Packaging a R (CRAN) package#

Packaging a R package is similar to packaging a Python package!

Generating a starting point#

You can use rattler-build to generate a starting point for your recipe from the metadata on CRAN.

rattler-build generate-recipe cran r-knitr

Building a R Package#

recipe.yaml

context:
  version: "1.47"

package:
  name: r-knitr
  version: ${{ version }}
  noarch: generic  # (4)!

source:
- url: https://cran.r-project.org/src/contrib/Archive/knitr/knitr_${{ version }}.tar.gz
  sha256: fadd849bf94a4e02520088a6626577c3c636227fe11c5cd7e8fcc5d51a7aa6cf

build:
  script: R CMD INSTALL --build .  # (1)!

requirements:
  host:
  - r-base  # (2)!
  - r-evaluate >=0.15
  - r-highr >=0.11
  - r-xfun >=0.44
  - r-yaml >=2.1.19
  run:
  - r-base
  - r-evaluate >=0.15
  - r-highr >=0.11
  - r-xfun >=0.44
  - r-yaml >=2.1.19

tests:
# This is a shorthand test for R packages to ensure that the library loads correctly.
- r:
    libraries:
      - knitr
# You can also run arbitrary R code in the test section.
- script: test_package.R  # (3)!

about:
  homepage: https://yihui.org/knitr/
  summary: A General-Purpose Package for Dynamic Report Generation in R
  description: |-
    Provides a general-purpose tool for dynamic report
    generation in R using Literate Programming techniques.
  license: GPL-2.0
  repository: https://github.com/cran/knitr

The script section is where you specify the build commands to run. In this case, we are using R CMD INSTALL --build . to build the package.
The r-base package is required to run R and is specified in the host requirements.
The script key automatically detects the language based on the file extension. In the case of .R, it will execute the R script with rscript.
The noarch: generic directive indicates that the package is architecture-independent. This is useful for R packages that do not contain compiled code and can run on any architecture. It allows the package to be installed on any platform without needing to rebuild it for each architecture.