Skip to contents

Contributing to ofemeantest

Source: CONTRIBUTING.md

Thanks for your interest in improving the package. This document covers the local dev setup we expect contributors to have before opening a PR.

Development setup

  1. Clone and open the project in your R IDE:

    git clone https://github.com/PPaccioretti/ofemeantest.git
cd ofemeantest

  2. Install the dev dependencies (one-off):

    install.packages(c("devtools", "roxygen2", "testthat", "pkgdown"))

  3. Enable the project’s git hooks (one-off per clone). This points git at the versioned .githooks/ directory so the pre-push check runs on every git push:

    git config core.hooksPath .githooks

    You can verify it took effect with git config --get core.hooksPath (should print .githooks).

What the pre-push hook does

Before every git push the hook runs devtools::document() and checks whether anything in man/, NAMESPACE or DESCRIPTION changed. If it did, the push is blocked with a list of files to commit. This prevents the most common source of “stale docs” PRs — editing a roxygen comment in R/foo.R and forgetting to regenerate the .Rd.

To bypass the hook for a single push (e.g., emergency hotfix where the doc regen is unrelated):

git push --no-verify

The same check runs in GitHub Actions on every PR (.github/workflows/document.yaml), so even pushes that bypass the local hook will be caught at review time.

Running tests locally

devtools::test()       # run testthat suite
devtools::check()      # full R CMD check
pkgdown::build_site()  # rebuild the docs site under docs/

The docs/ directory is gitignored — it’s only used for local preview and is rebuilt by CI on push to main.

Style

  • Keep functions documented with roxygen2; mark internals with @keywords internal.
  • Add tests for new public behaviour. Contract tests in tests/testthat/test-dependency-contracts.R guard the interfaces we rely on from spdep, spatialreg, permuco and multcompView — add a new entry there when you start depending on a new upstream function.
  • Prefer base R over dplyr/tidyr to keep the Imports list small.