Docker

 

These exercises follow the material in the Docker section of Reproducible R.

1. Run R using a Docker container from Docker Hub

• Pull the Rocker R image for version 4.2.3 from Docker Hub
• Confirm that this image is available on your computer
• Launch a container from this image
  • HINT: is the -i and -t flags in the docker run command to run R interactively in container
• execute R code from within the container

docker pull rocker/r-ver:4.2.3
docker images
docker run --rm -ti rocker/r-ver:4.2.3
# an R prompt should appear in the terminal allowing you to write R code

2. Customize the Rocker R Docker image

• Make a Dockerfile that contains R version 4.2.3, the CRAN packages dplyr and ggplot2, the Bioconductor packages DESeq2 and rtracklayer, and the conda package samtools.
  • HINT: Make sure you install BiocManager from CRAN so that you can install Bioconductor packages and also install Herper so that you can install samtools
• Build an image using this Dockerfile

# navigate to 'r_course' directory in downloaded material
 cd /PathToDownloadedCourse/Reproducible_R-master/r_course
 
docker build -t rocker/r-ver:4.2.3_v2 ./data/docker_exercise
docker images

• Launch a container with this image with a volume from your local computer mounted
• Confirm that the container is running
• Confirm that you can see the files in the mounted folder
  • HINT: The R Rocker image does not set the working directory. You need to mount the local directory somewhere within the docker image and then navigate to that location from R to see those files. I recommend mounting your local directory to ‘/home/local_data’ within that container’s file system (this folder will automatically be made in container)
  • HINT #2: you can also use the ‘docker exec’ command from a separate terminal tab to run a command within a container. You can type ‘docker exec IMAGEID ls’ to list the file system.

# Code in terminal 
docker run --rm -ti -v ./data:/home/local_data rocker/r-ver:4.2.3_v2
docker ps

# see the mounted files
docker exec 13223bfef159 ls home # this command contains the container ID from 'docker ps'
# you should see both the miniconda directory that was made in the Dockerfile and the 'local_data' directory that contians the mounted directory

• make a plot from the ‘mtcars’ data set shows the relationship between weight (wt) and miles per gallon (mpg) that comes with R and save to your computer
  • HINT: this table is available as a variable called ‘mtcars’

# R code 

# set working directory to be the mounted directory
setwd("home/local_data")
# see files in mounted folder
list.files()

library(ggplot2)

# make and save plot
ggplot(mtcars, aes(x = wt, y = mpg)) + 
  geom_point()
ggsave("mpg_vs_weight.pdf")

# now look in the directory on your computer that is mounted and this plot should be there

• confirm that samtools is installed

# R code
library(Herper)

# the environment name and miniconda path set in the Dockerfile
Herper::local_CondaEnv(new = "docker_exercises", 
                       pathToMiniConda = "/home/miniconda")
# test out samtools
system("samtools --version")

- push this image to Docker Hub

# Code in terminal 

# log in and provide credentials used to sign into Docker Hub
# this will prompt you to enter username and password
docker login

# get image ID to tag and push 
docker images

# tag the image you want to push with your Docker Hub username and a tag name after the colon
# the ID is from the 'docker images' command
docker tag a8954d8caa31 dougbarrows/r-ver_4.2.3_v2:topush

# push to Docker Hub
docker push dougbarrows/r-ver_4.2.3_v2:topush