Dev Container for Reproducible Data Analysis

README.md

Containerized & Reproducible Data Analysis Environment

About

This gist contains Docker and Dev Container configs for R, Python, Rstudio Server, JupyterLab, Quarto, and renv.

Acknowledgements

Many thanks to the following people and organizations for sharing their code and tutorials openly. Please see below for a non-exhaustive list:

• David Smith (@revodavid) for his presentation and the code he provided on Zero-setup R workshops with GitHub Codespaces.

• えいつぴ (@eitsupi) for the helpful info on using RStudio in a Rocker container.

• Eric Nantz (R-Podcast) for the episode, Fully containerized R dev environment with Docker, RStudio, and VS-Code.

• Rocker for the Rocker Docker Images.

• Posit for Quarto, renv, and Rstudio.

• Project Jupyter for JupyterLab.

• Docker for the Docker Container Platform.

• Microsoft for the VS Code IDE and Dev Containers on GitHub.

Of course, I must also acknowledge the many open source contributers who helped build the software necessary for this workflow. Without their generosity, this would not be possible. Many thanks!

devcontainer.json

{
	"name": "Data Analysis Env",
	"build": {
		"dockerfile": "Dockerfile",
		// Update VARIANT to pick a specific R version: 4, 4.1, 4.2
		// More info: https://github.com/rocker-org/devcontainer-images/pkgs/container/devcontainer%2Fgeospatial
		"args": { "VARIANT": "4.2" }
	},

	// Install Dev Container Features. More info: https://containers.dev/features
	"features": {
		"ghcr.io/rocker-org/devcontainer-features/quarto-cli:1": {
			"installTinyTex": true
		},
		// Install JupyterLab and IRkernel.
		// More info: https://github.com/rocker-org/devcontainer-templates/tree/main/src/r-ver
		"ghcr.io/rocker-org/devcontainer-features/r-rig:1": {
			"version": "none",
			"installJupyterlab": true,
			"vscodeRSupport": "full",
			"installDevTools": true,
			"installRMarkdown": true,
			"installRadian": true,
			"installVscDebugger": true
		},
			
		// Install R dependencies
		"ghcr.io/rocker-org/devcontainer-features/r-packages:1": {
			"packages": "downlit"
		},
			
		// Install Apt packages.
			"ghcr.io/rocker-org/devcontainer-features/apt-packages:1": {
				"packages": "curl, nano, neofetch, tmux, libxt6"
		}
	},

	"customizations": {
		"vscode": {
			"extensions": [

				// infra tools
				"donjayamanne.githistory",
				"GitHub.codespaces",
				"ms-azuretools.vscode-docker",
				"GitHub.copilot",
				"GitHub.copilot-chat",
				"peakchen90.open-html-in-browser",
				"tomoki1207.pdf",
				"mechatroner.rainbow-csv",

				// language support
				"quarto.quarto",
				"ms-toolsai.jupyter",
				"ms-toolsai.jupyter-renderers",
				"ms-python.python",
				"ms-python.vscode-pylance",
				"REditorSupport.r",
				"RDebugger.r-debugger",
				"Ikuyadeu.r-pack",
				"DavidAnson.vscode-markdownlint",

				// theming
				"GitHub.github-vscode-theme",
				"cosmicsarthak.cosmicsarthak-neon-theme",
				"hassanoof.theme",
				"PKief.material-icon-theme",
				
				// vscode behavior
				"vsls-contrib.codetour",
				"dqisme.sync-scroll"
			]
		}
	},

	// Forward Jupyter and RStudio ports
	"forwardPorts": [8787, 8888],
	"portsAttributes": {
		"8787": {
			"label": "Rstudio",
			"requireLocalPort": true,
			"onAutoForward": "ignore"
		},
		"8888": {
			"label": "Jupyter",
			"requireLocalPort": true,
			"onAutoForward": "ignore"
		}
	},

	// Use 'postAttachCommand' to run commands after the container is started.
	"postAttachCommand": "R -q -e 'renv::repair() ; renv::restore()'"

	// Uncomment to connect as root instead. More info: https://aka.ms/dev-containers-non-root
	// "remoteUser": "root"
}

Dockerfile

# Pre-built Dev Container Image for R. More info: https://github.com/rocker-org/devcontainer-images/pkgs/container/devcontainer%2Fgeospatial
# Available R version: 4, 4.1, 4.2
ARG VARIANT="4.2"
FROM ghcr.io/rocker-org/devcontainer/geospatial:${VARIANT}

RUN install2.r --error --skipinstalled -n -1 \
        trackdown \
        gptstudio \
    && rm -rf /tmp/downloaded_packages \
    && R -q -e 'remotes::install_github("https://github.com/dcomtois/summarytools/tree/0-8-9")'

# Install Python packages
COPY requirements.txt /tmp/pip-tmp/
RUN python3 -m pip --disable-pip-version-check --no-cache-dir install -r /tmp/pip-tmp/requirements.txt

requirements.txt

pybryt
pylint
datascience
otter-grader
numpy
pandas
scipy
folium>=0.9.1
matplotlib
ipywidgets>=7.0.0
bqplot
nbinteract>=0.0.12
okpy
scikit-learn