R & Python Docker Image

This repo contains a Dockerfile used to build a debian-based Docker image with R and Python installed.

It includes the following R packages:

• tidyverse
• dplyr
• devtools
• AlphaSimR
• mvtnorm
• tmvtnorm
• jeanlucj/BreedSimCost (from GitHub)

And the following python packages:

• botorch
• numpy
• rpy2

All commands should be run within the directory containing the Dockerfile. If running on BioHPC, replace the docker command with docker1.

Specifying R and Python versions

The R and Python versions are specified as build args when building the docker image.

The R version must match an available Docker r-base tag. The Python version must be available to download from the Python ftp directory.

Specifying R and Python packages

When the image is built, it will install the packages defined in the ./build/R_packages.txt and ./build/python_requirements.txt files. The python_requirements.txt file can include specific versions to install.

Building the Image

To build the image:

1. First, set the R and Python versions to use:

export R_VERSION=4.2.0
export PYTHON_VERSION=3.9.13

2. (Optional) Change the R and/or python packages to include.

3. Then, run the docker build command:

docker build -t triticeaetoolbox/r_python:R${R_VERSION}-python${PYTHON_VERSION} --build-arg R_VERSION=${R_VERSION} --build-arg PYTHON_VERSION=${PYTHON_VERSION} .

NOTE: You may need to change the last . (representing the current working directory) to the full/absolute path of the directory containing the Dockerfile.

This will create an image named triticeaetoolbox/r_python with the tag R4.2.0-python3.9.13 to specify which versions of R and python were used in that build. Use docker images to display the locally available images.

Transferring an Image

To transfer an image built on one machine to run on a different machine, you can create a .tar archive of the image, transfer it to another machine, and then load the .tar file on the new machine.

1. List the locally available images on the first machine to get the name and tag of the image to transfer

$ docker images
REPOSITORY                       TAG                   IMAGE ID       CREATED         SIZE
triticeaetoolbox/r_python        R4.1.0-python3.8.0    5fa9fa43651a   14 hours ago    4.28GB
triticeaetoolbox/r_python        R4.2.0-python3.9.13   6ed0ef719546   16 hours ago    4.14GB

2. Create a .tar archive of the image

docker save -o r_python.tar triticeaetoolbox/r_python:R4.2.0-python3.9.13

3. Copy the .tar file to the new machine (ie using scp or any other file transfer tool)

4. Load the image on the new machine

docker load -i r_python.tar

Use a Pre-Built Image

To use a pre-built image from Docker Hub, view the list of available tags for the triticeaetoolbox/r_python image.

To download a specific image, run docker pull triticeaetoolbox/r_python:R4.2.0-python3.9.13.

Starting a Container

To run the image in a new container:

docker run --name jannink_analysis -v $(pwd)/scripts:/root/scripts -d triticeaetoolbox/r_python:R4.2.0-python3.9.13

This will create and run a new container named jannink_analysis using the image triticeaetoolbox/r_python with the tag R4.2.0-python3.9.13.

It will also bind-mount the ./scripts directory from the host into the /root/scripts directory within the container. You can place scripts and data in this directory and they will be available within the Docker container. Any output from the container saved in the /root/scripts directory will also be saved on the host.

Using the Container

By default, the container will not execute any commands or scripts on its own. To run scripts within the container, start a bash session within the container by running docker exec -it jannink_analysis bash and then manually execute the scripts in the /root/scripts directory.

Stopping the Container

To stop the container, run docker stop jannink_analysis.

To restart the container, run docker start jannink_analysis.

To remove the stopped container, run docker container rm jannink_analysis.