def run(self, basename, roi, cond, smooth=False, filtfile=None, event=False):
# Save here....
table = join_by_underscore(False, basename, roi, cond)
# and reinit write flags.
roicount = 0
mode = 'w'
header = True
# Getting to work, find subjects data
paths = self.data.get_roi_data_paths(roi)
if not event:
metas = self.data.get_metapaths_containing(cond)
else:
metas = self.data.get_RT_metadata_event_paths()
# And decompose it
for path, meta in zip(paths, metas):
roiname = get_roiname(path)
print("\t{0}".format(roiname)) ## ...progress
# If were past the first Ss data, append.
if roicount > 0:
mode = 'a'
header = False
# Get data
targets = csv_to_targets(meta)
X = load_nii(path, clean=True, sparse=False, smooth=smooth)
targets = tr_pad_targets(targets, "TR", X.shape[0], pad=np.nan)
# Preprocess labels
if filtfile is not None:
targets = reprocess_targets(filtfile, targets, np.nan)
assert targets["TR"].shape[0] == X.shape[0], ("target"
"reprocessing is broken")
norm = MinMaxScaler((0,1))
X = norm.fit_transform(X.astype(np.float))
Xcs, csnames, ti_cs = self.spacetime.fit_transform(
X, targets[cond], targets["trialcount"],
self.window, self.tr)
# and write.
dataname = join_by_underscore(False, roiname)
known = []
for Xc, csname, ti in zip(Xcs, csnames, ti_cs):
if not csname in known:
save_tcdf(
name=join_by_underscore(True, table, csname),
X=Xc,
cond=csname,
dataname=dataname,
index=ti.astype(np.int),
header=header,
mode=mode,
float_format="%.{0}f".format(self.nsig))
known.append(csname)
else:
save_tcdf(
name=join_by_underscore(True, table, csname),
X=Xc,
cond=csname,
dataname=dataname,
index=ti.astype(np.int),
header=False,
mode='a',
float_format="%.{0}f".format(self.nsig))
roicount += 1
if __name__ == '__main__':
from wheelerdata.load.fh import FH
from fmrilearn.preprocess.labels import csv_to_targets
from fmrilearn.load import load_meta
from fmrilearn.load import load_nii
from fmrilearn.preprocess.labels import filter_targets
data = FH()
metas = data.get_metapaths_containing('rt')
targets = csv_to_targets(metas[0])
paths = data.get_roi_data_paths('Insula')
X = load_nii(paths[0], clean=True, sparse=False, smooth=False)
scaler = MinMaxScaler(feature_range=(0, 1))
X = scaler.fit_transform(X.astype(np.float))
X = X[targets['TR'], :]
X = X.mean(1)[:, np.newaxis]
y = targets['rt']
tc = targets['trialcount']
Xfir, flfir = fir(X, y, tc, 20, 1.5)
#Xeva, fleva = eva(X, y, tc, 11, 1.5)
import matplotlib.pyplot as plt
fig = plt.figure()
# ax1 = fig.add_subplot(211)
# ax1.plot(Xeva)
if __name__ == '__main__':
from wheelerdata.load.fh import FH
from fmrilearn.preprocess.labels import csv_to_targets
from fmrilearn.load import load_meta
from fmrilearn.load import load_nii
from fmrilearn.preprocess.labels import filter_targets
data = FH()
metas = data.get_metapaths_containing('rt')
targets = csv_to_targets(metas[0])
paths = data.get_roi_data_paths('Insula')
X = load_nii(paths[0], clean=True, sparse=False, smooth=False)
scaler = MinMaxScaler(feature_range=(0, 1))
X = scaler.fit_transform(X.astype(np.float))
X = X[targets['TR'],:]
X = X.mean(1)[:,np.newaxis]
y = targets['rt']
tc = targets['trialcount']
Xfir, flfir = fir(X, y, tc, 20, 1.5)
#Xeva, fleva = eva(X, y, tc, 11, 1.5)
import matplotlib.pyplot as plt
fig = plt.figure()
# ax1 = fig.add_subplot(211)
# ax1.plot(Xeva)