jeylau
commented
1 year ago Hello @NielsWW, thanks for reporting this bug. I had encountered it about 2 months ago and addressed it with https://github.com/DeepLabCut/napari-deeplabcut/commit/a65f78e578549b02d11fa0dd3afd94e9d6390ce1. Please run pip install -U napari-deeplabcut to make sure you get the latest version and the fix ☺️
NielsWW
Hi,
I would like to point out an issue when new videos are analysed on a Linux-driven machine, but outlier frames are extracted and the labels refined on a Windows machine.
I will first describe the workflow I used to train a model with minimal resources:
For each 'port', I ensured that the project_path was properly set. Now, when I try to save the refined layer, I get an error as the config file cannot be found:
Traceback (most recent call last): File "C:\Users\NielsW\AppData\Local\miniconda3\envs\bradykinesia\lib\site-packages\napari_deeplabcut_widgets.py", line 207, in
lambda: _save_layers_dialog(
File "C:\Users\NielsW\AppData\Local\miniconda3\envs\bradykinesia\lib\site-packages\napari_deeplabcut_widgets.py", line 94, in _save_layers_dialog
self.viewer.layers.save("", selected=True, plugin="napari-deeplabcut")
File "C:\Users\NielsW\AppData\Local\miniconda3\envs\bradykinesia\lib\site-packages\napari\components\layerlist.py", line 476, in save
return save_layers(path, layers, plugin=plugin, _writer=_writer)
File "C:\Users\NielsW\AppData\Local\miniconda3\envs\bradykinesia\lib\site-packages\napari\plugins\io.py", line 229, in save_layers
_written, writer_name = _write_single_layer_with_plugins(
File "C:\Users\NielsW\AppData\Local\miniconda3\envs\bradykinesia\lib\site-packages\napari\plugins\io.py", line 442, in _write_single_layer_with_plugins
written_paths, writer_name = _npe2.write_layers(
File "C:\Users\NielsW\AppData\Local\miniconda3\envs\bradykinesia\lib\site-packages\napari\plugins_npe2.py", line 101, in write_layers
paths, writer = io_utils.write_get_writer(
File "C:\Users\NielsW\AppData\Local\miniconda3\envs\bradykinesia\lib\site-packages\npe2\io_utils.py", line 115, in write_get_writer
return _write(path, layer_data, plugin_name=plugin_name, return_writer=True)
File "C:\Users\NielsW\AppData\Local\miniconda3\envs\bradykinesia\lib\site-packages\npe2\io_utils.py", line 224, in _write
res = writer.exec(args=args)
File "C:\Users\NielsW\AppData\Local\miniconda3\envs\bradykinesia\lib\site-packages\npe2\manifest\utils.py", line 62, in exec
return self.get_callable(_registry)(*args, **kwargs)
File "C:\Users\NielsW\AppData\Local\miniconda3\envs\bradykinesia\lib\site-packages\napari_deeplabcut_writer.py", line 65, in write_hdf
config = _load_config(os.path.join(project_folder, "config.yaml"))
File "C:\Users\NielsW\AppData\Local\miniconda3\envs\bradykinesia\lib\site-packages\napari_deeplabcut_reader.py", line 153, in _load_config
with open(config_path) as file:
FileNotFoundError: [Errno 2] No such file or directory: 'C:/Users/NielsW/Documents/Research Project/Bradykinesia-NielsW-2023-03-08/labeled-data/PD2_4_L_ON\config'
This must be since somewhere the DLCHeader is constructed on Ubuntu and the current method to split the last two directories from the root path relies on the path separators of the currently used OS. I have an easy fix that solves this issue, but I'm not well-versed with GitHub so I'm uncertain how to make a pull request. My proposal is as follows:
in _writer.py, line 65 was: project_folder = root.rsplit(os.sep, 2)[0]
I propose to replace it with: project_folder = os.path.dirname(os.path.dirname(root))
Maybe I'm overlooking some problems this might create, or I did something wrong when trying to port the project. My apologies if this issue is already addressed. I use napari-deeplabcut 0.0.9