[QUESTION] BraTS24_2 Zenodo file not found

[KSchmidtACR]

Your question

Hi I've been using BraTS24_2 in a pipeline without issue. Today when I try to run my pipeline, I'm seeing a Zenodo issue. When I search Zenodo for the record id in the Traceback, the record is not found. Any idea what is going on? Traceback:

Unknown issue encountered: Traceback (most recent call last): File "/gable-trial/orchestrate_gable.py", line 239, in run_orchestrator run_brainles_segmentor(preprocessed_filepaths, os.path.join(base_dir, brats_seg_dir, patient, patient + '_' + sub, 'BraTS24_2')) File "/gable-trial/gable_brainles_segmentation.py", line 31, in run_brainles_segmentor segmenter.infer_single( File "/home/kschmidt/miniconda3/envs/new-brats/lib/python3.10/site-packages/brats/core/segmentation_algorithms.py", line 179, in infer_single self._infer_single( File "/home/kschmidt/miniconda3/envs/new-brats/lib/python3.10/site-packages/brats/core/brats_algorithm.py", line 188, in _infer_single run_container( File "/home/kschmidt/miniconda3/envs/new-brats/lib/python3.10/site-packages/brats/core/docker.py", line 404, in run_container additional_files_path = _get_additional_files_path(algorithm) File "/home/kschmidt/miniconda3/envs/new-brats/lib/python3.10/site-packages/brats/core/docker.py", line 141, in _get_additional_files_path return check_additional_files_path( File "/home/kschmidt/miniconda3/envs/new-brats/lib/python3.10/site-packages/brats/utils/zenodo.py", line 37, in check_additional_files_path zenodo_response = _get_zenodo_metadata_and_archive_url(record_id=record_id) File "/home/kschmidt/miniconda3/envs/new-brats/lib/python3.10/site-packages/brats/utils/zenodo.py", line 131, in _get_zenodo_metadata_and_archive_url raise ZenodoException(error_msg) brats.utils.exceptions.ZenodoException: Can not find additional files for record_id '14413387' on Zenodo (response.status_code=403). Exiting...

[KSchmidtACR]

Is there a way to store all necessary files locally and run this algorithm in "offline mode" to avoid reaching out to Zenodo?

[neuronflow]

Does the issue persist? Might be an uptime issue with zenodo.

You could run the pipeline once on a machine with internet connection so it will download the model weights/images and helper files, then copy your Python and docker environment.

[KSchmidtACR]

Thanks for responding.

The issue persisted through this morning, but I just tried again and it is running fine. I'm not sure if this is a coincidence or not, but the server running my pipeline was just upgraded from Ubuntu 22.04 to 24.04.

I have successfully run this algorithm numerous times with an internet connection, so it's confusing to me why I wouldn't have had all the necessary files. Plus I had an internet connection when experiencing the Zenodo errors as well.

[neuronflow]

Great that is working again.

From my understanding the algorithms should not require an internet connection once the files are downloaded, correct @MarcelRosier ?