cortex-lab / phy

phy: interactive visualization and manual spike sorting of large-scale ephys data
BSD 3-Clause "New" or "Revised" License
314 stars 157 forks source link

segmentation fault #1231

Closed Kayv-cmb closed 11 months ago

Kayv-cmb commented 11 months ago

When I try to open the gui-template with phy template-gui ./params.py nothing happen, I have an error message saying segmentation fault. Could I have some help on how to resolve that ? I tried different dataset, I tried on a cluster with 250GB of RAM, I tried locally on my mac with 16 GB of ram, I have the access right to the file. I am kinda lost on what to do?

full error message :

phy template-gui ./params.py
15:51:40.373 [W] model:603            Unreferenced clusters found in templates (generally not a problem)
15:51:42.821 [W] model:625            Unreferenced clusters found in spike_clusters (generally not a problem)
zsh: segmentation fault  phy template-gui 
zm711 commented 11 months ago

@Kayv-cmb.

Couple questions.

How did you install phy? Could you do

conda list

inside your environment if you installed with anaconda.

then could you run

phy template-gui ./params.py --debug

and copy the output so we can get more resolution of what exactly is calling the seg fault.

Kayv-cmb commented 11 months ago

Thanks for answers,

Before I had installed it with the pip install git+https://github.com/cortex-lab/phy.git

I managed to solve the SegFault by installing through the .yml file. Now I have this issue

QOpenGLFramebufferObject: Unsupported framebuffer format.
QOpenGLFramebufferObject: Framebuffer incomplete, missing attachment.

I think it's because I run it on a cluster. I managed to run it locally so everything is fine.

Also I think phy is not compatible with python 3.12 as they removed imp module

zm711 commented 11 months ago

@Kayv-cmb,

yeah, I'm working on updating both the pip install (the requirements.txt is out of date so I have an open PR to fix that) and the imp for 3.12 is another problem that I have an open PR to fix. So hopefully they are merged soon, but for now the yaml is the best fix. Happy it works -- and yeah I think it could be a cluster issue if it is working fine locally.