Fix electrode selection in Axona raw recording #1520
Conversation
| cntr = (itetr * elec_per_tetrode) + ielec | ||
| ch_name = f"{itetr + 1}{letters[ielec]}" | ||
| cntr = (itetr * elec_per_tetrode) + ielec + first_channel | ||
| ch_name = "{}{}".format(itetr + active_tetrode_set[0], letters[ielec]) |
There was a problem hiding this comment.
| ch_name = "{}{}".format(itetr + active_tetrode_set[0], letters[ielec]) | |
| ch_name = f"{itetr + active_tetrode_set[0]}{letters[ielec]}" |
we try to keep the codebase on f-strings since they are faster and easier to read :)
| @@ -558,6 +557,20 @@ def get_active_tetrode(self): | |||
| tetrode_id = int(key.strip("collectMask_")) | |||
| active_tetrodes.append(tetrode_id) | |||
| return active_tetrodes | |||
|
|
|||
| def get_active_channels(self): | |||
There was a problem hiding this comment.
public or private? what do you think?
There was a problem hiding this comment.
I advice to make things private by default and move away from it when you have a good reason to.
There was a problem hiding this comment.
That's how I was leaning too. I don't see why the user would run this function.
There was a problem hiding this comment.
Yes, I agree with making them private
| chan = self._get_channel_from_tetrode(tetrode) | ||
| active_channels.append(chan) | ||
|
|
||
| return np.concatenate(active_channels) |
There was a problem hiding this comment.
what's the benefit of concat? I need to doublecheck what is chan is it a list of all channels from a tetrode. If that is the case I would call it channels or chans instead so we know it's plural.
There was a problem hiding this comment.
Yes, chan is a list of all channels from a tetrode, so I agree to call it channels or chans.
Since I was using this function only for the _get_analogsignal_chunk function and it needed the list of active channels as an array I returned directly active_channels as an array. But we can also do it in _get_analogsignal_chunk and keep this function consistent with get_active_tetrodes and return directly active_channels
|
@alejoe91 do you know this datalad error? it only happened for one dataset on the 3.9 python vs the 3.11 python runs. It's completely unrelated to this PR, but if you recognize the issue then I won't spend time digging into it. |
|
@zm711 maybe just a random failure, re-triggered to see what happens! |
|
I tried retriggering once, but yeah let's see if maybe it was just a timing issue. EDIT; --Yep okay just a random ci failure. cool cool |
|
Gin data for this? : O |
You mean you want extra gin data? This was a bug in the original implementation just because we didn't know which channel was which. If you use the current gin data with the spikeinterface extractor you'll see that the channels are just 1-16 or 1-32 rather than being meaningful. |
|
Thanks. I did not realize we already had the gin data as I came from the neuroconv trail. |
I'm proposing this PR to fix #1454
@zm711