For starting contributing, the submodules must also be cloned.

git clone ... --recurse-submodules (<path>)

For updating Raidionics after backend update, performing a git pull inside raidionics_rads_lib/ is necessary.

pip install -e git+https://github.com/dbouget/raidionics_rads_lib.git@master#egg=raidionicsrads
pip install -e git+https://github.com/dbouget/raidionics_seg_lib.git@master#egg=raidionicsseg

Including the rads/seg backend the first time in a new project.

git submodule add -b submodule_seg https://github.com/dbouget/raidionics_rads_lib.git
git submodule update --init --recursive

Use the requirements.txt file to create a virtual environment with the required libraries.

virtualenv -p python3 venv  
source venv/bin/activate  
pip install -r requirements.txt  
deactivate  

The command line input parameters are:

• -g [--use_gui]: Must be set to 0 to disable the gui, otherwise 1.
• -i [--input_filename]: Complete path to the MRI volume to process.
• (optional) -s [--input_tumor_segmentation_filename]: Complete path to the corresponding tumor mask, to avoid re-segmentation.
• -o [--output_folder]: Main destination directory. A unique timestamped folder will be created inside for each run.
• -d [--gpu_id]: Number of the GPU to use for the segmentation task. Set the value to -1 to run on CPU.
• -t [--task]: Process to perform, either segmentation or diagnosis (for generating the standardized report).
• -m [--model_segmentation]: Name of the trained model to use (from the list of automatically downloadable models).

To run segmentation directly from command line, without the use of the GUI, run the following:

source venv/bin/activate  
python main.py -g 0 -i /path/to/volume/T1.nii.gz -o /path/to/output/ -d 0 -t segmentation -m MRI_Meningioma 
deactivate