Problem trying to run prepare-data

[baptisteCD]

Hi,
I am trying to run in an interactive session the following:

clinicadl prepare-data image --custom_suffix 'graymatter_space-Ixi549Space_modulated-on_probability.nii.gz' --extract_json MOODS_2024 ${CAPS_DIR} custom

I get the following error, which is not very clear.

Traceback (most recent call last):
  File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/clinicadlEnv/bin/clinicadl", line 5, in <module>
    from clinicadl.cmdline import cli
  File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/clinicadlEnv/lib/python3.8/site-packages/clinicadl/cmdline.py", line 4, in <module>
    from clinicadl.generate.generate_cli import cli as generate_cli
  File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/clinicadlEnv/lib/python3.8/site-packages/clinicadl/generate/generate_cli.py", line 3, in <module>
    from clinicadl.generate.generate_artifacts_cli import cli as generate_artifacts_cli
  File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/clinicadlEnv/lib/python3.8/site-packages/clinicadl/generate/generate_artifacts_cli.py", line 3, in <module>
    from clinicadl.generate.generate import generate_artifacts_dataset
  File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/clinicadlEnv/lib/python3.8/site-packages/clinicadl/generate/generate.py", line 32, in <module>
    from .generate_utils import (
  File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/clinicadlEnv/lib/python3.8/site-packages/clinicadl/generate/generate_utils.py", line 157, in <module>
    def generate_scales(size: str) -> tuple[float, float]:
TypeError: 'type' object is not subscriptable

Can you please point me to what could be the problem?

Also, how is clinicaDL finding the image based on the "graymatter_space-Ixi549Space_modulated-on_probability.nii.gz"? What it I have two projections to several dartel templates in the caps? (note that this is not the case, at the time of the error).

Thanks

[NicolasGensollen]

Hello @baptisteCD

For the error that you have, it is due to your version of Python (3.8). The support for typing using builtins like list has been introduced in Python 3.9. In addition, the official support of Python 3.8 is ending this month, so it is strongly advised to update to a more recent version of Python.

If you are using Conda, you can do:

conda create -n my_env python=3.11
conda activate my_env
pip install clinicadl

Hope this helps !

[thibaultdvx]

Hi @baptisteCD, regarding your last question, could you please give us an insight of how your caps folder is organised?

With prepare-data, ClinicaDL will look for all the nifti files that match custom_suffix inside your sub-x/session-y folder. You will have an error if two files match the suffix in the same sub-x/session-y folder, even if these two files are in different subfolders (let's say sub-x/session-y/projection_1 and sub-x/session-y/projection_2).
If the projections you mentioned don't have the same suffix, you just need to run prepare-data twice, once with each suffix. Is it your case?

[baptisteCD]

Hi @thibaultdvx ,
I have used clinica (t1 volume) with two different dartel templates. My folder, for one individual looks like:

/caps/subjects/sub-002/ses-M000/t1/spm/dartel/
-> group-moodsAll/{custom_suffix}
-> group-ukb1000/{custom_suffix}

So I should probably rename one of the outputs, for clinicaDL to handle this case
Cheers
B

[baptisteCD]

After a lot of debugging, I now have another cryptic error when trying to run:

clinicadl predict ${maps_dir} "ResNet3D_weighted_epoch50_dropout0.1_MOODS_prediction" --participants_tsv ${BIDS_DIR}/participants_diag.tsv --caps_directory ${CAPS_DIR} --no_labels --no-gpu

[rank0]: Traceback (most recent call last):
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/bin/clinicadl", line 8, in <module>
[rank0]:     sys.exit(cli())
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/click/core.py", line 1157, in __call__
[rank0]:     return self.main(*args, **kwargs)
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/click/core.py", line 1078, in main
[rank0]:     rv = self.invoke(ctx)
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/click/core.py", line 1688, in invoke
[rank0]:     return _process_result(sub_ctx.command.invoke(sub_ctx))
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/click/core.py", line 1434, in invoke
[rank0]:     return ctx.invoke(self.callback, **ctx.params)
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/click/core.py", line 783, in invoke
[rank0]:     return __callback(*args, **kwargs)
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/predict/predict_cli.py", line 128, in cli
[rank0]:     predict(
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/predict/predict.py", line 65, in predict
[rank0]:     maps_manager.predict(
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/utils/maps_manager/maps_manager.py", line 366, in predict
[rank0]:     data_test = return_dataset(
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/utils/caps_dataset/data.py", line 787, in return_dataset
[rank0]:     return CapsDatasetImage(
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/utils/caps_dataset/data.py", line 330, in __init__
[rank0]:     super().__init__(
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/utils/caps_dataset/data.py", line 88, in __init__
[rank0]:     self.size = self[0]["image"].size()
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/utils/caps_dataset/data.py", line 349, in __getitem__
[rank0]:     image_path = self._get_image_path(participant, session, cohort)
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/utils/caps_dataset/data.py", line 169, in _get_image_path
[rank0]:     results = clinicadl_file_reader(
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/utils/clinica_utils.py", line 1058, in clinicadl_file_reader
[rank0]:     results, errors_encountered = file_reader(
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/utils/clinica_utils.py", line 1118, in _read_files_sequential
[rank0]:     find_sub_ses_pattern_path(
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/utils/clinica_utils.py", line 626, in find_sub_ses_pattern_path
[rank0]:     if _are_multiple_runs(current_glob_found):
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/utils/clinica_utils.py", line 668, in _are_multiple_runs
[rank0]:     if len(files) < 2 or any(["_run-" not in f.name for f in files]):
[rank0]:   File "/network/lustre/iss02/home/baptiste.couvy/.conda/envs/my_env/lib/python3.10/site-packages/clinicadl/utils/clinica_utils.py", line 668, in <listcomp>
[rank0]:     if len(files) < 2 or any(["_run-" not in f.name for f in files]):
[rank0]: AttributeError: 'str' object has no attribute 'name'

[thibaultdvx]

Hi @baptisteCD, I have the same error when two files match custom_suffix in a subject/session folder. Are you sure this is not your case?

[baptisteCD]

Hi @thibaultdvx, I think I passed that stage. I renamed the files from clinica so the suffix is unique. I get the error when trying to run clinicadl predict. There is only one .pt (tensor) file per individual.

[thibaultdvx]

Hi @baptisteCD, it seems that you found an error in the expression any(["_run-" not in f.name for f in files])(penultimate line of your error message). We will deal with that error.

The question that remains is why you don't have len(files) < 2, which means that at least two files match the patten stores in you maps.json (probably the custom suffix graymatter_space-Ixi549Space_modulated-on_probability.nii.gz in your case).

You need to have only one .pt file per individual but also one .nii.gz file per individual. If both these conditions are satisfied, we run out of ideas and we would need the maps.json file store in your maps folder, as well as a description of your caps structure (use for example the command tree in your terminal).

Sorry for the trouble!