feat(runtime): add R provider

[CalvinAllen]

Add R runtime provider to dtvem for managing R versions.

Overview

Implement an R provider that allows users to install, manage, and switch between R versions using dtvem. R is widely used in data science, statistics, and academic research where reproducibility is critical.

Why R Needs Version Management

• Research Reproducibility: Published papers must be reproducible with exact R version
• Package Compatibility: R packages often break between R versions
• Legacy Analysis: Need to run old analysis scripts with original R version
• Production Environments: Match development R version to production servers
• Academic Requirements: Thesis/dissertation code must work years later
• CRAN Packages: Over 19,000 packages with version-specific dependencies

Implementation Checklist

• Create src/runtimes/r/provider.go
• Implement all 19 Provider interface methods
• Add init() function for auto-registration: runtime.Register(NewProvider())
• Import in src/main.go: _ "github.com/CodingWithCalvin/dtvem/src/runtimes/r"
• Update schemas/runtimes.schema.json to add "r" to runtime enum
• Implement Shims() method to return: ["R", "Rscript"]
• Implement ShouldReshimAfter() (likely false - no global executable installs)
• Add detection logic for existing R installations (system, RVM)
• Implement download/installation logic from official sources
• Ensure cross-platform support (Windows, macOS, Linux)
• Add tests using provider contract harness (internal/runtime/provider_test_harness.go)
• Update documentation with R examples

Version Sources

• CRAN: https://cran.r-project.org/
• CRAN Mirrors: https://cran.r-project.org/mirrors.html
• R for Windows: https://cran.r-project.org/bin/windows/base/
• R for macOS: https://cran.r-project.org/bin/macosx/
• R Source: https://cran.r-project.org/src/base/

Download Sources by Platform

Windows

• Installers: https://cran.r-project.org/bin/windows/base/
• Format: R-{version}-win.exe (e.g., R-4.3.2-win.exe)
• Need to extract or install silently
• Base R only (Rtools separate for package compilation)

macOS

• Installers: https://cran.r-project.org/bin/macosx/
• Format: R-{version}.pkg or R-{version}-arm64.pkg
• Intel (x86_64) and Apple Silicon (arm64) builds available
• Need to extract .pkg without system installation

Linux

• Source Tarballs: https://cran.r-project.org/src/base/R-{major}/
• Format: R-{version}.tar.gz
• Must compile from source (dependencies required)
• Or use prebuilt binaries from distributions

Migration Sources

Detect and migrate from:

• System installations:
  • Windows: C:\Program Files\R\R-{version}\
  • macOS: /Library/Frameworks/R.framework/Versions/{version}/
  • Linux: /usr/bin/R, /usr/local/bin/R
• RVM (R Version Manager): If exists
• RStudio: May bundle R versions

Package Libraries

R packages are typically installed per major.minor version:

• User Library: ~/R/{platform}-library/{major}.{minor}/
• Site Library: System-wide package location
• Decision: Each R version should have its own library
  • Location: ~/.dtvem/versions/r/{version}/library/
  • Set R_LIBS_USER environment variable

Shims to Create

• R - Main R interpreter (case-sensitive!)
• Rscript - Command-line R script execution

Automatic Reshim Detection

Implement ShouldReshimAfter():

• Returns false - R doesn't install global executables
• R packages are libraries, not executables
• Exception: Some packages install command-line tools, but rare

Cross-Platform Considerations

Windows

• Installer Extraction: May need to extract from .exe installer
• Rtools: Separate download for compiling packages (optional)
• PATH: Add bin/, bin/x64/ to PATH
• Registry: System R writes to registry, dtvem versions shouldn't

macOS

• Framework Bundle: R.framework structure
  • /Resources/bin/R
  • /Resources/bin/Rscript
• Architecture: Intel vs Apple Silicon
  • R 4.1+ has arm64 builds for M1/M2
• Dependency: May need Xcode Command Line Tools for package compilation

Linux

• Compilation: Must compile from source or use distro packages
• Dependencies: Many (gcc, gfortran, libreadline, libcurl, etc.)
• Package Compilation: Need development headers for system libraries

Version Format

• R uses semantic-like versioning: 4.3.2, 4.2.1, 3.6.3
• Format: {major}.{minor}.{patch}
• Major version changes are rare (last was 3.x → 4.0 in 2020)
• Minor versions: 4.0, 4.1, 4.2, 4.3, 4.4 (annual releases)

R Package Compatibility

• Packages compiled for R 4.3.x often work on R 4.3.y (same minor)
• Packages must be recompiled for major.minor changes (4.2 → 4.3)
• renv: R's equivalent to Python's virtualenv
  • Per-project package libraries
  • Works alongside dtvem's version management

Testing Requirements

• Use provider contract harness in runtimes/r/provider_test.go
• Test all 19 interface methods
• Test version detection from system installations
• Test installation on all platforms
• Test shim creation and execution
• Test R_LIBS_USER environment variable setup

Special Considerations

RStudio Integration

• RStudio detects R installations automatically
• Should work with dtvem-managed R versions
• RStudio preferences allow selecting R version

Package Installation

• Packages installed with install.packages()
• Stored in version-specific library directories
• Binary packages (Windows/macOS) vs source packages (Linux)

Bioconductor

• Separate repository for bioinformatics packages
• Version tied to R version (Bioconductor 3.18 → R 4.3)

CRAN Task Views

• Curated package collections for specific domains
• Not managed by dtvem

Environment Variables

• R_HOME: Points to R installation directory
• R_LIBS_USER: User package library location
• R_LIBS_SITE: Site-wide package library
• dtvem should set these per version

Configuration Files

• .Rprofile: R startup configuration (often in project directory)
• .Renviron: Environment variables (often in project directory)
• ~/.Rprofile: User-level startup (shared across versions)

Integration with Data Science Workflows

• Jupyter: R kernel (IRkernel) for Jupyter notebooks
• RMarkdown: Literate programming documents
• Shiny: Web applications built with R
• Posit (RStudio) Products: Workbench, Connect, Package Manager

renv Integration

R's renv package provides project-level package management:

• Similar to Python's venv or Node's node_modules
• renv::init() creates project-specific library
• Works alongside dtvem (dtvem manages R version, renv manages packages)

Academic/Research Context

• Reproducibility: Core requirement in research
• Long-term Archival: Code must run years after publication
• Thesis/Dissertation: PhD students need code to work for defense
• Collaborators: Different lab members may use different R versions

References

• Provider interface: src/internal/runtime/provider.go
• Example implementations: src/runtimes/node/, src/runtimes/python/
• Adding providers guide: .claude/CLAUDE.md
• R Project: https://www.r-project.org/
• CRAN: https://cran.r-project.org/
• renv: https://rstudio.github.io/renv/