def test_normalize_max(random_df):
data = random_df
normalized = utils.normalize(data, "max")
# check max index
didx = data.idxmax(axis=1)
nidx = normalized.idxmax(axis=1)
assert (didx == nidx).all()
# check max equals to 1
assert np.isclose(normalized.max(axis=1).values, 1).all()
def test_normalize_sum(random_df):
data = random_df
normalized = utils.normalize(data, "sum")
# check max index
didx = data.idxmax(axis=1)
nidx = normalized.idxmax(axis=1)
assert (didx == nidx).all()
# check that each row sums 1
assert np.isclose(normalized.sum(axis=1).values, 1).all()
def test_normalize_invalid_mode(random_df):
data = random_df
with pytest.raises(ValueError):
utils.normalize(data, "invalid_mode")
def test_normalize_feature(random_df):
data = random_df
ft = data.columns[25]
normalized = utils.normalize(data, "feature", ft)
assert np.isclose(normalized[ft], 1).all()
def test_normalize_euclidean(random_df):
data = random_df
normalized = utils.normalize(data, "euclidean")
norm = normalized.apply(lambda x: np.linalg.norm(x), axis=1)
assert np.isclose(norm, 1).all()
def test_normalize_max(random_df):
data = random_df
normalized = utils.normalize(data, "max")
assert np.isclose(normalized.max(axis=1).values, 1).all()
def test_normalize_sum(random_df):
data = random_df
normalized = utils.normalize(data, "sum")
assert np.isclose(normalized.sum(axis=1).values, 1).all()