The repository contains general tools (tools folder) for the analysis of imaging studies, as well as scripts and processing results of specific studies (analysis folder).

Frequently used data and selected processing results are stored as part of the repository in the data folder. Those (larger) files are managed using git-lfs.

Various 'global' settings, such as local project paths, are defined in config.py and can be made available throughout the project using import config.

After normal git clone, the data directory may still be empty. In this case, perform

git lfs pull

If git-lfs is not activated on your system, follow installation instructions at (https://git-lfs.github.com).

If you are only interested in the final statistical analyses, without prior image processing, you can use the Jupyter notebooks in the relvant analysis subfolders. These use processing results from the data folder.

The file environment_notebook_analysis.yml defines a minimum CONDA environment for running the Jupyter notebooks included in this project.

Steps for set-up:

• Go to the root directory of the project project-root
• The command

  conda env create -f environment_notebook_analysis.yml
  

  will create a CONDA environment with the name 'notebook-analysis'.
• Activate this environment by

  source activate notebook-analysis
  

• Start Jupyter notebook

  jupyter notebook
  

  Importing custom project libraries should work from within the ipython notebooks, if jupyter has been started from project-root. Jupyter's working directory can be controlled by:

  jupyter notebook --notebook-dir=<path-to-dir>
  

For image processing more libraries are needed. Major dependencies are:

• always:
  • pandas
  • matplotlib
  • numpy
• most image processing:
  • grabbit
  • SimpleITK
  • pydicom
  • dicom2nifti
  • vtk

If running CONDA python on MacOS, you can create a suitable python environment by

conda env create -f environment.yml

For image registration & LVd computation, the following additional dependencies need to be installed and path / environment settings in config.py be updated for the specific local installation:
- ANTs

These dependencies can be installed using the install_dependencies.sh script.