docker-rshiny

This serves as a GitHub repository where images can either be built directly from or pulled from our gru/wu/rshiny repository on dockerhub. The docker images serve images for our RShiny-applications, for example, materialflows.net and SCP Hotspot Analysis.

How to use it?

Remote

In your docker-compose.yml use gruwu/rshiny as your image.

e.g.

services:
  app:
    container_name: shiny
    image: gruwu/rshiny
    # or with tag to get specific version:
    # image: gruwu/rshiny:2019-04-02

After a download / git clone

You can either build it with docker build . -t example_name (make sure to navigate to the directory first) and then use this example_name as your image: as in Remote.

You can also state a build context. Just create a directory (e.g. docker-rshiny) with a Dockerfile, shiny-server.conf, and shiny-server.sh, inside your project directory, adjust packages to your needs and then modify the docker-compose.yml to use build instead of image.

services:
  app:
    container_name: shiny
    # image: gruwu/rshiny
    build:
      context: ./docker-rshiny/

How to adjust it?

The most common use case will be that you need to add or remove a package dependency. If you can think of anything else that might be of use, open an issue.

Add or remove a package dependency

For this, you just add your desired package at the end of the additions to && install2.r --error \, where there is no \ anymore (e.g. after leaflet.extras). Or, in case you want to delete it, you remove the line of the package that you do not need.

&& install2.r --error \
  ...
  leaflet.extras \
  the_cran_package_i_need

Alternatively, or if you need a specific version or a GitHub package, you can add the following:

# install plotly version 4.9
RUN R -e "install.packages('https://cran.r-project.org/src/contrib/plotly_4.9.0.tar.gz', repos = NULL)"

Test any changes

Navigate to the test_app directory, then type docker-compose up -d. This builds the app inside the test_app folder with every docker-setup in the directory. With that you can test compatibility of your app with new versions or also older versions, other R versions or package versions, etc.

The apps will be deployed between the ports :80 (build in the main directory), :81 (2019-04-02 version) and :84 (2019-05-28 version).