Installing pyAFQ

The pyAFQ software works (at least) on Python 3.7 and 3.8.

Installing the release version

The released version of the software is the one that is officially supported, and if you are getting started with pyAFQ, this is probably where you should get started

To install it, in a shell or command line, issue the following:

pip install pyAFQ

Installing the development version

The development version is probably less stable, but might include new features and fixes. There are two ways to install this version. The first uses pip:

pip install git+

The other requires that you clone the source code to your machine:

git clone

Then, change your working directory into the top-level directory of this repo and issue:

pip install -e .[dev,fury]

which will install pyAFQ locally in editible mode and include extensions.


The project follows the standard github fork and pull request workflow. So if you plan on contributing to pyAFQ it is recommended that you fork the repository and issue pull requests.

See Contributing to pyAFQ


It is also recommended that you utilize python virtual environment and package mangagement tools and begin with a clean environment.

Installing using Docker

pyAFQ automatically builds and pushes a Docker image with pyAFQ installed for every commit. The images can be found here: To pull an image, you can either pull the latest:

docker pull

or specify the commit using its hash:

docker pull

pyAFQ also automatically builds and pushes a Docker image with pyAFQ and QSIprep ( installed for every commit to master. This image may be useful if you want an all-in-one image for pre-processing and tractometry. You can pull the latest of this image or use a specific commit as well:

docker pull