ClinicaDL Zoo

ClinicaDL Zoo hosts a collection of transforms and models compatible with ClinicaDL.

Contributing to ClinicaDL Zoo will make your work fully compatible with ClinicaDL, and thus fully reproducible. It will also benefit the community.

Getting started

Currently, ClinicaDL 2.0 is in beta mode, so make sure you are on the branch clinicadl_v2 and in the right ClinicaDL environment, and run the command:

poetry install --extras zoo

NOTE: In the future, ClinicaDL Zoo will be downloaded via: pip install clinicadl[zoo]

Then, you can easily access all transforms and models in the zoo, and use them in clinicadl:

>>> from clinicadl_zoo.transforms.config import NanRemovalConfig
>>> from clinicadl.transforms.config import ZNormalizationConfig
>>> from clinicadl.transforms import Transforms

>>> transforms = Transforms(image_transforms=[NanRemovalConfig(), ZNormalizationConfig()])
>>> transforms.get_transforms()[0]
Compose([NanRemoval(nan=0.0, posinf=None, neginf=None), ZNormalization(masking_method=None)])

Contributing

To add your transforms or models to ClinicaDL Zoo, you need to open a Pull Request on this repository. To do this, please follow to relevant tutorial:

• add a transform

Prerequisite

To contribute, you need:

1. a Python environment with ClinicaDL installed in developer mode (i.e. via poetry and not via pip);

2. to clone this repository (ideally, fork it first):

   git clone https://github.com/aramis-lab/clinicadl-zoo.git
   

3. to install some developing tools:

   cd clinicadl-zoo
   pre-commit install