Python KeyError at final segmentation: Manually initialized 'segmentations' in config file

[mrheinnecker]

I ran into the following issue while running: run.py -t nuc_it00 (running nucleus segmentation after annotation of gt_cube)

The prompted Error is: Traceback (most recent call last): File "/home/rheinnec/miniconda3/envs/cebra-em-env/bin/run.py", line 7, in exec(compile(f.read(), file, 'exec')) File "/home/rheinnec/repos/CebraEM/cebra-em/bin/run.py", line 325, in run( File "/home/rheinnec/repos/CebraEM/cebra-em/bin/run.py", line 211, in run if bm not in config_seg['segmentations'].keys(): KeyError: 'segmentations'

I saw that in config/config_nuc_it00.json there is no 'segmentations' entry, which I think the script is looking for. I think actually this is only needed to check if it is already there and otherwise it will be created. So I included a dummy-segmentation entry in the config file (or just copied and renamed another dictonary entry):

"segmentations": { "cpuxxxs": 1 }

After that the segmentation started and is currently running. I am not sure if it is again on my side that there is no entry, but also creating another iteration for nuc (nuc_it01) with init_segmentation.py and linking the ground thruths with link_gt.py did not add an 'segmentations' entry to the config. Maybe i am missing another function call?

Best wishes! Marco

[jhennies]

Can you do a git pull again? This issue should already be fixed with one of the latest merges to main.

[mzugravu]

@mrheinnecker

Hi Marco, I was wondering if you managed to complete the whole CebraEM workflow. I am encountering some issues. I would be really grateful if you could let me know.

Best wishes, Bianca

[mrheinnecker]

Yes i managed to run it.