def _get_box_bounds(bunchs, channel_ids):
cp = {}
for d in bunchs:
cp.update({cid: pos
for cid, pos in zip(d.channel_ids,
d.channel_positions)})
box_pos = np.stack([cp[cid] for cid in channel_ids])
return _get_boxes(box_pos, margin=.1)
def _get_box_bounds(bunchs, channel_ids):
cp = {}
for d in bunchs:
cp.update({cid: pos
for cid, pos in zip(d.channel_ids,
d.channel_positions)})
box_pos = np.stack([cp[cid] for cid in channel_ids])
return _get_boxes(box_pos, margin=.1)
def _iter_channel(positions):
size = 100
margin = 5
boxes = _get_boxes(positions, keep_aspect_ratio=False)
xmin, ymin = boxes[:, :2].min(axis=0)
xmax, ymax = boxes[:, 2:].max(axis=0)
x = boxes[:, [0, 2]].mean(axis=1)
y = - boxes[:, [1, 3]].mean(axis=1)
positions = np.c_[x, y]
tr = [margin, margin, size - margin, size - margin]
positions = Range(NDC, tr).apply(positions)
for x, y in positions:
yield x, y
def _iter_channel(positions):
size = 100
margin = 5
boxes = _get_boxes(positions, keep_aspect_ratio=False)
xmin, ymin = boxes[:, :2].min(axis=0)
xmax, ymax = boxes[:, 2:].max(axis=0)
x = boxes[:, [0, 2]].mean(axis=1)
y = - boxes[:, [1, 3]].mean(axis=1)
positions = np.c_[x, y]
tr = [margin, margin, size - margin, size - margin]
positions = Range(NDC, tr).apply(positions)
for x, y in positions:
yield x, y
def __init__(self,
waveforms=None,
channel_positions=None,
channel_order=None,
best_channels=None,
**kwargs):
self._key_pressed = None
self._overlap = False
self.do_zoom_on_channels = True
self.do_show_labels = False
self.filtered_tags = ()
self.best_channels = best_channels or (lambda clusters: [])
# Channel positions and n_channels.
assert channel_positions is not None
self.channel_positions = np.asarray(channel_positions)
self.n_channels = self.channel_positions.shape[0]
# Initialize the view.
box_bounds = _get_boxes(channel_positions, margin=.1)
super(WaveformView, self).__init__(layout='boxed',
box_bounds=box_bounds,
**kwargs)
self.events.add(channel_click=ChannelClick)
# Box and probe scaling.
self._box_scaling = np.ones(2)
self._probe_scaling = np.ones(2)
# Make a copy of the initial box pos and size. We'll apply the scaling
# to these quantities.
self.box_pos = np.array(self.boxed.box_pos)
self.box_size = np.array(self.boxed.box_size)
self._update_boxes()
# Data: functions cluster_id => waveforms.
self.waveforms = waveforms
# Channel positions.
assert channel_positions.shape == (self.n_channels, 2)
self.channel_positions = channel_positions
channel_order = (channel_order if channel_order is not None
else np.arange(self.n_channels))
assert channel_order.shape == (self.n_channels,)
self.channel_order = channel_order
def __init__(self,
waveforms=None,
channel_positions=None,
waveform_lims=None,
best_channels=None,
**kwargs):
self._key_pressed = None
self._overlap = False
self.do_zoom_on_channels = True
self.data_index = 0
self.best_channels = best_channels or (lambda clusters: [])
# Channel positions and n_channels.
assert channel_positions is not None
self.channel_positions = np.asarray(channel_positions)
self.n_channels = self.channel_positions.shape[0]
# Initialize the view.
box_bounds = _get_boxes(channel_positions)
super(WaveformView, self).__init__(layout='boxed',
box_bounds=box_bounds,
**kwargs)
self.events.add(channel_click=ChannelClick)
# Box and probe scaling.
self._box_scaling = np.ones(2)
self._probe_scaling = np.ones(2)
# Make a copy of the initial box pos and size. We'll apply the scaling
# to these quantities.
self.box_pos = np.array(self.boxed.box_pos)
self.box_size = np.array(self.boxed.box_size)
self._update_boxes()
# Data: functions cluster_id => waveforms.
self.waveforms = waveforms
# Waveform normalization.
assert len(waveform_lims) == 2
self.data_bounds = [-1, waveform_lims[0], +1, waveform_lims[1]]
# Channel positions.
assert channel_positions.shape == (self.n_channels, 2)
self.channel_positions = channel_positions