Plot with Mayavi in Jupyter notebook on Docker for Mac¶

Last update: 14 January, 2017

1. Install Docker for Mac

2. Install Xquartz

3. Install socat:

   $ brew install socat
   

4. Build docker image:

   FROM jupyter/datascience-notebook
   MAINTAINER Taku Yoshioka <taku.yoshioka.4096@gmail.com>
   
   USER root
   
   RUN apt-get update && \
       apt-get install -y \
       libglu1-mesa
   
   RUN apt-get update && \
       apt-get install -y \
       libvtk5-dev
   
   RUN apt-get update && \
       apt-get install -y \
       mayavi2
   
   RUN rm /bin/sh && ln -s /bin/bash /bin/sh
   
   USER jovyan
   
   ENV LIBGL_ALWAYS_INDIRECT 1
   
   RUN source activate python2 && \
       conda install -y mayavi && \
       conda install -y -c conda-forge jupyter_contrib_nbextensions
   RUN source activate python2 && \
       jupyter nbextension install --py mayavi --user
   

   $ build -t jupyter-mayavi .
   

5. Set indirect GLX flag to be used for Xquartz

   $ defaults write org.macosforge.xquartz.X11 enable_iglx -bool true
   

   • Hoffman2 Cluster User Guide

6. Run Xquartz

7. Run socat

   $ socat TCP-LISTEN:6000,reuseaddr,fork UNIX-CLIENT:\"$DISPLAY\"
   

8. Run docker container

   $ ip=$(ifconfig en0 | grep inet | awk '$1=="inet" {print $2}')
   $ docker run --rm -p 9000:8888 -e DISPLAY=$ip:0 --name test jupyter-mayavi
   

9. Launch a notebook with Python2 kernel

10. Run the following commands in the notebook:

    from mayavi import mlab
    mlab.init_notebook()
    mlab.test_plot3d()