/rap-init — Create a Reproducible Analytical Pipeline

/rap-init — Create a Reproducible Analytical Pipeline

User Input

$ARGUMENTS

Instructions

You are scaffolding a new R project that follows the UK Government Reproducible Analytical Pipelines (RAP) methodology. The generated project will immediately satisfy all Baseline RAP requirements and provide scaffolding toward Silver.

1. Parse project name

Extract the project name from $ARGUMENTS.

• If provided, sanitise it: lowercase, replace spaces with hyphens, remove special characters.
• If empty, use the current directory name as the project name.
• Store as PROJECT_NAME.

2. Check the target directory

Inspect the current working directory for existing files:

ls -A

• If the directory already contains R files, DESCRIPTION, or renv.lock, warn the user that files exist and ask for confirmation before proceeding.
• If .git/ already exists, note this — skip git init later.

3. Create directory structure

Create the RAP project directories:

mkdir -p R data output tests/testthat reports

4. Write template files

For each template file listed below, read the template from the plugin's template directory, replace all occurrences of {{PROJECT_NAME}} with the actual project name, and write the result to the target path.

Template Source	Target Path
plugins/rap/templates/r-project/DESCRIPTION.template	DESCRIPTION
plugins/rap/templates/r-project/README.md.template	README.md
plugins/rap/templates/r-project/.gitignore.template	.gitignore
plugins/rap/templates/r-project/R/01_read_data.R.template	R/01_read_data.R
plugins/rap/templates/r-project/R/02_process_data.R.template	R/02_process_data.R
plugins/rap/templates/r-project/R/03_analyse_data.R.template	R/03_analyse_data.R
plugins/rap/templates/r-project/R/04_create_output.R.template	R/04_create_output.R
plugins/rap/templates/r-project/data/README.md.template	data/README.md

Additionally, create the RStudio project file:

File: PROJECT_NAME.Rproj

Version: 1.0

RestoreWorkspace: No
SaveWorkspace: No
AlwaysSaveHistory: No

EnableCodeIndexing: Yes
UseSpacesForTab: Yes
NumSpacesForTab: 2
Encoding: UTF-8

AutoAppendNewline: Yes
StripTrailingWhitespace: Yes

BuildType: Package
PackageUseDevtools: Yes

5. Execute setup commands

Run the following commands to initialise the project environment:

Git initialisation (only if .git/ does not already exist):

git init

renv initialisation (only if renv.lock does not already exist):

Rscript -e "renv::init(bare = TRUE)"

Use bare = TRUE so renv creates the infrastructure without trying to discover and install packages from the skeleton scripts. The user will add packages incrementally.

If R is not available on the system, skip renv::init() and inform the user they should run it manually when R is available.

6. Display summary

After all files are created, display a summary:

## RAP Project Created: PROJECT_NAME

### Files created

- `DESCRIPTION` — R package metadata
- `README.md` — Project documentation with reproduction steps
- `.gitignore` — Version control exclusions (protects sensitive data)
- `PROJECT_NAME.Rproj` — RStudio project file
- `R/01_read_data.R` — Data ingestion functions
- `R/02_process_data.R` — Data cleaning and transformation
- `R/03_analyse_data.R` — Statistical analysis
- `R/04_create_output.R` — Output generation
- `data/README.md` — Data provenance documentation

### Directories created

- `R/` — Analysis functions (numbered for pipeline order)
- `data/` — Input data (excluded from Git)
- `output/` — Generated outputs (excluded from Git)
- `tests/testthat/` — Unit tests
- `reports/` — R Markdown reports

### RAP Compliance

Your project now satisfies these **Baseline RAP** requirements:
- ✓ **B1**: Code in an open-source language (R)
- ✓ **B2**: Version controlled with Git
- ✓ **B3**: README with reproduction steps

To complete Baseline, you also need:
- ○ **B4**: Peer review (open a PR for review)
- ○ **B5**: Publish openly (push to GitHub/GitLab)

### Next steps

1. Edit `data/README.md` to document your data sources
2. Update `R/01_read_data.R` with your actual data ingestion code
3. Run `/rap-check` to audit your RAP compliance level
4. Run `/rap-test` to set up unit testing (Silver requirement)
5. Run `/rap-pipeline` to configure {targets} build automation

7. Edge Case Handling

Handle these situations gracefully:

• Git not installed: If git init fails with "command not found", display:

  Git is not installed. Install Git from https://git-scm.com/ then run `git init` manually.
  Your project files have been created successfully — Git can be added later.
  

• R not installed: If Rscript is not available, skip renv::init() and display:

  R is not installed or not on PATH. Install R from https://cran.r-project.org/
  After installing R, run: Rscript -e "renv::init(bare = TRUE)"
  

• renv not installed: If renv::init() fails with "there is no package called 'renv'", display:

  The {renv} package is not installed. Run in R:
    install.packages("renv")
    renv::init(bare = TRUE)
  

• Permission errors: If file creation fails, display the specific error and suggest checking directory permissions: ls -la .

• packrat detected: If packrat/ directory exists, warn:

  This project uses {packrat} for dependency management. Consider migrating
  to {renv} which is its successor: https://rstudio.github.io/renv/articles/packrat.html
  Skipping renv::init() to avoid conflicts.
  

• Non-empty directory with conflicts: If existing files would be overwritten, list them and ask for confirmation before proceeding. Never silently overwrite.

8. RAP Context

Include this brief context at the end of your response:

What is RAP? Reproducible Analytical Pipelines apply software engineering practices to statistical production. Developed by the UK Government Analysis Function, RAP ensures official statistics are produced transparently, reproducibly, and efficiently.

Learn more:

• UK Government Analysis Function RAP page
• NHS RAP Community of Practice
• Building RAPs with R
• RAP Companion