Changes in version 0.3.3 (2026-07-22)                  

Infrastructure

  - Add pre-commit hooks for repository hygiene: check-added-large-files
    (2 MB cap, --enforce-all) and detect-private-key from
    pre-commit/pre-commit-hooks (v6.0.0), and codespell (v2.4.2) for
    spell-checking. codespell runs with --write-changes and an
    --ignore-words-list for domain terms (acn, ths); renv.lock and
    release_github.sh are excluded to avoid false positives on generated
    content and shell identifiers.
  - Fix two spelling typos surfaced by codespell, in NEWS.md and a
    comment in R/glimpse_utils.R.

                        Changes in version 0.3.2                        

Bug fixes

  - Fix generate_study_config() writing unparsable YAML on Windows.
    Paths were emitted as double-quoted scalars, so backslash sequences
    in Windows paths (e.g. C:\Users\...) were interpreted as YAML
    escapes and simulate_study() failed to read the config. Paths are
    now single-quoted, which passes backslashes through verbatim.

                        Changes in version 0.3.1                        

Infrastructure

  - Add a pkgdown GitHub Actions workflow
    (.github/workflows/pkgdown.yaml) that builds the documentation site
    and deploys it to the gh-pages branch, and set the site url in
    _pkgdown.yml.
  - Add an R-hub v2 GitHub Actions workflow
    (.github/workflows/rhub.yaml) for multi-platform CRAN checks,
    dispatched via rhub::rhub_check().

                        Changes in version 0.3.0                        

Summary

  - Split the study workflow into a glimpse phase and a simulate phase:
    add glimpse_study() and simulate_study_from_summary(), allowing the
    glimpse (which needs the real .sas7bdat files) to run independently
    from simulation.
  - Fall back to a default seed of 123 when a seed is not set in the
    study config, and make the seed argument optional in
    generate_study_config() (also defaulting to 123).
  - Add .assert_required_id_cols() to safeguard against ID leakage. It
    now errors if USUBJID/SUBJID are present in the data but missing
    from id_cols, ensuring subject IDs are regenerated synthetically
    rather than value-sampled from real data.
  - Refactor generate_study_config() to take a required output_dir (with
    an optional config_yaml_name), writing the config into that
    directory alongside where synthetic data will be saved.
  - Drop the synadam::/synadam::: syntax to avoid the CRAN self-:::
    NOTE.
  - Add runnable roxygen @examples across the exported
    glimpse_*/simulate_* functions. Introduce a shared
    man-roxygen/setup_adam_datasets.R template that stages bundled
    fixtures.
  - Relocate the MIT license text to LICENSE.md and switch LICENSE to
    the YEAR/COPYRIGHT HOLDER format, raise the minimum R version
    to 4.1.0 (native pipe), and extend .Rbuildignore.

                        Changes in version 0.2.0                        

New features

  - Keep all PARAMCDs in BDS simulation instead of requiring a
    paramcd_filter argument.
  - Simulate flag columns based on the proportion of "Y"/"N" values
    rather than preserving exact positions.

                        Changes in version 0.1.0                        

  - Initial release.