dcbia-ortholab / slicerdentalmodelseg Goto Github PK

It is my opinion that "prediction" is too generic of a module name to have in an extension. The module name should likely relate to the extension name.

that's awsom work, if need help, contact me

If I recall, you opted for using the system python environment so you could get newer packages than were available in Python 3.6. With the update of Slicer's (not SlicerSALT's yet) Python to 3.9.10, can you get everything you need into the Slicer python environment?

The GUI of the new version differs from the one explained in the README file.
Could you please provide some information on how to run the updated extension?

Can you share the dataset of teeth?

This is a collection of random small comments or questions that don't necessarily warrant their own issue yet. I am putting individual topics in separate comments. If any of them rise to level of an individual issue, we use the "Reference in new issue" feature to create a new issue for it.

The instantiation failure of the CrownSegmentation module in Slicer is attributed to the following line:

Should I proceed to fix this issue by submitting a PR?

For an example on how to do this see here.

CrownSegmentationcli module is certainly not an example, so update, probably to Segmentation.Advanced

https://github.com/MathieuLeclercq/SlicerJawSegmentation/blob/e89fd4821a6ac19732ac21d7f35385fd3212aba6/CrownSegmentationcli/CrownSegmentationcli.xml#L3

There is nothing in any of the CMake to actually bring the seg_code folder into Slicer. Thus, when I try to run I get

When the extension is actually being installed into / built into Slicer, the source repo won't be there. I don't know exactly how you were adding the extension to slicer, via the Extension Wizard or what, but I think it allowed your sys.path shenanigans to work.

When actually installed closer to how it would be from the extension manager, it can't find the code.

I will fix this and put out PR. There is not a clean slicer way to do this, but we can add some CMake that makes it work.

• In help, separate the different options into different lines. If I am looking for "Number of views" to see what it does, it is hard to find since it is halfway through some paragraph
• In README, same as above, separate the different options into different lines.
• In help say something about clicking the Advanced->Install/Check dependencies button the first time through
• In README say something about clicking the Advanced->Install/Check dependencies button the first time through

If you aren't going to write tests for the prediction module, then remove the test that comes by default with the extension wizard, so it doesn't fail.
https://github.com/MathieuLeclercq/SlicerJawSegmentation/blob/e6cfcd3fad261feecc65b6a01d939213960a3e3f/prediction/prediction.py#L544-L594

This should not be hardcoded to your machine. Likely it should be blank to start.

https://github.com/MathieuLeclercq/SlicerJawSegmentation/blob/8411a7fdc91757fa87010c5312d31a5440bc391d/CrownSegmentation/Resources/UI/CrownSegmentation.ui#L208

It would be good to be able to run without CUDA. I got the following error:

CLI execution failed: 
 
CrownSegmentationcli standard error:

Traceback (most recent call last):
  File "/data/Projects-build/salt-general/SlicerJawSegmentation-build/lib/SlicerSALT-5.1/cli-modules/CrownSegmentationcli.py", line 297, in <module>
    main(sys.argv[1], sys.argv[2], int(sys.argv[3]), int(sys.argv[4]), sys.argv[5],sys.argv[6], sys.argv[7], sys.argv[8])
  File "/data/Projects-build/salt-general/SlicerJawSegmentation-build/lib/SlicerSALT-5.1/cli-modules/CrownSegmentationcli.py", line 145, in main
    model.load_state_dict(torch.load(unet_model))
  File "/home/connor/Projects-build/salt-general/build-5-2/python-install/lib/python3.9/site-packages/torch/serialization.py", line 712, in load
    return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args)
  File "/home/connor/Projects-build/salt-general/build-5-2/python-install/lib/python3.9/site-packages/torch/serialization.py", line 1046, in _load
    result = unpickler.load()
  File "/home/connor/Projects-build/salt-general/build-5-2/python-install/lib/python3.9/site-packages/torch/serialization.py", line 1016, in persistent_load
    load_tensor(dtype, nbytes, key, _maybe_decode_ascii(location))
  File "/home/connor/Projects-build/salt-general/build-5-2/python-install/lib/python3.9/site-packages/torch/serialization.py", line 1001, in load_tensor
    wrap_storage=restore_location(storage, location),
  File "/home/connor/Projects-build/salt-general/build-5-2/python-install/lib/python3.9/site-packages/torch/serialization.py", line 176, in default_restore_location
    result = fn(storage, location)
  File "/home/connor/Projects-build/salt-general/build-5-2/python-install/lib/python3.9/site-packages/torch/serialization.py", line 152, in _cuda_deserialize
    device = validate_cuda_device(location)
  File "/home/connor/Projects-build/salt-general/build-5-2/python-install/lib/python3.9/site-packages/torch/serialization.py", line 136, in validate_cuda_device
    raise RuntimeError('Attempting to deserialize object on a CUDA '
RuntimeError: Attempting to deserialize object on a CUDA device but torch.cuda.is_available() is False. If you are running on a CPU-only machine, please use torch.load with map_location=torch.device('cpu') to map your storages to the CPU.

As part of SlicerSALT, create a google slides tutorial for using the module. As examples, see the following:

SRep: https://bit.ly/3sTEG3H
Groups: https://bit.ly/2WsFiun
SPHARM-PD Generator: https://bit.ly/2Fyn97v

For the install from extension manager slides, use the SRep as basis, since it is up to date with the newer Slicer look.

For now, focus on getting the content of the tutorial (words, screenshots, etc). @bpaniagua will work on getting a google slides template.

You can make a setting save across runs of Slicer similar to this: https://slicer.readthedocs.io/en/latest/developer_guide/script_repository.html#set-a-new-default-module-at-startup

As the title says, remove all the pycache folders which contain the compiled python code. Recommend adding a .gitignore file at the root of the repo with *.pyc in it.