Fast Azimuthal Integration in Python

Main development website: https://github.com/silx-kit/pyFAI

PyFAI is an azimuthal integration library designed for high-performance, achieving performance comparable to C and even greater through OpenCL-based GPU acceleration. It is based on histogramming the 2θ/Q positions of each pixel centre, weighted by pixel intensity, whereas the parallel version performs a SparseMatrix-DenseVector multiplication. Both method achieve the same numerical result. Neighboring output bins also receive contributions from pixels adjacent to the border through pixel splitting. PyFAI also provides tools to calibrate the experimental setup using Debye-Scherrer rings of a reference compound.

References

• The philosophy of pyFAI is described in the proceedings of SRI2012
• Implementation in parallel is described in the proceedings of EPDIC13
• Benchmarks and optimization procedure are described in the proceedings of EuroSciPy2014
• Calibration procedures are described in J. Synch. Radiation (2020)
• Application of signal separation to diffraction image compression and serial crystallography in J. Appl. Cryst. (2025)

Installation

Using PIP (python-package installer)

As with most Python packages, pyFAI is available via pip:

pip install pyFAI[gui]

It is recommended to run this in a virtual environment. Provide the --user option to perform an installation local to your user-space (not recommended). Under UNIX, you may have to run the command via sudo to gain root access and perform a system wide installation (which is neither recommended).

Using conda installer

PyFAI is also available via the conda installer from Anaconda:

conda install pyfai -c conda-forge

To install conda please see either conda or Anaconda.

From source code

The current development version of pyFAI can be downloaded from GitHub. The source code is currently distributed as a zip package.

Download and unpack it:

unzip pyFAI-main.zip
cd pyFAI-main

Install dependencies:

pip install -r requirements.txt

Build and test it:

python run_tests.py

For its tests, pyFAI downloads test images from the internet. Depending on your network connection and your local network configuration, you may have to set up a proxy configuration like this (not needed at ESRF):

export http_proxy=http://proxy.site.org:3128

Finally, install pyFAI in the virtualenv after testing it:

pip install .

The latest development version is available by checking out the Git repository:

git clone https://github.com/silx-kit/pyFAI.git
cd pyFAI
pip install .

To enable GPU acceleration in pyFAI, please install pyopencl.

Documentation

Documentation can be built using this command and Sphinx (installed on your computer):

python build-doc.py

Dependencies

Python 3.10 ... 3.14 are well tested and officially supported (thread-free is untested).

For full functionality of pyFAI, the following modules need to be installed:

• numpy
• scipy
• matplotlib
• fabio
• h5py
• pyopencl
• pyside6
• silx
• numexpr

Those dependencies can simply be installed by:

pip install -r requirements.txt

Ubuntu and Debian-like Linux distributions

On Ubuntu or Debian, the required Python modules for pyFAI can be installed either via the Synaptic Package Manager (under System → Administration) or from the command line using apt-get:

sudo apt-get install pyfai

MacOSX

On macOS, a recent version of Python (≥3.10) must be installed before installing pyFAI. Apple provides only an outdated version of Python 2.7 which is deprecated. To build pyFAI from source, you will also need Xcode, which is available from the Mac App Store. The binary extensions will use only a single core due to the limitation of the compiler from Apple. OpenCL is hence greatly advised on Apple systems. Next, install the missing dependencies using pip:

pip install -r requirements.txt

Windows

On Windows, a recent version of Python (>=3.10) must be installed before pyFAI. The Visual Studio C++ compiler is required when building from source Next, install any missing dependencies using pip:

pip install -r requirements.txt

Getting Help

A mailing-list, pyfai@esrf.fr, is available to get help on the program and how to use it. One needs to subscribe by sending an email to sympa@esrf.fr with the subject "subscribe pyfai". There is also a discussion space at GitHub and an issue tracker where bugs can be reported.

Maintainers

• Jérôme Kieffer (ESRF)
• Edgar Gutierrez Fernandez (ESRF)
• Loïc Huder (ESRF)

Contributors

Thanks to all who have contributed to pyFAI!

Indirect contributors (ideas, ...)

• Peter Boesecke
• Manuel Sánchez del Río
• Thomas Vincent
• Vicente Armando Solé
• Brian Pauw
• Veijo Honkimaki