def setUp(self):
self.df = pd.DataFrame({
'id': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
'category': ['a', 'b', 'c', 'd', 'a', 'b', 'c', 'd', 'a', 'b'],
'gender': ['F', 'M', 'F', 'M', 'F', 'M', 'F', 'M', 'F', 'M'],
'age': [1, 2, 3, 1, 2, 3, 1, 2, 3, 1]
})
self.cardinality = cardinality.calculate_cardinality(self.df)
def setUp(self):
self.df = pd.DataFrame({
'id': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
'category': ['a', 'b', 'c', 'd', 'a', 'b', 'c', 'd', 'a', 'b'],
'gender': ['F', 'M', 'F', 'M', 'F', 'M', 'F', 'M', 'F', 'M'],
'age': [1, 2, 3, 1, 2, 3, 1, 2, 3, 1]
})
self.cardinality = cardinality.calculate_cardinality(self.df)
def test_calculates_cardinality(self):
expected = pd.DataFrame({
'Feature Name': ['id', 'age', 'category', 'gender', 'boring'],
'unique_value_count': [4, 3, 3, 2, 1],
'unique_ratio': [1, 0.75, 0.75, 0.5, 0.25]
})
result = cardinality.calculate_cardinality(self.df)
for column in expected:
self.assertEqual(result[column].all(), expected[column].all())
def test_calculates_cardinality(self):
expected = pd.DataFrame({
'Feature Name': ['id', 'age', 'category', 'gender', 'boring'],
'unique_value_count': [4, 3, 3, 2, 1],
'unique_ratio': [1, 0.75, 0.75, 0.5, 0.25]
})
result = cardinality.calculate_cardinality(self.df)
for column in expected:
self.assertEqual(result[column].all(), expected[column].all())
def setUp(self):
self.df = pd.DataFrame({
'id': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
'category': ['a', 'b', 'c', 'd', 'a', 'b', 'c', 'd', 'a', 'b'],
'gender': ['F', 'M', 'F', 'M', 'F', 'M', 'F', 'M', 'F', 'M'],
'age': [1, 2, 3, 1, 2, 3, 1, 2, 3, 1],
'boring': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
})
self.df['bad_string'] = 'yup'
self.df['bad_float'] = 3.33
self.df['bad_int'] = 1
self.df['bad_bool'] = False
self.cardinality = cardinality.calculate_cardinality(self.df)
def setUp(self):
self.df = pd.DataFrame({
'id': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
'category': ['a', 'b', 'c', 'd', 'a', 'b', 'c', 'd', 'a', 'b'],
'gender': ['F', 'M', 'F', 'M', 'F', 'M', 'F', 'M', 'F', 'M'],
'age': [1, 2, 3, 1, 2, 3, 1, 2, 3, 1],
'boring': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
})
self.df['bad_string'] = 'yup'
self.df['bad_float'] = 3.33
self.df['bad_int'] = 1
self.df['bad_bool'] = False
self.cardinality = cardinality.calculate_cardinality(self.df)
def test_returns_dataframe(self):
self.assertIsInstance(
cardinality.calculate_cardinality(self.df),
pd.DataFrame)
def test_returns_dataframe(self):
self.assertIsInstance(
cardinality.calculate_cardinality(self.df),
pd.DataFrame)