[Bug]No dataset.json after runing prepare_mic.py for Task012_LIDC

[hushunda]

Task: Task012_LIDC
When run nndet_prep, a bug has occurred—— Did not find dataset.json or dataset.yaml .
I think the dataset.yml should be generated when running prepare_mic.py. However, i do not find the generated code in the prepare_mic.py.

[hushunda]

In addition, there are a lot of bugs when processing data of Task012_LIDC

[mibaumgartner]

Dear @hushunda ,

indeed, I forgot to add the dataset.yaml generation to the script (I generated it manually). I'll update the code to include the data, in the meantime you could generate it yourself, here is the info:

task: Task012_LIDC

name: LIDC
target_class:
test_labels: False # manually splitted test set

remove_classes: []
reorder_classes: {}
subtract_one_from_classes: False  # subtract one from all classes in mapping

labels:
    "0": "benign"
    "1": "malignant"

modalities:
    "0": "CT"

dim: 3

Most of the code is adapted from the original script of the medicaldetectiontoolkit (from here) and it worked on the data I received from my colleague. Can you specify in more detail which other bugs you mean?

I can see that the current preparation of the data is also somewhat cumbersome and I hope to provide an updated version with the help of newer tools like pylidc sometime in the future (PRs from the community are highly appreciated as well).