def test_diff(self):
"""
Makes sure the IntColumn Diff() works appropriately.
"""
data = [2, 'not an int', 6, 4]
df = pd.Series(data).apply(str)
profiler1 = IntColumn("Int")
profiler1.update(df)
data = [1, 15]
df = pd.Series(data).apply(str)
profiler2 = IntColumn("Int")
profiler2.update(df)
# Assert the difference report is correct
expected_diff = {
'max': -9.0,
'mean': -4.0,
'min': 1.0,
'stddev': -7.899494936611665,
'sum': -4.0,
'variance': -94.0,
'median': -4,
'mode': [[2, 6, 4], [], [1, 15]],
'median_absolute_deviation': -5,
't-test': {
't-statistic': -0.5638091828819275,
'conservative': {
'df': 1,
'p-value': 0.6731699660830497
},
'welch': {
'df': 1.0547717074524683,
'p-value': 0.6691886269547123
}
}
}
profile_diff = profiler1.diff(profiler2)
self.assertAlmostEqual(expected_diff.pop('median'),
profile_diff.pop('median'),
places=2)
expected_diff_mode = expected_diff.pop('mode')
diff_mode = profile_diff.pop('mode')
for i in range(len(expected_diff_mode)):
np.testing.assert_almost_equal(sorted(expected_diff_mode[i]),
sorted(diff_mode[i]), 2)
self.assertAlmostEqual(expected_diff.pop('median_absolute_deviation'),
profile_diff.pop('median_absolute_deviation'),
places=2)
self.assertDictEqual(expected_diff, profile_diff)
# Assert type error is properly called
with self.assertRaises(TypeError) as exc:
profiler1.diff("Inproper input")
self.assertEqual(
str(exc.exception),
"Unsupported operand type(s) for diff: 'IntColumn' and"
" 'str'")
def test_diff(self):
"""
Makes sure the IntColumn Diff() works appropriately.
"""
data = [2, "not an int", 6, 4]
df = pd.Series(data).apply(str)
profiler1 = IntColumn("Int")
profiler1.update(df)
data = [1, 15]
df = pd.Series(data).apply(str)
profiler2 = IntColumn("Int")
profiler2.update(df)
# Assert the difference report is correct
expected_diff = {
"max": -9.0,
"mean": -4.0,
"min": 1.0,
"stddev": -7.899494936611665,
"sum": -4.0,
"variance": -94.0,
"median": -4,
"mode": [[2, 6, 4], [], [1, 15]],
"median_absolute_deviation": -5,
"t-test": {
"t-statistic": -0.5638091828819275,
"conservative": {
"df": 1,
"p-value": 0.6731699660830497
},
"welch": {
"df": 1.0547717074524683,
"p-value": 0.6691886269547123
},
},
}
profile_diff = profiler1.diff(profiler2)
try:
json.dumps(profile_diff)
except TypeError as e:
self.fail("JSON Serializing issue with the profile diff. "
"Exception raised: {}".format(str(e)))
self.assertAlmostEqual(expected_diff.pop("median"),
profile_diff.pop("median"),
places=2)
expected_diff_mode = expected_diff.pop("mode")
diff_mode = profile_diff.pop("mode")
for i in range(len(expected_diff_mode)):
np.testing.assert_almost_equal(sorted(expected_diff_mode[i]),
sorted(diff_mode[i]), 2)
self.assertAlmostEqual(
expected_diff.pop("median_absolute_deviation"),
profile_diff.pop("median_absolute_deviation"),
places=2,
)
self.assertDictEqual(expected_diff, profile_diff)
# Assert type error is properly called
with self.assertRaises(TypeError) as exc:
profiler1.diff("Inproper input")
self.assertEqual(
str(exc.exception),
"Unsupported operand type(s) for diff: 'IntColumn' and"
" 'str'",
)