def get_epochs_and_cov(X, y, window=500):
"""return epochs from array."""
raw_train = toMNE(X, y)
picks = range(len(getChannelNames()))
events = list()
events_id = dict()
for j, eid in enumerate(getEventNames()):
tmp = find_events(raw_train, stim_channel=eid, verbose=False)
tmp[:, -1] = j + 1
events.append(tmp)
events_id[eid] = j + 1
# concatenate and sort events
events = np.concatenate(events, axis=0)
order_ev = np.argsort(events[:, 0])
events = events[order_ev]
epochs = Epochs(raw_train,
events,
events_id,
tmin=-(window / 500.0) + 1 / 500.0 + 0.150,
tmax=0.150,
proj=False,
picks=picks,
baseline=None,
preload=True,
add_eeg_ref=False,
verbose=False)
cov_signal = compute_raw_data_covariance(raw_train, verbose=False)
return epochs, cov_signal
def get_epochs_and_cov(X, y, window=500):
"""return epochs from array."""
raw_train = toMNE(X, y)
picks = range(len(getChannelNames()))
events = list()
events_id = dict()
for j, eid in enumerate(getEventNames()):
tmp = find_events(raw_train, stim_channel=eid, verbose=False)
tmp[:, -1] = j + 1
events.append(tmp)
events_id[eid] = j + 1
# concatenate and sort events
events = np.concatenate(events, axis=0)
order_ev = np.argsort(events[:, 0])
events = events[order_ev]
epochs = Epochs(raw_train, events, events_id,
tmin=-(window / 500.0) + 1 / 500.0 + 0.150,
tmax=0.150, proj=False, picks=picks, baseline=None,
preload=True, add_eeg_ref=False, verbose=False)
cov_signal = compute_raw_data_covariance(raw_train, verbose=False)
return epochs, cov_signal
def _fit(self, X, y):
"""fit and return epochs."""
epochs, cov_signal = get_epochs_and_cov(X, y, self.window)
xd = Xdawn(n_components=self.nfilters, signal_cov=cov_signal,
correct_overlap=False)
xd.fit(epochs)
P = []
for eid in getEventNames():
P.append(np.dot(xd.filters_[eid][:, 0:self.nfilters].T,
xd.evokeds_[eid].data))
self.P = np.concatenate(P, axis=0)
self.labels_train = epochs.events[:, -1]
return epochs
def _fit(self, X, y):
"""fit and return epochs."""
epochs, cov_signal = get_epochs_and_cov(X, y, self.window)
xd = Xdawn(n_components=self.nfilters,
signal_cov=cov_signal,
correct_overlap=False)
xd.fit(epochs)
P = []
for eid in getEventNames():
P.append(
np.dot(xd.filters_[eid][:, 0:self.nfilters].T,
xd.evokeds_[eid].data))
self.P = np.concatenate(P, axis=0)
self.labels_train = epochs.events[:, -1]
return epochs
def toMNE(X, y=None):
"""Tranform array into MNE for epoching."""
ch_names = deepcopy(getChannelNames())
montage = read_montage('standard_1005', ch_names)
ch_type = ['eeg']*len(ch_names)
data = X.T
if y is not None:
y = y.transpose()
ch_type.extend(['stim']*N_EVENTS)
event_names = getEventNames()
ch_names.extend(event_names)
# concatenate event file and data
data = np.concatenate((data, y))
info = create_info(ch_names, sfreq=128.0, ch_types=ch_type,
montage=montage)
raw = RawArray(data, info, verbose=False)
return raw
def toMNE(X, y=None):
"""Tranform array into MNE for epoching."""
ch_names = getChannelNames()
montage = read_montage('standard_1005', ch_names)
ch_type = ['eeg']*len(ch_names)
data = X.T
if y is not None:
y = y.transpose()
ch_type.extend(['stim']*6)
event_names = getEventNames()
ch_names.extend(event_names)
# concatenate event file and data
data = np.concatenate((data, y))
info = create_info(ch_names, sfreq=500.0, ch_types=ch_type,
montage=montage)
raw = RawArray(data, info, verbose=False)
return raw
def _fit(self, X, y):
"""fit and return epochs."""
epochs, cov_signal = get_epochs_and_cov(X, y, self.window)
# pdb.set_trace()
# TODO: Try to understand what is Xdawn (something that improve SSRN by spatial information?)
xd = Xdawn(n_components=self.nfilters, signal_cov=cov_signal,
correct_overlap=False)
xd.fit(epochs)
P = []
for eid in getEventNames():
# set_breakpoint('xd info in getEVetnNames for-lop');
# pdb.set_trace()
P.append(np.dot(xd.filters_[eid][:, 0:self.nfilters].T,
xd.evokeds_[eid].data))
self.P = np.concatenate(P, axis=0)
self.labels_train = epochs.events[:, -1]
# pdb.set_trace()
return epochs
from sklearn.metrics import roc_auc_score
from preprocessing.aux import getEventNames, load_raw_data
from multiprocessing import Pool
from read_adapter import subjects_path_list
from eeg_config import subjects
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('--n_subjects', type=int, action='store', default=2)
args, unknown = parser.parse_known_args()
subjects = range(1, args.n_subjects + 1)
cols = getEventNames()
def _from_yaml_to_func(method, params):
"""go from yaml to method.
Need to be here for accesing local variables.
"""
prm = dict()
if params is not None:
for key, val in params.iteritems():
prm[key] = eval(str(val))
return eval(method)(**prm)
def doCols(col):
ensemble = yml['Model'][modelName]['ensemble']
addSubjectID = True if 'addSubjectID' in yml.keys() else False
mode = sys.argv[2]
if mode == 'val':
test = False
elif mode == 'test':
test = True
else:
raise('Invalid mode. Please specify either val or test')
print('Running %s in mode %s, will be saved in %s' % (modelName,mode,fileName))
######
cols = getEventNames()
ids = np.load('../infos_test.npy')
subjects_test = ids[:, 1]
series_test = ids[:, 2]
ids = ids[:, 0]
labels = np.load('../infos_val.npy')
subjects = labels[:, -2]
series = labels[:, -1]
labels = labels[:, :-2]
allCols = range(len(cols))
# ## loading prediction ###
files = getLvl1ModelList()
