def dummy_metrics():
"""Create a Dict of dummy metric."""
metrics = {'with float': Metric(data=np.array([29., 31., 39., 41., 50.])),
'with int': Metric(data=np.array([99, 100, 101, 102, 103])),
'with nan': Metric(data=np.array([99, np.nan, 101, 102, 103])),
'inconsistent length': Metric(data=np.array([99, 100])),
'with string': Metric(data=np.array([99, "boo!", 101, 102, 103]))}
return metrics
def dummy_data_and_labels_2d():
"""Create a dummy 2d data with associated 2d label, for testing extract_metric()."""
data = Metric(data=np.array([[5, 5],
[5, 5]]))
labels = np.array([[1, 1],
[1, 1]]).T # need to transpose because last dim are labels
labels = np.expand_dims(labels, axis=2) # because ndim(label) = ndim(data)+1
# Create label_struc{}
label_struc = {0: aggregate_slicewise.LabelStruc(id=0, name='mask')}
return data, labels, label_struc
def dummy_data_and_labels():
"""Create a dummy data with partial volume effect, with associated mask, for testing extract_metric()."""
data = Metric(data=np.array([19., 21., 30., 39., 41.]))
# Create 3 labels. The last label has very small volume fraction to assess the efficacy of MAP estimation.
labels = np.array([[0., 0., 0.5, 1., 1.],
[0.9, 1., 0.5, 0., 0.],
[0.1, 0., 0., 0., 0.]]).T # need to transpose because last dim are labels
# Create label_struc{}
label_struc = {0: aggregate_slicewise.LabelStruc(id=0, name='label_0', map_cluster=0),
1: aggregate_slicewise.LabelStruc(id=1, name='label_1', map_cluster=1),
2: aggregate_slicewise.LabelStruc(id=2, name='label_2', map_cluster=1),
99: aggregate_slicewise.LabelStruc(id=[1, 2], name='label_1,2', map_cluster=None)}
return data, labels, label_struc
def dummy_metrics():
"""Create a Dict of dummy metric."""
metrics = {
'with float':
Metric(data=np.array([29., 31., 39., 41., 50., 51., 59., 62., 70.])),
'with int':
Metric(data=np.array([99, 100, 101, 102, 103, 104, 105, 106, 107])),
'with nan':
Metric(data=np.array([99, np.nan, 101, 102, 103, 104, 105, 106, 107])),
'inconsistent length':
Metric(data=np.array([99, 100])),
'with string':
Metric(data=np.array([99, "boo!", 101, 102, 103, 104, 105, 106, 107])),
'3D':
Metric(data=np.resize(
np.array([99, 100, 101, 102, 103, 104, 105, 106, 107]), [4, 5, 9]))
}
return metrics