Add GitHub Action to regenerate example_data when convert_output.R changes

[Copilot]

data/example_data.rda had no automated path to stay in sync when convert_output.R was updated — it required manual execution of data-raw/DATASET.R.

Changes

• .github/workflows/update-example-data.yml — new workflow that:
  • Triggers on push to main with path filter R/convert_output.R, plus workflow_dispatch
  • Installs R dependencies (including usethis) via r-lib/actions
  • Regenerates example_data using the same pattern as data-raw/DATASET.R to preserve the correct object name in the .rda:

    example_data <- stockplotr::convert_output(file = "inst/extdata/Report.sso")
    usethis::use_data(example_data, overwrite = TRUE)

  • Commits data/example_data.rda as github-actions[bot] with [skip ci] to avoid loops; skips the commit entirely if the data is unchanged
  • Requires contents: write permission to push back to the branch

[Schiano-NOAA]

@copilot by using this action, is the script data-raw/DATASET.R unnecessary?

[Copilot]

@copilot by using this action, is the script data-raw/DATASET.R unnecessary?

Not entirely — data-raw/DATASET.R still serves two purposes:

1. R package convention / documentation: The data-raw/ directory is the standard place to keep scripts that document how a package's data/ files were generated. Tools like usethis::use_data_raw() create it for exactly this reason, and CRAN/package reviewers expect it.
2. Local development: Developers can still run it by hand when they want to regenerate the data locally without pushing to main.

The GitHub Action automates the same regeneration in CI, but keeping data-raw/DATASET.R is still recommended as the human-readable record of how example_data.rda is produced. You could remove the devtools::install_github(...) line from it (since locally you'd use devtools::load_all() instead), but the script itself is worth keeping.

[github-actions]

Code Metrics Report

Coverage	Code to Test Ratio	Test Execution Time
69.3%	1:0.2	4m30s

---

Reported by octocov