Changes in version 0.1.0.9000                      

  - read_ema() no longer errors when a timestamp cannot be parsed at
    all; such records now come back as NA with the documented warning,
    so one corrupted row cannot take down the import of an otherwise
    good export.
  - The README now distinguishes nof1kit from overlapping single-case
    packages and documents the privacy and ethics considerations of the
    bundled self-experiment data.
  - Added an R-native inst/CITATION entry alongside CITATION.cff.

                 Changes in version 0.1.0 (2026-08-08)                  

First release.

  - design_schedule(): exact uniform sampling of balanced,
    run-constrained randomization schedules by dynamic programming,
    reproducible from a required seed.
  - check_schedule(): design diagnostics (counts, run lengths,
    alternations, lag-1 autocorrelation).
  - write_schedule(): JSON export in the format read by the companion
    mobile collection tools.
  - sim_power(): simulation-based power and false-positive rates under
    AR(1) errors, analysed with and without the dependence, for both
    randomized and user-supplied schedules.
  - read_ema(): tidy import of EMA exports (CSV or JSON), with a
    study-day index derived from the start date and an error when a
    file's own index disagrees with the calendar.
  - validate_ema(): structured integrity checks, including day indices
    that have drifted out of step with the calendar.
  - compliance(): prompt-level compliance against response windows, with
    off-window records reported separately.
  - Vignette running the full lifecycle on the bundled 70-day study.