Docker image for analyses using RStudio and Python-Conda

The motivation for this project is to encourage the use of portable development environments in research and engineering. The environment should be intuitive to use so that anyone can deploy it again and reproduce your results - even you six months from now!

This project provides a portable development environment that enables you to use R and Python together. The Docker image tschaffter/rstudio offered by this project is based on the image rocker/rstudio.

Features:

• Use sentitive information like credentials without specifying them in your notebooks, hence preventing the risk of publishing this information publicly.
• Create and manage Conda environments (Miniconda) using the R library reticulate to run and/or develop Python programs that require different version of Python or packages.
• Render Rmd notebook to HTML using the Docker image provided in this project, e.g. to generate HTML notebooks in GitHub workflows before publishing them to GitHub Pages.
• Benefit from regular updates of the image tschaffter/rstudio which will bring the latest versions of R/RStudio and other dependencies (Miniconda, R and Python packages).
• You only need the Docker Engine on your system to develop code in R and Python (see Requirements).

This image includes the following common Sage Bionetworks software:

• R libraries
  • sagethemes: Sage-branded plot themes.
• Python packages
  • challengeutils: Synapse challenge utility functions.
  • synapseclient: Programmatic interface to Synapse services for Python.

All packages:

• R (see renv.lock).
• Python (see conda/sage-bionetworks/environment.yml).

• Docker Engine >=19.03.0

The example notebooks below are rendered to HTML and published to GitHub Pages by the CI/CD workflow of this repository.

Rmd Notebook	Description	HTML Notebook
notebook.Rmd	Default RStudio notebook.
r-and-python.Rmd	Shows how to use R and Python together.
sagethemes.Rmd	Example notebook provided by the R library sagethemes.

Important: Please make sure when you write your own notebooks that no sensitive information ends up being publicly available. Please check with the information security officer of your organization to confirm that the approach described here can be applied to your use case.

1. Create and edit the configuration file. You can initially start RStudio using this configuration as-is.

   cp .env.example .env
   

2. Start RStudio. Add the option -d or --detach to run in the background.

   docker compose up
   

RStudio is now available at http://localhost. On the login page, enter the default username (rstudio) and the password specified in .env.

To stop RStudio, enter Ctrl+C followed by docker compose down. If running in detached mode, you will only need to enter docker compose down.

You can use the image tschaffter/rstudio as-is to start an instance of RStudio and develop tools that interact with Sage Bionetworks services, e.g. Synapse.

If you want to create a portable development environment, start by creating a new GitHub repository from this template. You can then customize your environment by specifying the R and Python packages to include with your image. Finally, edit the the GitHub workflow .github/workflows/ci.yml to indicates the Docker repository where the image should be pushed (see Section Versioning).

Example projects that use this repository / image:

• Sage-Bionetworks/rstudio
• Sage-Bionetworks-Challenges/challenge-analysis
• nlpsandbox/notebooks

In RStudio, use the following options to add and update libraries:

• Tools > Install Packages...
• Tools > Check for Package Updates...

Run the command renv::snapshot() to update the file renv.lock, which is used in Dockerfile to install the required R libraries.

See the content of the folder conda for an example of how to define a conda environment. The packages to add to this environment must be added to the file requirements.txt. The creation of one or more Conda environments can be specified in Dockerfile.

Set the environment variables SYNAPSE_TOKEN to the value of one of your Synapse Personal Access Tokens. If this variable is set, it will be used to create the configuration file ~/.synapseConfig when the container starts.

This Docker image comes with Miniconda installed (see below) and an example Conda environment named sage-bionetworks. This environment includes packages used to interact with the collaborative platform Synapse developed by Sage Bionetworks.

Attach to the RStudio container (here assuming that rstudio is the name of the container). For better safety, it is recommended to work as a non-root user. You can then list the environments available, activate an existing environment or create a new one.

$ docker exec -it rstudio bash
container # su yourusername
container $ conda env list
container $ conda activate sage-bionetworks

The R code below lists the environment available before activating the existing environment named sage-bionetworks.

> library(reticulate)
> conda_list()
    name                              python
1 miniconda           /opt/miniconda/bin/python
2      sage-bionetworks /opt/miniconda/envs/sage-bionetworks/bin/python
> use_condaenv("sage-bionetworks", required = TRUE)

When using Docker volumes, permissions issues can arise between the host OS and the container. You can avoid these issues by letting RStudio know the User ID (UID) and Group ID (GID) it should use when creating and editting files so that these IDs match yours, which you can get using the command id:

$ id
uid=1000(kelsey) gid=1000(kelsey) groups=1000(kelsey)

In this example, we would set USERID=1000 and GROUPID=1000.

Set the environment variable ROOT=TRUE (default is FALSE).

docker logs --follow rstudio

This Docker image provides the command render that generates an HTML or PDF notebook from an R notebook (.Rmd). Run the command below from the host to mount the directory $(pwd)/notebooks where the R notebook is and generate the HTML notebook that will be saved to the same directory with the extension .nb.html.

docker run --rm \
    --env-file .env \
    -v $(pwd)/notebooks:/data \
    tschaffter/rstudio:4.0.5 \
    render /data/examples/*.Rmd

This repository uses semantic versioning to track the releases of this project. This repository uses "non-moving" GitHub tags, that is, a tag will always point to the same git commit once it has been created.

The artifact published by this repository is the Docker image tschaffter/rstudio. The versions of the image are aligned with the versions of R/RStudio, not the GitHub tags of this repository.

The table below describes the image tags available.

You should avoid using a moving tag like latest when deploying containers in production, because this makes it hard to track which version of the image is running and hard to roll back.

Apache License 2.0