PyCharm, Vagrant, Fabric & Anaconda

Development Workspace Creation

• On your local filesystem, in your development workspace, create a directory colour-science and cd into it:

$ mkdir colour-science
$ cd colour-science

• Clone the colour-vagrant repository:

$ git clone git://github.com/colour-science/colour-vagrant.git

• You should now have a colour-vagrant directory nested into the colour-science one.

  Let's open the colour-vagrant directory into PyCharm.

Vagrant Installation

We will loosely follow Jetbrain's PyCharm Documentation.

• Install VirtualBox.

• Install Vagrant.

• Install Fabric, this is the provider needed for the colour-vagrant environment. Initially, I was provisioning with Puppet. It was cumbersome to use, so I decided to go for Fabric and the vagrant-fabric plugin:

  $ pip install --user fabric
  

• Install XQuartz: This is the X11 display server for macOS. The virtual machine will export the display to it so that you can see the figures from Matplotlib.

• VirtualBox directories syncing performance degrades quickly with large number of files.

  As a result the directories syncing is done with NFS.

  You will however be asked for your macOS password at each virtual machine spin up because Vagrant needs to modify configuration files on the macOS host.

  In order to avoid that and following the Vagrant Documentation, you can edit your macOS /etc/sudoers file and append the following content:

  Cmnd_Alias VAGRANT_EXPORTS_ADD = /usr/bin/tee -a /etc/exports
  Cmnd_Alias VAGRANT_NFSD = /sbin/nfsd restart
  Cmnd_Alias VAGRANT_EXPORTS_REMOVE = /usr/bin/sed -E -e /*/ d -ibak /etc/exports
  %admin ALL=(root) NOPASSWD: VAGRANT_EXPORTS_ADD, VAGRANT_NFSD, VAGRANT_EXPORTS_REMOVE

PyCharm Vagrant Configuration

The canonical way is to use Vagrant from the command line as described in Vagrant Documentation, but here we will leverage the PyCharm integration.

• Set the Vagrant Executable field to your Vagrant executable, e.g. /usr/local/bin/vagrant. This should not be required as vagrant should be enough, but I encountered issues lately.

• Set the Instance Folder field to your colour-vagrant directory.

• Add a new Vagrant box to PyCharm in the Boxes tab and use the following image: bento/ubuntu-16.04

• Add the vagrant-fabric plugin to PyCharm in the Plugins tab.

Your configuration screen should be approximately as follows:

Vagrant Up

With everything setup, hit the PyCharm --> Tools --> Vagrant --> Up menu item to spin up the virtual machine.

You can go grab a coffee, the initial provisioning will take roughly 25-30 minutes.

If the provisioning ended smoothly, you should have the following elements available:

• colour and colour-science.org repositories cloned into the colour-science directory right next to the colour-vagrant one.

• Remote Python environments installed into the virtual machine at the following location with all the dependencies needed:

$ cd /home/vagrant/miniconda/envs/
$ ls -l
total 16
drwxrwxr-x  4 vagrant vagrant 4096 Apr 29 03:12 ./
drwxrwxr-x 13 vagrant vagrant 4096 Apr 29 03:02 ../
drwxrwxr-x 20 vagrant vagrant 4096 Apr 29 03:09 python2.7/
drwxrwxr-x 20 vagrant vagrant 4096 Apr 29 03:30 python3.5/

• The colour-science.org website served from the virtual machine at the following address: http://localhost:8080/

PyCharm Environment Configuration

• Add the various remote Python interpreters to PyCharm the Configure a remote interpreter using Vagrant guide:

  • /home/vagrant/miniconda/envs/python2.7/bin/python

  • /home/vagrant/miniconda/envs/python3.5/bin/python

• Add the paths mappings from the macOS host to the virtual machine in the Defaults configurations, in my case the mappings are as follows:

  /Users/KelSolaar/Documents/Development/colour-science = /colour-science

Your Run/Debug configuration screen should be approximately looking like that:

Usage

$ ssh -X vagrant@192.168.32.64

$ xhost + 192.168.32.64

$ cd /colour-science
$ ipython notebook --pylab=inline --ip=0.0.0.0

$ source activate python2.7

$ source activate python3.5