def test_clean_db_drop_na():
d = pd.DataFrame()
N = 100
d['quality'] = np.linspace(0, 1E-3, 5 * N)
d['rhythm'] = np.asarray([1, 2, 3, 4, np.nan] * N)
d['weight'] = np.asarray(np.asarray([
0,
] * N + list(np.arange(N)) * 4))
d['age'] = d['weight']
df = clean_db.clean_db(d,
drop_columns=None,
drop_na_in=['rhythm'],
drop_zeroes_in=None,
quality_threshold=None)
assert len(df) == 4 * N == len(df.dropna(subset=('rhythm', )))
def test_clean_db_quality():
d = pd.DataFrame()
N = 100
d['quality'] = np.linspace(0, 1E-3, 5 * N)
d['rhythm'] = np.asarray([1, 2, 3, 4, np.nan] * N)
d['weight'] = np.asarray(np.asarray([
0,
] * N + list(np.arange(N)) * 4))
d['age'] = d['weight']
df = clean_db.clean_db(d,
drop_columns=None,
drop_na_in=None,
drop_zeroes_in=None,
quality_threshold=1E-4)
assert max(df.quality) < 1E-4
assert len(df) < len(d)
assert len(df) == sum(np.where(d.quality < 1E-4, 1, 0))
def test_clean_db_drop_zeroes():
d = pd.DataFrame()
N = 100
d['quality'] = np.linspace(0, 1E-3, 5 * N)
d['rhythm'] = np.asarray([1, 2, 3, 4, np.nan] * N)
d['weight'] = np.asarray(np.asarray([
0,
] * N + list(np.arange(N)) * 4))
d['age'] = d['weight']
zeroes_labels = ['age', 'weight']
df = clean_db.clean_db(d,
drop_columns=None,
drop_na_in=None,
drop_zeroes_in=zeroes_labels,
quality_threshold=None)
assert len(df) == len(d) - 104 # 100 zeroes + 4 zeroes
for _ in zeroes_labels:
assert min(abs(df[_])) > 0
def test_clean_db_drop_columns():
d = pd.DataFrame()
N = 100
d['quality'] = np.linspace(0, 1E-3, 5 * N)
d['rhythm'] = np.asarray([1, 2, 3, 4, np.nan] * N)
d['weight'] = np.asarray(np.asarray([
0,
] * N + list(np.arange(N)) * 4))
d['age'] = d['weight']
drop_columns_df = ['rhythm', 'weight']
df = clean_db.clean_db(d,
drop_columns=drop_columns_df,
drop_na_in=None,
drop_zeroes_in=None,
quality_threshold=None)
df_col = list(set(df.columns) & set(drop_columns_df))
assert len(df_col) == 0
assert len(df) == len(d)