============
Docker image
============
We can easily build a docker image using the following *Dockerfile*:
::
	
	# **************************************************************************
	# Purpose ...: jfricas (https://github.com/fricas/jfricas)
	# Build .... : docker build -t jfricas:latest .
	# Test ..... : docker run -ti --rm --network=host 
	#                --env DISPLAY=:0 fricas/jfricas:latest jupyter notebook
	# Version .. : 0.2.17 -- 17-SEP-2019
	# OS ....... : ubuntu:latest
	# FriCAS ... : fricas:latest 
	# **************************************************************************
	
	FROM fricas/fricas
	
	LABEL fricas-devel <fricas-devel@googlegroups.com>
	
	# ENV DEBIAN_FRONTEND noninteractive
	# discouraged: see https://docs.docker.com/engine/faq/
	
	ENV JFRICAS_GITHUB https://github.com/fricas/jfricas
	ENV JFRICAS_VERSION 0.2.17
	
	# ======================================
	# Get packages and install missing tools
	# ======================================
	
	RUN apt-get update && apt-get install -y -q \
	    python3.6 \
	    python3-pip \
	    cl-asdf \
	    cl-hunchentoot \
	    firefox \
	    gnuplot
	
	
	
	# ===============================
	# Install jfricas.pip from Github
	# ===============================
	
	RUN cd /root && \
	    pip3 install jupyter && \
	    git clone $JFRICAS_GITHUB && \
	    cd jfricas && \
	    pip3 install . && \
	    python3 -m jfricas.install_kernel && \
	    cd /root
	
Build
-----
Go to the directory containing the ``Dockerfile`` and type
::

    docker build -t jfricas:latest .
    
    or with logfile:
    
     docker build -t jfricas:latest . > log

Alternatively you want to use the ``build.sh`` script:
::

	#!/bin/sh
	sudo docker rmi fricas/jfricas:latest
	sudo docker build -t fricas/jfricas:latest . > log
	sudo docker images


Test
----
There is a ``test.sh`` in the docker subfolder, or you can enter the command
by hand (one line, otherwise use line continuation):
::

    docker run -ti --rm --network=host --env DISPLAY=:0 fricas/jfricas:latest 
	  jupyter notebook --no-browser --allow-root
	  
.. image:: pics/docker_run_jfc.png

To merely access the container you can also use
::

   docker run -ti --rm fricas/jfricas:latest /bin/bash
   
to get a ``bash``, from where you can start things manually.


Right click on a link and choose ``open``.

.. image:: pics/docker_open_jfc.png

Your browser should open and display the ``root`` directory of the image.

.. image:: pics/docker_jfc_home.png

Choose the kernel by ``New->FriCAS``

.. image:: pics/docker_new_fricas.png

then you can use the notebook:

.. image:: pics/docker_fricas_nb.png


**Note**: the ``--rm`` option means that the container is deleted after closing
the notebook server. This is for testing only, so that your computer will 
not populated with a lot of orphaned containers. If you want to save your
work (``*.ipynb``, etc.), then remove ``--rm`` and use ``--name`` for named 
containers if you like.

Push
----
To push the image you can use the script ``push.sh``, or the command:
::

      sudo docker push fricas/jfricas:latest

.. image::  pics/docker_push.png


RUN
---
There are two scripts: ``run.sh`` and ``runx.sh`` which may be used to start
the ``jupyter notebook`` inside the ``jfricas`` docker image. The latter can
be used if you want to have ``Hyperdoc`` and ``Draw``, otherwise ``run.sh``
is safer as it will not set any permissions (``xhost local:root``).

Example for running a named container (jfc):
::

       docker run -ti --name jfc --network=host 
	     --env DISPLAY=:0 fricas/jfricas:latest 
		 jupyter notebook --no-browser  --allow-root
		 
To save the changes in ``jfc`` to the image, do
::

    sudo docker commit jfc fricas/jfricas
    [sudo] password for user: 
    sha256:bea4f6525c7b2f03f6893bec5a03646a1ba69cb8f669890c14105fa61010c1dd

Then you can remove the ```jfc`` container (if you like):
::

    docker rm jfc
    
Do not forget do push the changed image (only if you want to, of course):

    sudo docker push fricas/jfricas

The commands ``docker ps -a`` will list all active containers, while
``docker images`` will show the images.