spikeinterface: Incorrect 'parent_recording' arg for multiprocessing when skipping kilosort 2.5 preprocessing
Hi everyone !
I was trying to launch kilosort 2.5 with skip_kilosort_preprocessing = True, which will use the TracePaddedRecording class, which have an parent_recording key in the __init__ part.
I got the following error :
File "<string>", line 1, in <module>
File "C:\Users\DeepLabCut\Anaconda3\envs\spykeline\lib\multiprocessing\spawn.py", line 116, in spawn_main
exitcode = _main(fd, parent_sentinel)
File "C:\Users\DeepLabCut\Anaconda3\envs\spykeline\lib\multiprocessing\spawn.py", line 126, in _main
self = reduction.pickle.load(from_parent)
File "C:\Users\DeepLabCut\Anaconda3\envs\spykeline\lib\site-packages\spikeinterface\core\base.py", line 529, in from_dict
extractor = _load_extractor_from_dict(dictionary)
File "C:\Users\DeepLabCut\Anaconda3\envs\spykeline\lib\site-packages\spikeinterface\core\base.py", line 1135, in _load_extractor_from_dict
extractor = extractor_class(**new_kwargs)
TypeError: __init__() got an unexpected keyword argument 'parent_recording'
Is it something wrong on my side or should we change the parent_recording key?
Thanks !!
About this issue
- Original URL
- State: open
- Created 2 years ago
- Comments: 16 (16 by maintainers)
I thought we left those for backward compatability because we thought they wouldn’t cause issues. But clearly that is not the case. I know we discussed which ones to changes and which to leave the same at Edinburgh, but it was too long ago now.
Up to you @alejoe91 a global fix is probably safer just in case we do have older files floating around or other preprocessing with the old naming. I could try to track down where we are incorrectly writing the json if you think that would be better.
Thanks @OlivierPeron we will work it on shortly 😃
It is always worth trying the most recent (because I can’t remember when we switched form parent_recording to just recording. But I would try my other thing first and see what happens.
Could you try loading the file after transferring just the intan file to a new folder. So we make sure there isn’t any cache of json files that might be providing the wrong name. If the error still occurs in a clean folder with just the intan file then it’s a bug for us. If that works then it means an old file was confusing things.