Introducing this chapter:

• Preparation

• Setting up accounts

• Installing software

Create an Account on Docker Hub. This is helpful for several reasons:

• We can save and keep our Container Image in the cloud repository (1 private repository is given for free)

• We can share our work (note that image can also be stored persistently) if needed

Installation of Docker Desktop - Application to develop or run Docker Containers

Reviewing several settings of the Docker Desktop application:

• keep it starting up on re-start or not

• allocated disk memory

• allocated CPU/resources

Instructions to install Docker on Linux

The best and easiest way to follow this course is to use any User Friendly Version Control Software!

My preferred one is GitHub Desktop

Goals of this section - build our personalized Docker Image with R-Studio and all our favorite packages!

Re-using Dockerfile for our new project

Put our project under version control! It really takes 5 minutes!

Starting building a Docker Image

docker login

docker build -t <docker hub user name/repository name> .

mind the dot!

While our image is being built we will update our Readme file.

Why? For a sole purpose of be able to repeat the procedure and re-build the same image some time after!

Use command docker images to list available images

Use command docker run with options to run the container

Map folders content to use them in our container

Updating readme file with an example of instructions learned so far

Creating executable file using vim editor

Saving this image in the Docker Hub [optional]

Saving image locally has it's advantages...

Image can be removed from our computer to optimize space. Good thing is that image can be easily restored either from file or from the docker hub. It is enough to run executable file and docker will check for the image presence first locally and then on Docker hub... and it will download it!

Practicing to restore image from file. Not relevant in case of:

• image is stored in Docker Hub

• good internet connection

• no concern over publishing image

Using another terminal to fetch container ID

Change docker image can be also done after building it. For example:

• adding new package

• updating something on Linux machine

• adding our research document, run it and saving results into the image...

Update Image on Docker Hub

Save a new version of an image. Useful to support lifecycle of the software

taking advantage of the automated builds:

adding the R package with much less effort!

Multiple advantages of automated build are including significant reduction of effort to maintain released images

A very practical example on how to use R-Studio in a Docker Container:

• Map local folder to container

• Create multiple project using R-Studio

• Add project under version control

Summary of this section

Introduction to the section. What are the advantages to have our own image template with R installation?

Reviewing repository containing:

1. R script with commands to install favorite R packages

2. Dockerfile with commands to start building image from verified publisher 'rocker/r-base'

Use feature Autobuilds to build image directly on Dockerhub infrastructure

Recommendation to use verified publisher of the docker images to start working on your own container

docker pull vladdsm/docker-r-h2o

testing r container

Intermediate summary

• Having IDE in container

• Having r-base with h2o

we can update our images using Autobuilds!

Visual blueprint explaining management of docker images

• Re-use dockerfile

• Focus on research using Version Control, automated builds

• Share results to others

visual representation of steps

Added reproducible research doc

• Adding R markdown document to the image

• Add R script to render this document

testing our container, results will be generated in the folder mapped to the output folder of the container

push the image to Dockerhub

Reproducing research by running container on Windows 10 machine

How to use our image to add any of our scripts and run it!

#Create a new dockerfile using

FROM vladdsm/docker-r-h2o:latest

# add our script

ADD *.RMD document

# execute the script as soon as container runs

CMD RScript run_analysis_script.R

Building and pushing the image to the Docker Hub

Testing the container... expected result is to see the output file generated in the folder

Publish our work in GitHub

In this Section we will look into specific case study aimed to demonstrate usage of Docker Containers to publish and consume data via Application Programming Interface [API].

Key features discussed:

• Setup Docker Container to publish API

• Setup Docker Network

• Use Shiny App in another container to consume data from the API

In this Section we will look into specific case study aimed to demonstrate usage of Docker Containers to publish and consume data via Application Programming Interface [API].

Key features discussed:

• Setup Docker Container to publish API

• Setup Docker Network

• Use Shiny App in another container to consume data from the API

Develop Plumber API code using Environment running R-Studio in the Docker container

Build docker image with Plumber API

Creating and inspecting docker network

Testing connectivity between working containers

Getting code to the toy Shiny Application

Test multi container application

Introduction to this chapter:

• Goal is to place our Shiny Application into the portable Docker Image

Quick win:

• Starting our first rocker shiny server

Video explanation:

• Getting started with Shiny Server

• Possible options:

  • rocker/shiny-verse

  • rocker/shiny

  • openanalytics/shinyproxy

Using Shiny Server on Docker Container to run locally stored Shiny Apps

Embedding Shiny Apps into Docker Image and then run Starting Shiny App on Shiny Server running on Docker Container ...

... in English:

1. Storing Shiny App into Docker Image

2. Running Shiny Server on Docker Container

3. Accessing Shiny Application from Browser

A quick introduction to {golem} package that helps to professionally deploy ShinyApp

Planned section

Develop a blueprint and working example to setup a workflow of:

• Development of ShinyApp as an R Package

• Testing

• Build image

• Deploy Container to production

Note: planned content

TDL: add links from UseR2019