def test_create_residuals_df_covars_none(subject_array, subject_data):
names, covars = ['noggin_left', 'noggin_right'], []
array_resids = [residuals(subject_array[:, 2:2].T, subject_array[:, i])
for i in [0, 1]]
np.testing.assert_almost_equal(
np.array(create_residuals_df(subject_data, names, covars)[names]),
np.array(array_resids).T)
def test_create_residuals_df_covars_singular():
names, covars = ['noggin_left', 'noggin_right'], ['day of week']
array_resids = [
residuals(subject_array()[:, 2:3].T,
subject_array()[:, i]) for i in [0, 1]
]
np.testing.assert_almost_equal(
np.array(create_residuals_df(subject_data(), names, covars)[names]),
np.array(array_resids).T)
def create_residuals_df(df, names, covars=[]):
'''
Calculate residuals of columns specified by names, correcting for the
columns in covars_list.
Parameters
----------
df : :class:`pandas.DataFrame`
A pandas data frame with subjects as rows and columns including
brain regions and covariates. Should be numeric for the columns in
`names` and `covars`.
names : list
A list of the brain regions you wish to correlate.
covars: list, optional
A list of covariates to correct for before correlating
the regional measures. Each element should correspond to a
column heading in `df`.
Default is an empty list.
Returns
-------
:class:`pandas.DataFrame`
Residuals of columns `names` of `df`, correcting for `covars`
Raises
------
TypeError
if there are non numeric entries in the columns specified by `names` or
`covars`
'''
# Raise TypeError if any of the relevant columns are nonnumeric
non_numeric_cols = get_non_numeric_cols(df[names + covars])
if non_numeric_cols:
raise TypeError('DataFrame columns {} are non numeric'.format(
', '.join(non_numeric_cols)))
# Make a new data frame that will contain
# the residuals for each column after correcting for
# the covariates in covars
df_res = df[names + covars].copy()
# Create your covariates array
if len(covars) > 1:
x = np.vstack([df[covars]])
elif len(covars) == 1:
x = df[covars]
else:
x = np.ones_like(df.iloc[:, 0])
# Calculate the residuals
for name in names:
df_res.loc[:, name] = residuals(x.T, df.loc[:, name])
# Return the residuals data frame
return df_res