)
profiler.update(df)
self.assertEqual(4, profiler.precision['sample_size'])
# Trun on precision, set 0.5 sample_size
options = FloatOptions()
options.set({"precision.sample_ratio": 0.5})
profiler = FloatColumn(df.name, options=options)
profiler.update(df)
self.assertEqual(2, profiler.precision['sample_size'])
def test_profiled_max(self):
data = np.linspace(-5, 5, 11)
df = pd.Series(data).apply(str)
profiler = FloatColumn(df.name)
profiler.update(df[:-1])
self.assertEqual(profiler.max, 4)
profiler.update(df)
self.assertEqual(profiler.max, 5)
profiler.update(pd.Series(['4']))
self.assertEqual(profiler.max, 5)
# empty data
data = pd.Series([], dtype=object)
profiler = FloatColumn(data.name)
profiler.update(data)
self.assertEqual(profiler.max, None)
# data with None value
df = pd.Series([2.0, 3.0, None, np.nan]).apply(str)
profiler = FloatColumn(df.name)
profiler.update(df)
self.assertEqual(profiler.max, 3.0)
# data with one value
df = pd.Series([2.0]).apply(str)
profiler = FloatColumn(df.name)
profiler.update(df)
self.assertEqual(profiler.max, 2.0)
# data with unique value
df = pd.Series([2.0, 2.0, 2.0, 2.0, 2.0]).apply(str)
profiler = FloatColumn(df.name)
profiler.update(df)
self.assertEqual(profiler.max, 2.0)
# data with unique value as zero
df = pd.Series([0.0, 0.0, 0.0, 0.0, 0.0]).apply(str)
profiler = FloatColumn(df.name)
profiler.update(df)
self.assertEqual(profiler.max, 0.0)
def test_profile(self):
data = [2.5, 12.5, 'not a float', 5, 'not a float']
df = pd.Series(data).apply(str)
profiler = FloatColumn(df.name)
expected_profile = dict(
min=2.5,
max=12.5,
mean=20/3.0,
median=None,
variance=27 + 1/12.0,
stddev=np.sqrt(27+1/12.0),
histogram={
'bin_counts': np.array([1, 1, 0, 1]),
'bin_edges': np.array([2.5, 5.0, 7.5, 10.0, 12.5]),
},
quantiles={
0: 3.75,
1: 5.0,
2: 10.0
},
times=defaultdict(float, {'histogram_and_quantiles': 15.0,\
'precision': 1.0, 'max': 1.0, 'min': 1.0,\
'sum': 1.0, 'variance': 1.0}),
precision=1.0
)
time_array = [float(i) for i in range(100, 0, -1)]
with mock.patch('time.time', side_effect=lambda: time_array.pop()):
# Validate that the times dictionary is empty
self.assertEqual(defaultdict(float), profiler.profile['times'])
profiler.update(df)
profile = profiler.profile
# pop out the histogram to test separately from the rest of the dict
# as we need comparison with some precision
histogram = profile.pop('histogram')
expected_histogram = expected_profile.pop('histogram')
quantiles = profile.pop('quantiles')
expected_quantiles = expected_profile.pop('quantiles')
actual_quartiles = {
0: quantiles[249],
1: quantiles[499],
2: quantiles[749]
}
self.assertDictEqual(expected_profile, profile)
self.assertEqual(expected_profile['precision'], 1.0)
self.assertCountEqual(expected_histogram['bin_counts'],
histogram['bin_counts'])
self.assertCountEqual(
np.round(expected_histogram['bin_edges'], 12),
np.round(histogram['bin_edges'], 12))
self.assertDictEqual(actual_quartiles, expected_quantiles)
# Validate time in datetime class has expected time after second update
profiler.update(df)
expected = defaultdict(float, {'min': 2.0, 'max': 2.0, 'sum': 2.0,\
'variance': 2.0, 'precision': 2.0,\
'histogram_and_quantiles': 30.0})
self.assertEqual(expected, profiler.profile['times'])
def test_profiled_mean_and_variance(self):
"""
Checks the mean and variance of profiled numerical columns.
:return:
"""
def mean(df):
total = 0
for item in df:
total += item
return total / len(df)
def var(df):
var = 0
mean_df = mean(df)
for item in df:
var += (item - mean_df)**2
return var / (len(df) - 1)
def batch_variance(mean_a, var_a, count_a, mean_b, var_b, count_b):
delta = mean_b - mean_a
m_a = var_a * (count_a - 1)
m_b = var_b * (count_b - 1)
M2 = m_a + m_b + delta**2 * count_a * count_b / (count_a + count_b)
return M2 / (count_a + count_b - 1)
data = np.linspace(-5, 5, 11).tolist()
df1 = pd.Series(data)
data = np.linspace(-3, 2, 11).tolist()
df2 = pd.Series(data)
data = np.full((10, ), 1)
df3 = pd.Series(data)
num_profiler = FloatColumn(df1.name)
num_profiler.update(df1.apply(str))
self.assertEqual(mean(df1), num_profiler.mean)
self.assertEqual(var(df1), num_profiler.variance)
self.assertEqual(np.sqrt(var(df1)), num_profiler.stddev)
variance = batch_variance(mean_a=num_profiler.mean,
var_a=num_profiler.variance,
count_a=num_profiler.match_count,
mean_b=mean(df2),
var_b=var(df2),
count_b=df2.count())
num_profiler.update(df2.apply(str))
df = pd.concat([df1, df2])
self.assertEqual(mean(df), num_profiler.mean)
self.assertEqual(variance, num_profiler.variance)
self.assertEqual(np.sqrt(variance), num_profiler.stddev)
variance = batch_variance(mean_a=num_profiler.mean,
var_a=num_profiler.variance,
count_a=num_profiler.match_count,
mean_b=mean(df3),
var_b=var(df3),
count_b=df3.count())
num_profiler.update(df3.apply(str))
df = pd.concat([df1, df2, df3])
self.assertEqual(mean(df), num_profiler.mean)
self.assertEqual(variance, num_profiler.variance)
self.assertEqual(np.sqrt(variance), num_profiler.stddev)
def test_profiled_histogram(self):
"""
Checks the histogram of profiled numerical columns.
:return:
"""
list_data_test = []
# this data has 4 bins, range of 3
# with equal bin size, each bin has the width of 0.75
df1 = pd.Series(["1.0", "2.0", "3.0", "4.0"])
expected_histogram1 = {
'bin_counts': np.array([1, 1, 1, 1]),
'bin_edges': np.array([1.0, 1.75, 2.5, 3.25, 4.0]),
}
list_data_test.append([df1, expected_histogram1])
# this data has 4 bins, range of 12
# with equal bin size, each bin has the width of 3.0
df2 = pd.Series(["1.0", "5.0", "8.0", "13.0"])
expected_histogram2 = {
'bin_counts': np.array([1, 1, 1, 1]),
'bin_edges': np.array([1.0, 4.0, 7.0, 10.0, 13.0]),
}
list_data_test.append([df2, expected_histogram2])
# this data has 3 bins, range of 3
# with equal bin size, each bin has the width of 1
df3 = pd.Series(["1.0", "1.0", "3.0", "4.0"])
expected_histogram3 = {
'bin_counts': np.array([2, 0, 2]),
'bin_edges': np.array([1.0, 2.0, 3.0, 4.0]),
}
list_data_test.append([df3, expected_histogram3])
# this data has only one unique value, not overflow
df4 = pd.Series([-10.0, -10.0, -10.0]).apply(str)
expected_histogram4 = {
'bin_counts': np.array([3]),
'bin_edges': np.array([-10.0, -10.0]),
}
list_data_test.append([df4, expected_histogram4])
# this data has only one unique value, overflow
df5 = pd.Series([-10.0**20]).apply(str)
expected_histogram5 = {
'bin_counts': np.array([1]),
'bin_edges': np.array([-10.0**20, -10.0**20]),
}
list_data_test.append([df5, expected_histogram5])
for i, (df, expected_histogram) in enumerate(list_data_test):
profiler = FloatColumn(df.name)
profiler.update(df)
profile = profiler.profile
histogram = profile['histogram']
self.assertCountEqual(expected_histogram['bin_counts'],
histogram['bin_counts'])
if i != 4:
self.assertCountEqual(
np.round(expected_histogram['bin_edges'], 12),
np.round(histogram['bin_edges'], 12))
else: # for overflow, dont use np.round
self.assertCountEqual(expected_histogram['bin_edges'],
histogram['bin_edges'])
def test_profiled_precision(self):
"""
Checks whether the precision for the profiler is correct.
:return:
"""
df_1 = pd.Series([0.4, 0.3, 0.1, 0.1, 0.1]).apply(str)
df_2 = pd.Series([0.11, 0.11, 0.12, 2.11]).apply(str)
df_3 = pd.Series([4.114, 3.161, 2.512, 2.131]).apply(str)
df_mix = pd.Series([4.1, 3., 2.52, 2.13143]).apply(str)
float_profiler = FloatColumn("Name")
float_profiler.update(df_3)
self.assertEqual(3, float_profiler.precision)
float_profiler.update(df_2)
self.assertEqual(2, float_profiler.precision)
float_profiler.update(df_1)
self.assertEqual(1, float_profiler.precision)
float_profiler = FloatColumn("Name")
float_profiler.update(df_mix)
self.assertEqual(1, float_profiler.precision)
def test_float_column_with_wrong_options(self):
with self.assertRaisesRegex(
ValueError, "FloatColumn parameter 'options' must be of"
" type FloatOptions."):
profiler = FloatColumn("Float", options="wrong_data_type")
def test_histogram_with_varying_number_of_bin(self):
"""
Checks the histogram with large number of bins
"""
# this data use number of bins less than the max limit
df1 = pd.Series([1, 2, 3, 4]).apply(str)
profiler1 = FloatColumn(df1.name)
profiler1.max_histogram_bin = 50
profiler1.update(df1)
num_bins = len(profiler1.profile['histogram']['bin_counts'])
self.assertEqual(num_bins, 4)
# this data uses large number of bins, which will be set to
# the max limit
df2 = pd.Series(
[3.195103249264023e+18, 9999995.0, 9999999.0, 0.0,
-10**10]).apply(str)
profiler2 = FloatColumn(df2.name)
profiler2.max_histogram_bin = 50
profiler2.update(df2)
num_bins = len(profiler2.profile['histogram']['bin_counts'])
self.assertEqual(num_bins, 50)
# max number of bin is increased to 10000
profiler2 = FloatColumn(df2.name)
profiler2.max_histogram_bin = 10000
profiler2.update(df2)
num_bins = len(profiler2.profile['histogram']['bin_counts'])
self.assertEqual(num_bins, 10000)
