Python versions and dependencies#

DIPY follows the Scientific Python SPEC 0 — Minimum Supported Versions recommendation as closely as possible, including the supported Python and dependencies versions.

Further information can be found in 🛠️ Toolchain Roadmap.


Depends on a few standard libraries: python (the core language), numpy (for numerical computation), scipy (for more specific mathematical operations), cython (for extra speed), nibabel (for file formats; we require version 2.4 or higher), h5py (for handling large datasets), tqdm and trx-python. Optionally, it can use fury (for visualisation), matplotlib (for scientific plotting), and ipython (for interaction with the code and its results). cvxpy, scikit-learn, statsmodels, pandas and tensorflow are required for some modules.