def join_segments(time_segments): from mypy.utils import group # check which should be joined join_segments = (time_segments[:-1, 1] - time_segments[1:, 0]) == 0 segment_groups = group(join_segments) segment_groups[:, 1] += 1 # join segments final_segments = np.vstack([time_segments[segment_groups[:, 0], 0], time_segments[segment_groups[:, 1], 1]]).T # we missed single-segments, search for them prev = 0 missed_segments = list() for row in segment_groups: if not row[0] == prev: missed_segments.extend(list(range(prev + 1, row[0]))) prev = row[1] if prev < len(time_segments): missed_segments.extend(list(range(prev + 1, len(time_segments)))) # append missed single segments to multi-segments and sort final_segments = np.vstack( [final_segments, time_segments[missed_segments, :]]) final_segments = final_segments[final_segments[:, 0].argsort(), :] return final_segments
def join_segments(time_segments):
from mypy.utils import group
# check which should be joined
join_segments = (time_segments[:-1, 1] - time_segments[1:, 0]) == 0
segment_groups = group(join_segments)
segment_groups[:, 1] += 1
# join segments
final_segments = np.vstack([
time_segments[segment_groups[:, 0], 0],
time_segments[segment_groups[:, 1], 1]
]).T
# we missed single-segments, search for them
prev = 0
missed_segments = list()
for row in segment_groups:
if not row[0] == prev:
missed_segments.extend(list(range(prev + 1, row[0])))
prev = row[1]
if prev < len(time_segments):
missed_segments.extend(list(range(prev + 1, len(time_segments))))
# append missed single segments to multi-segments and sort
final_segments = np.vstack(
[final_segments, time_segments[missed_segments, :]])
final_segments = final_segments[final_segments[:, 0].argsort(), :]
return final_segments
def _to_annotation(self, bad_bool, description='BAD_'): import mne groups = group(bad_bool) if len(groups) > 0: onset = groups[:, 0] * (self.step / self.sfreq) duration = (self.window / self.sfreq) + np.diff(groups).ravel() * (self.step / self.sfreq) return mne.Annotations(onset, duration, [description] * groups.shape[0]) else: return mne.Annotations([], [], [])
def apply_artifacts_to_tfr(tfr, artif, orig_time, fillval=np.nan): '''Fill tfr data with `fillval` according to `artif`. Operates in-place.''' n_items, n_samples = artif.shape bad_inds = np.any(artif, axis=1).nonzero()[0] for bad in bad_inds: limits = group(artif[bad, :]) for lim in limits: slc = find_range(tfr.times, list(orig_time[lim])) tfr.data[bad, :, :, slc] = fillval return tfr
def _to_annotation(self, bad_bool, description='BAD_'):
import mne
groups = group(bad_bool)
if len(groups) > 0:
onset = groups[:, 0] * (self.step / self.sfreq)
duration = (self.window / self.sfreq
) + np.diff(groups).ravel() * (self.step / self.sfreq)
return mne.Annotations(onset, duration,
[description] * groups.shape[0])
else:
return mne.Annotations([], [], [])
def apply_artifacts_to_tfr(tfr, artif, orig_time, fillval=np.nan):
'''Fill tfr data with `fillval` according to `artif`.
Operates in-place.'''
n_items, n_samples = artif.shape
bad_inds = np.any(artif, axis=1).nonzero()[0]
for bad in bad_inds:
limits = group(artif[bad, :])
for lim in limits:
slc = find_range(tfr.times, list(orig_time[lim]))
tfr.data[bad, :, :, slc] = fillval
return tfr
def extend_bads(rej_win, extend, copy=True): '''Extend rejection periods by specfied number of samples.''' if copy: rej_win = rej_win.copy() n_samples = rej_win.shape[1] bad_inds = np.any(rej_win, axis=1).nonzero()[0] for bad in bad_inds: slices = group(rej_win[bad, :], return_slice=True) for slc in slices: slc = extend_slice(slc, extend, n_samples) rej_win[bad, slc] = True return rej_win
def extend_bads(rej_win, extend, copy=True):
'''Extend rejection periods by specfied number of samples.'''
if copy:
rej_win = rej_win.copy()
n_samples = rej_win.shape[1]
bad_inds = np.any(rej_win, axis=1).nonzero()[0]
for bad in bad_inds:
slices = group(rej_win[bad, :], return_slice=True)
for slc in slices:
slc = extend_slice(slc, extend, n_samples)
rej_win[bad, slc] = True
return rej_win
def highlight(x_values,
which_highligh,
kind='patch',
color=None,
alpha=0.3,
axis=None,
level=0.04,
height=0.03):
'''Highlight ranges along x axis.
Parameters
----------
x_values : numpy array
Values specifying x axis points along which which_highligh operates.
which_highligh : numpy array of bool
Boolean values - each entry specifies whether corresponding value for
x_values belongs to highlighted ranges.
'''
from matplotlib.patches import Rectangle
from mypy.utils import group
if color is None:
color = 'orange' if kind == 'patch' else 'k'
axis = plt.gca() if axis is None else axis
ylims = axis.get_ylim()
y_rng = np.diff(ylims)
hlf_dist = np.diff(x_values).mean() / 2
grp = group(which_highligh, return_slice=True)
for slc in grp:
this_x = x_values[slc]
start = this_x[0] - hlf_dist
length = np.diff(this_x[[0, -1]]) + hlf_dist * 2
ptch = Rectangle((start, ylims[0]),
length,
y_rng,
lw=0,
facecolor=color,
alpha=alpha)
axis.add_patch(ptch)