def _test_datalad_openfmri_dataset(d):
of = ofm.OpenFMRIDataset(d)
# smoke tests that we can load the dataset's attributes etc
#assert(of.get_model_descriptions())
sub_ids = of.get_subj_ids()
assert (sub_ids)
assert ('TR' in of.get_scan_properties())
tasks = of.get_task_descriptions()
assert (tasks)
task = tasks.keys()[0]
run_ids = of.get_bold_run_ids(sub_ids[0], task)
assert (run_ids)
task_runs = of.get_task_bold_run_ids(task)
assert (task_runs)
model_ids = of.get_model_ids()
assert (model_ids)
from datalad.auto import AutomagicIO
with AutomagicIO(): # so necessary files get fetched if necessary
# try loading first run for some task of the first subject
data = of.get_model_bold_dataset(model_ids[0], sub_ids[0],
task_runs.values()[0][:1])
assert (data.shape)
def test_openfmri_dataset():
skip_if_no_external('nibabel')
of = ofm.OpenFMRIDataset(pathjoin(pymvpa_dataroot, 'haxby2001'))
assert_equal(of.get_model_descriptions(), {1: 'visual object categories'})
sub_ids = of.get_subj_ids()
assert_equal(sub_ids, [1, 'phantom'])
assert_equal(of.get_scan_properties(), {'TR': '2.5'})
tasks = of.get_task_descriptions()
assert_equal(tasks, {1: 'object viewing'})
task = tasks.keys()[0]
run_ids = of.get_bold_run_ids(sub_ids[0], task)
assert_equal(run_ids, range(1, 13))
task_runs = of.get_task_bold_run_ids(task)
assert_equal(task_runs, {1: range(1, 13)})
# test access anatomy image
assert_equal(
of.get_anatomy_image(1, fname='lowres001.nii.gz').shape, (6, 10, 10))
# try to get an image that isn't there
assert_raises(IOError, of.get_bold_run_image, 1, 1, 1)
# defined model contrasts
contrast_spec = of.get_model_contrasts(1)
# one dict per task
assert_equal(len(contrast_spec), 1)
assert_true(1 in contrast_spec)
# six defined contrasts
assert_equal(len(contrast_spec[1]), 1)
# check one
assert_array_equal(contrast_spec[1]['face_v_house'],
[-1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0])
orig_attrs = SampleAttributes(
pathjoin(pymvpa_dataroot, 'attributes_literal.txt'))
for subj, runs in task_runs.iteritems():
for run in runs:
# load single run
ds = of.get_bold_run_dataset(subj,
task,
run,
flavor='1slice',
mask=pathjoin(pymvpa_dataroot,
'mask.nii.gz'),
add_sa='bold_moest.txt')
# basic shape
assert_equal(len(ds), 121)
assert_equal(ds.nfeatures, 530)
# functional mapper
assert_equal(ds.O.shape, (121, 40, 20, 1))
# additional attributes present
moest = of.get_bold_run_motion_estimates(subj, task, run)
for i in range(6):
moest_attr = 'bold_moest.txt_%i' % (i, )
assert_true(moest_attr in ds.sa)
assert_array_equal(moest[:, i], ds.sa[moest_attr].value)
# check conversion of model into sample attribute
events = of.get_bold_run_model(subj, task, run)
for i, ev in enumerate(events):
# we only have one trial per condition in the demo dataset
assert_equal(ev['conset_idx'], 0)
# proper enumeration and events sorted by time
assert_equal(ev['onset_idx'], i)
onsets = [e['onset'] for e in events]
sorted_onsets = sorted(onsets)
assert_array_equal(sorted_onsets, onsets)
targets = events2sample_attr(events,
ds.sa.time_coords,
noinfolabel='rest')
assert_array_equal(
orig_attrs['targets'][(run - 1) * 121:run * len(ds)], targets)
assert_equal(ds.sa['subj'][0], subj)
# check that we can get the same result from the model dataset
# (make it exercise the preproc interface too)
def preproc_img(img):
return img
modelds = of.get_model_bold_dataset(1,
subj,
flavor='1slice',
preproc_img=preproc_img,
modelfx=assign_conditionlabels,
mask=pathjoin(
pymvpa_dataroot,
'mask.nii.gz'),
add_sa='bold_moest.txt')
modelds = modelds[modelds.sa.run == run]
targets = np.array(targets, dtype='object')
targets[targets == 'rest'] = None
assert_array_equal(targets, modelds.sa.targets)
# more basic access
motion = of.get_task_bold_attributes(1, 'bold_moest.txt', np.loadtxt)
assert_equal(len(motion), 12) # one per run
# one per subject, per volume, 6 estimates
assert_equal([(len(m), ) + m[1].shape for m in motion], [(1, 121, 6)] * 12)