Fix electrode selection in Axona raw recording #1520
Conversation
zm711
left a comment
zm711
left a comment
There was a problem hiding this comment.
This looks pretty good to me. The failing test seems to be a ci issue and not related to this PR so I'll try to figure that out. Let me know what you think about the comments :)
neo/rawio/axonarawio.py
Outdated
| 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 :)
neo/rawio/axonarawio.py
Outdated
| 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. |
|
@letiziasignorelli just wanted to see if you'd have bandwidth to work on this before the end of the year :) ? Otherwise I'll change the tag to |
|
@zm711 I replied to all comments, and from my side I was just waiting for approval :). Is there anything I should still work on/change? |
|
Did you forget to push the changes? I don't see any of the changes--(I see your comments, but not a commit where you have the changes). Once they are pushed here then the last thing is just resolving the conflict in the author doc :) |
|
Okok you're right, I'll push the changes in the next few days. Thanks :) |
| elif isinstance(channel_indexes, slice): | ||
| channel_indexes_all = [i for i in range(bin_dict["num_channels"])] | ||
| channel_indexes = channel_indexes_all[channel_indexes] | ||
| channel_indexes = self._get_active_channels() |
There was a problem hiding this comment.
@h-mayorquin do you want to doublecheck this. I'm just re-reading and I'm thinking of the case of a channel slice so would love another pair of eyes on the logic here.
There was a problem hiding this comment.
I am confused on why this function does not take any input, I would expect this to use the slice (channel_indexes) that it was passed?
There was a problem hiding this comment.
@h-mayorquin it was this part of the code which you indicated your confusion and I too wasn't completely sure.
There was a problem hiding this comment.
Seems OK to me. I re-read and it is just that they are getting it from the class attributes
|
Hi @letiziasignorelli. |
|
@samuelgarcia just so you know the background (although I really appreciate the extra read :)) In our original implementation we were just giving the channels arbitrary names like But I'm just trying to make sure the slicing of channels when some aren't turned on will be correct. |
|
Maybe we should merge this? It's been six months. |
|
Let's ping Sam for tomorrow and if he doesn't respond then I can fix the merge conflict. You want to give this one more careful read to make sure you understand the slicing logic? |
|
I remember reading it back then. I can take a second look if you pint-point the specific part that you don't feel confident about. |
|
OK lets merge. |
|
Thanks @letiziasignorelli and sorry for the delay! |
5 participants
I'm proposing this PR to fix #1454
@zm711