RRE – Reproducible Research Environment

RRE is a one-shot project creator that scaffolds a reproducible R (and optional Python) research project backed by Docker and GitLab. It automates the setup of repositories, Docker images, and project structure to ensure computational reproducibility. It works in a very simple way. The workflow is depicted in the figure below.

Key Features

• Automated Setup: Creates a brand-new GitLab repository (via OAuth or PAT) and a local project folder with a sensible structure.
• Docker Integration: Produces a pinned Docker image (RStudio Server) and runs scripts.
• CI/CD Ready: Optionally wires up GitLab CI to build and push an image to Docker Hub.
• Python Support: Add reticulate / conda support (optional) and set up renv.
• Cross-Platform: Ships with both a GUI and a CLI, including OS-specific bootstrap scripts for Windows, Linux, and macOS.

Prerequisites

• Git and Docker installed and running.
• Python 3.11+ (The Windows script will auto-install Python if missing).

Quick Start

Windows

1. Double-click install_and_run_windows.bat.
2. The script will install necessary dependencies and launch the GUI.

Linux / macOS

1. Open a terminal.
2. Run the bootstrap script:
   • Linux: ./install_and_run_linux.sh
   • macOS: ./install_and_run_mac.sh

What Gets Created

When you create a new project, RRE generates:

• A GitLab repository with an initial commit.
• A local folder containing:
  • Standard directories: Data/, Scripts/, Packages/.
  • main.Rmd: A starter manuscript.
  • Dockerfile: Based on a rocker image.
  • run_project.sh / run_project.bat: Scripts to start RStudio locally.
  • Configuration files: .Rprofile, .gitlab-ci.yml, etc.

Usage

Once the project is created, you can start your reproducible environment by running the run_project script in your project folder. This will launch RStudio Server inside a Docker container, accessible via your web browser (usually at http://localhost:8787).