def test_missing_task(default_param):
# input invalid value for tasks
default_param["load_data"]["tasks"] = ""
# Load data using the invalid field
with pytest.raises(Exception):
load.load_files(default_param["load_data"])
assert True
def test_missing_except_subj(default_param):
# input invalid value for tasks
default_param["load_data"]["exceptions"]["subjects"] = ""
default_param["load_data"]["exceptions"]["tasks"] = ["*"]
default_param["load_data"]["exceptions"]["runs"] = ["*"]
# Load data using the invalid field
with pytest.raises(Exception):
load.load_files(default_param["load_data"])
assert True
def test_missing_data(default_param, tmp_path):
# create empty and temporary directory
empty_data = tmp_path / "empty"
empty_data.mkdir()
# input invalid value for tasks
default_param["load_data"]["root"] = empty_data
# Load data using the invalid field
with pytest.raises(Exception):
load.load_files(default_param["load_data"])
assert True
def test_select_subj_task_except_task(default_param, subj_files):
# select subjects & tasks (randomize?)
selected_subjects = ["NDARAB793GL3"]
selected_tasks = ["ContrastChangeBlock1"]
default_param["load_data"]["subjects"] = selected_subjects
default_param["load_data"]["tasks"] = selected_tasks
# select excluded subjects (randomize?)
exc_task = ["ContrastChangeBlock1"]
default_param["load_data"]["exceptions"]["subjects"] = ["*"]
default_param["load_data"]["exceptions"]["tasks"] = exc_task
default_param["load_data"]["exceptions"]["runs"] = ["*"]
# Load data using the selected subjects & tasks
data = load.load_files(default_param["load_data"])
# check if the number of files loaded matches
# the total number of files for all selected subjects excluding exceptions
count = 0
for subject in selected_subjects:
for file in subj_files[subject]:
# split and find task of file (TODO: need to BIDSIFY)
task_id = [s for s in file.split("_") if "task" in s][0]
task = [s for s in task_id.split("-") if s != "task"][0]
if task in exc_task:
continue
elif task in selected_tasks:
count += 1
assert len(data) == count
def test_except_subj_task(default_param, avail_subj, subj_files):
# select excluded subjects & tasks (randomize?)
exc_subj = ["NDARAB793GL3"]
exc_task = ["ContrastChangeBlock1"]
default_param["load_data"]["exceptions"]["subjects"] = exc_subj
default_param["load_data"]["exceptions"]["tasks"] = exc_task
default_param["load_data"]["exceptions"]["runs"] = ["*"]
# Load data using the excluded subjects & tasks
data = load.load_files(default_param["load_data"])
# check if the number of files loaded matches
# the total number of all subjects & tasks apart from the excluded
count = 0
for subject in avail_subj:
for file in subj_files[subject]:
file_vals = file.split("_")
# split and find subject of file (TODO: need to BIDSIFY)
sub_id = [s for s in file_vals if "sub" in s][0]
sub = [s for s in sub_id.split("-") if s != "sub"][0]
# split and find task of file (TODO: need to BIDSIFY)
task_id = [s for s in file.split("_") if "task" in s][0]
task = [s for s in task_id.split("-") if s != "task"][0]
if sub in exc_subj and task in exc_task:
continue
count += 1
assert len(data) == count
def test_return_values(default_param, select_subjects, select_tasks):
default_param["load_data"]["subjects"] = select_subjects
default_param["load_data"]["tasks"] = select_tasks
# Load data using the selected subjects & tasks
data = load.load_files(default_param["load_data"])
# get the pipeline steps
feature_params = default_param["preprocess"]
interp_param = feature_params["interpolate_data"]
for file in data:
eeg_obj = mne_bids.read_raw_bids(file)
# generate epoched object to be interpolated
epo, _ = pre.segment_data(eeg_obj, **feature_params["segment_data"])
# interpolate data
interp_eeg, output_dict = pre.interpolate_data(epo, **interp_param)
# assert that None does not exist in final reject
assert None not in output_dict.values()
# assert object returned is epoch object
assert isinstance(interp_eeg, Epochs)
def test_select_all(default_param, avail_subj, subj_files):
# Load data using the default parameters
data = load.load_files(default_param["load_data"])
# check if the number of files loaded matches
# the number of all files available
count = 0
for subject in avail_subj:
count += len(subj_files[subject])
assert len(data) == count
def test_except_all(default_param):
# select all for exceptions
default_param["load_data"]["exceptions"]["subjects"] = ["*"]
default_param["load_data"]["exceptions"]["tasks"] = ["*"]
default_param["load_data"]["exceptions"]["runs"] = ["*"]
# Load data using the excluded subjects & tasks
data = load.load_files(default_param["load_data"])
# assert nothing is selected
assert len(data) == 0
def test_select_subj(default_param, subj_files):
# select participants (make this a random selection?)
selected_subjects = ["NDARAB793GL3"]
default_param["load_data"]["subjects"] = selected_subjects
# Load data using the selected participants
data = load.load_files(default_param["load_data"])
# check if the number of files loaded matches
# the total number of files for all selected subjects
count = 0
for subject in selected_subjects:
count += len(subj_files[subject])
assert len(data) == count
def test_return_values(default_param, select_subjects, select_tasks):
default_param["load_data"]["subjects"] = select_subjects
default_param["load_data"]["tasks"] = select_tasks
# Load data using the selected subjects & tasks
data = load.load_files(default_param["load_data"])
for file in data:
eeg_obj = mne_bids.read_raw_bids(file)
# reject epochs
badchans_raw, output_dict = pre.identify_badchans_raw(eeg_obj)
# assert that None does not exist in bad chans
assert None not in output_dict.values()
def test_except_subj(default_param, avail_subj, subj_files):
# select excluded subjects (randomize?)
exc_subj = ["NDARAB793GL3"]
default_param["load_data"]["exceptions"]["subjects"] = exc_subj
default_param["load_data"]["exceptions"]["tasks"] = ["*"]
default_param["load_data"]["exceptions"]["runs"] = ["*"]
# Load data using the excluded subjects
data = load.load_files(default_param["load_data"])
# get the excluded subset of subjects
subjects_diff = set(avail_subj).difference(set(exc_subj))
# check if the number of files loaded matches
# the total number of all subjects apart from the excluded
count = 0
for subject in subjects_diff:
count += len(subj_files[subject])
assert len(data) == count
def test_return_values(default_param, select_subjects, select_tasks):
default_param["load_data"]["subjects"] = select_subjects
default_param["load_data"]["tasks"] = select_tasks
# Load data using the selected subjects & tasks
data = load.load_files(default_param["load_data"])
# get the pipeline steps
feature_params = default_param["preprocess"]
ica_param = feature_params["ica_raw"]
for file in data:
eeg_obj = mne_bids.read_raw_bids(file)
# reject epochs
ica_raw, output_dict = pre.ica_raw(eeg_obj, **ica_param)
# assert that None does not exist in bad chans
assert None not in output_dict.values()
def test_return_values(default_param, select_subjects, select_tasks):
default_param["load_data"]["subjects"] = select_subjects
default_param["load_data"]["tasks"] = select_tasks
# Load data using the selected subjects & tasks
data = load.load_files(default_param["load_data"])
# get the pipeline steps
feature_params = default_param["preprocess"]
filt_param = feature_params["filter_data"]
for file in data:
eeg_obj = mne_bids.read_raw_bids(file)
# filter data
filt_eeg, output_dict = pre.filter_data(eeg_obj, **filt_param)
# assert that None does not exist in final reject
assert None not in output_dict.values()
def test_select_task(default_param, avail_subj, subj_files):
# select tasks (make this a random selection?)
selected_tasks = ["ContrastChangeBlock1"]
default_param["load_data"]["tasks"] = selected_tasks
# Load data using the selected task
data = load.load_files(default_param["load_data"])
# check if the number of files loaded matches
# the total number of files for all selected task
count = 0
for subject in avail_subj:
for file in subj_files[subject]:
# split and find task of file (TODO: need to BIDSIFY)
task_id = [s for s in file.split("_") if "task" in s][0]
task = [s for s in task_id.split("-") if s != "task"][0]
if task in selected_tasks:
count += 1
assert len(data) == count
def test_return_values(default_param, select_subjects, select_tasks):
default_param["load_data"]["subjects"] = select_subjects
default_param["load_data"]["tasks"] = select_tasks
# Load data using the selected subjects & tasks
data = load.load_files(default_param["load_data"])
# get the pipeline steps
feature_params = default_param["preprocess"]
reref_param = feature_params["reref_raw"]
for file in data:
eeg_obj = mne_bids.read_raw_bids(file)
# reref the data
reref_eeg, output_dict = pre.reref_raw(eeg_obj, **reref_param)
# assert that None does not exist in final reject
assert None not in output_dict.values()
# assert object returned is epoch object
assert isinstance(reref_eeg, Epochs)
def test_except_value(default_param, select_subjects, error_tasks):
default_param["load_data"]["subjects"] = select_subjects
default_param["load_data"]["tasks"] = error_tasks
# Load data using the selected subjects & tasks
data = load.load_files(default_param["load_data"])
# get the pipeline steps
feature_params = default_param["preprocess"]
for file in data:
eeg_obj = mne_bids.read_raw_bids(file)
# generate epoched object to be rejected
epo, _ = pre.segment_data(eeg_obj, **feature_params["segment_data"])
# attempt to reject epochs with data containing only one entire epoch
# across each channel
with pytest.raises(ValueError):
_, output_dict = pre.final_reject_epoch(epo)
assert True
assert isinstance(output_dict, dict)
# get data and metadata parameters
preprocess_params = user_params["preprocess"]
data_params = user_params["load_data"]
write_params = user_params["output_data"]
# get output root and channel type of data
ch_type = data_params["channel-type"]
output_path = write_params["root"]
# overwrite data_params using sys.argv arguments
if len(sys.argv) > 1:
data_params["subjects"] = [sys.argv[1]]
# get set of subjects & tasks to run while omitting existing exceptions
data = load.load_files(data_params)
# for each file in filtered data
for file in data:
# load raw data
eeg_obj = mne_bids.read_raw_bids(file)
outputs = [None] * len(preprocess_params)
# for each pipeline step in user_params, execute with parameters
for idx, (func, params) in enumerate(preprocess_params.items()):
eeg_obj, outputs[idx] = getattr(preprocess, func)(eeg_obj, **params)
# check if this is the fully preprocessed eeg object
final = idx == len(preprocess_params.items()) - 1
write.write_eeg_data(eeg_obj, func, file, ch_type, final, output_path)
