warutherford/rstudio-tidymodels repository overview

⁠rstudio-tidymodels

The purpose of this GitHub repo is to create a Dockerfile and eventual container based off the "rstudio-geospatial:3.6.3" Docker container to include R Tidymodels and dependents. The "rstudio-geospatial:3.6.3" container is sourced originally from the Rocker Project.⁠

⁠Methods for building a Docker container and deploying on CyVerse

Commands and workflow for running a Docker container either by pulling from CyVerse VICE and/or building a new Docker container based of a previous container called "rstudio-geospatial:3.6.3" to have R tidymodels already installed ("rstudio-tidymodels:3.6.3")

Text below provided by Tyson Swetnam on 2020-12-08 for specific use at University of Arizona and CyVerse:

Connect to the BIO5 VPN, then to connect to the machine over linux terminal you need to use the command:

ssh -p 1657 <your_cyverse_username>@gpu06.cyverse.org

You then enter your CyVerse password. Once you've logged in successfully, you should be ready to start the Docker container,

docker run -d -it --rm -v $HOME:/app --workdir /app -p 8799:80 -e REDIRECT_URL=http://gpu06.cyverse.org:8799 cyversevice/rstudio-geospatial:3.6.3

To mount your home directory as a folder called /app inside the container, so you can put your data there and it won't disappear when the container stops, remove the --rm tag:

docker run -d -it -v $HOME:/app --workdir /app -p 8799:80 -e REDIRECT_URL=http://gpu06.cyverse.org:8799 cyversevice/rstudio-geospatial:3.6.3

The -d flag will run the container as a detached background process, so you won't see anything after you execute it. You need to set it to an open port -p, suggested here as 8799 but it can be any four digit number that is not currently in use.

⁠Build a new docker container with tidymodels

add a tag of 3.6.3

docker build -t warutherford/rstudio-tidymodels:3.6.3

docker builds then push to docker hub

docker push warutherford/rstudio-tidymodels:3.6.3

To use this docker container in the future, sign back onto machine, and then run:

docker run -d -it --rm -v $HOME:/app --workdir /app -p 8799:80 -e REDIRECT_URL=http://gpu06.cyverse.org:8799 warutherford/rstudio-tidymodels:3.6.3 or remove the --rm flag to not remove container once stopped

Can give container a unique name for easier reference with --name flag, for example: docker run -d --name Rtidymodels_only -it -v $HOME:/app --workdir /app -p 8799:80 -e REDIRECT_URL=http://gpu06.cyverse.org:8799 warutherford/rstudio-tidymodels:3.6.3

⁠Finish and stop container

Check docker container ID or name docker ps

Stop the container by giving unique sequence of ID or name docker stop #### docker stop Rtidymodels_only (example)

The container is now stopped.

⁠Restart the container (if no --rm tag)

If stopped, restart by:

docker restart Rtidymodels_only or docker restart ####

⁠Rename the container

docker rename Rtidymodels_only Rtidymodels_complete

⁠Check docker manual for additional information/tags

man docker-run