Error in bc_waveformShape about bombcell HOT 11 CLOSED

Sweet! This ran smoothly on the dataset I could not run onto before. Thank you very much. Very appreciated! I'm going to close the issue

from bombcell.

Hi Valerio, thanks a lot for your message! It sounds like most likely no peak or no trough is detected in this waveform for some reason. Would you be able to send me the waveform for this unit so I can take a look? You can send me the templates.npy file and the unit number.
Cheers
Julie

from bombcell.

You can find it here:
https://drive.google.com/file/d/1n5cWXJink2VgVOgmDDrnzBdHboYs8Rye/view?usp=sharing
It's unit ('iUnit') 357 (Matlab indexing)
Thank you

from bombcell.

Great thank you!
Sorry I just realized - I also need the spike_templates.npy file. Would you be able to send it over?

from bombcell.

Hi Julie, don't worry and thank you for your help.
https://drive.google.com/drive/folders/12VnfWSyFBPD6FFCn6quv-wI5B3AhaPtF?usp=sharing

from bombcell.

Wonderful, thank you! Could you also tell me which Kilosort version you re using?

I reproduced the error you got and it is due to the template waveform for unit # 357 being all NaNs. This error is probably related to a previously signalled error, where some versions of Kilosort (2.5) can output templateWaveforms that are all NaNs. I still don't understand why this is the case, but, considering the waveforms are NaNs, I believe these units should be skipped and automatically classified as noise by bombcell.

I had previously created a branch to deal with scenario. Since it is now reproduced and affects more than one person, I will integrate these changes in the main branch later today.

from bombcell.

Amazing! Thank you :):)
Yes, you're right, I'm using Kilosort 2.5.
Thank you for the link to the other issue. I'm going to further hunt it down from the kilosort side but in the meantime, I really appreciate you dealing with it.
Super efficient. Thanks again!

from bombcell.

I've now updated bombcell to handle this situation in the main branch (pull request #61 ). Let me know if there are any issues.

As a sanity check, could you tell me, when you look at this unit in phy (it should be unit #359) does any waveform show up?

from bombcell.

Yeah, it's a weird one.

Thank you for updating the code! I'll try it right now

from bombcell.

I get the following error:
Unable to perform assignment because the left and right sides have a different number of elements.

Error in bc_computeAllEphysProperties (line 56)
[ephysProperties.waveformDuration_peakTrough_us(iUnit), ephysProperties.halfWidth_ms(iUnit), ...

Error in bc_ephysPropertiesPipeline (line 11)
ephysProperties = bc_computeAllEphysProperties(spikeTimes_samples, spikeTemplates, templateWaveforms,...

Error in Pipeline_code (line 111)
[ephysProperties, unitClassif] = bc_ephysPropertiesPipeline(ephysKilosortPath, savePath, rerunEP, region);

It happens on the incriminated iUnit :'(

from bombcell.

Sorry I missed this message.
Good point, I hadn't updated the ephysProperties to deal with these NaN waveforms. It should be good now! Let me know if there are any issues.

from bombcell.