SpikeGLX is loading `XD` channels both as events and signal

[h-mayorquin]

from neo.rawio import SpikeGLXRawIO
from pathlib import Path

folder_path = Path("/home/heberto/neuroconv_testing_data/ephy_testing_data/spikeglx/DigitalChannelTest_g0")
rawio = SpikeGLXRawIO(folder_path)
rawio.parse_header()
rawio

This is the digital only example on the gin tests:

So for SpikeGLX the XD digital channels (non-mux digital channel) is used to form a memmap and then extract every bit as a line:

python-neo/neo/rawio/spikeglxrawio.py

Lines 341 to 365 in ab821e7

	def _get_event_timestamps(self, block_index, seg_index, event_channel_index, t_start=None, t_stop=None):
	timestamps, durations, labels = [], None, []
	info = self.signals_info_dict[0, "nidq"] # There are no events that are not in the nidq stream
	dig_ch = info["digital_channels"]
	if len(dig_ch) > 0:
	event_data = self._events_memmap
	channel = dig_ch[event_channel_index]
	ch_idx = 7 - int(channel[2:]) # They are in the reverse order
	this_stream = event_data[:, ch_idx]
	this_rising = np.where(np.diff(this_stream) == 1)[0] + 1
	this_falling = (
	np.where(np.diff(this_stream) == 255)[0] + 1
	) # because the data is in unsigned 8 bit, -1 = 255!
	if len(this_rising) > 0:
	timestamps.extend(this_rising)
	labels.extend([f"{channel} ON"] * len(this_rising))
	if len(this_falling) > 0:
	timestamps.extend(this_falling)
	labels.extend([f"{channel} OFF"] * len(this_falling))
	timestamps = np.asarray(timestamps)
	if len(labels) == 0:
	labels = np.asarray(labels, dtype="U1")
	else:
	labels = np.asarray(labels)
	return timestamps, durations, labels

But it seems that we are also loading the whole channel in the signal.

I think this is OK and probably desired but I wanted to bring your attention that we are providing two ways of accessing the data here in case the discussion comes for another similar case.

[zm711]

I think this came up because some people want to treat this as an event and some people want the raw signal to process themselves.
If we do add a get digitalchunk then I think the water gets even muddier between signal, event, and digital.

[h-mayorquin]

Right, previous discussion here:

#1668 (comment)