Visualization Backends
~~~~~~~~~~~~~~~~~~~~~~

pyAFQ can generate figures to visualize the results of tractometry.
How these figures are generated depends on the choice of visualization
backend. Currently, there are three choices:

#. plotly: use `Plotly <https://plotly.com/python/>`_ to generate interactive
   figures that are exported as html files and rotating GIFs.
   The rotating GIFs can take a long time to generate (~2 min/gif)
   using this backend. No additional installations are
   required to use this backend. 

#. plotly_no_gif: use `Plotly <https://plotly.com/python/>`_ to generate
   interavtive figures that are exported as html files, but do not
   generate GIFs. No additional installations are required to use this
   backend.

#. fury: use `Fury <https://fury.gl/>`_ to generate rotating GIFs. Unlike
   our current setup in Plotly, Fury can generate GIFs quickly. To use this
   backend, install pyAFQ with the optional fury requirements:
      pip install pyAFQ[fury]
   And install `libGL <https://dri.freedesktop.org/wiki/libGL/>`_. If running
   fury on a headless environment, additionally install and use
   `Xvfb <https://www.x.org/releases/X11R7.6/doc/man/man1/Xvfb.1.xhtml>`_.
   Xvfb is used automatically in the api.GroupAFQ object if the virtual_frame_buffer
   argument is set to True.

By default, plotly_no_gif is used. Fury requires additional
installations and does not make interactive figures, and Plotly takes a
significant amount of time to generate rotating GIFs.


Fury Dockerfile for Cloudknot
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
If using the fury visualization backend while running pyAFQ on cloudknot, you
must provide a base image with the fury requiements installed.
Below is an example dockerfile that can be used to build that base image:: 

   # Use official python base image
   FROM python:3.9
   # Install libgl
   RUN apt-get update
   RUN apt-get install -y libgl1-mesa-glx
   RUN apt-get install -y xvfb