                     Changes in version 1.6.0.9000                      

  - Added sample_nhanes(), which simulates virtual pediatric populations
    (ages 2-17) with realistic body weight and/or height by MEC-weighted
    resampling of children from four NHANES releases (2013-2023),
    bundled as nhanes_peds. Resampling keeps each child's weight and
    height together; method = "smooth" adds correlated kernel noise for
    continuous values. compare_nhanes() and plot_nhanes() check the
    simulated population against the reference. A new vignette,
    vignette("nhanes", package = "pmxTools"), walks through them.

                  Changes in version 1.6 (2026-10-07)                   

  - Added cut_quantile(), which creates quantile-based bins for one or
    more continuous variables, with per-variable control over the number
    of groups.

  - Added datamap(), which plots individual dosing and observation
    records over time.

  - dgr_table() now takes fields as a named character vector mapping
    column names to display labels. The separate names argument is
    deprecated, but unnamed fields with names continues to work.

  - Added dependencies purrr and tidyr.

  - Fixed read_nm(), which passed the file extension to xml2::read_xml()
    as an encoding rather than appending it to the file name, and read
    the unresolved file name instead of the one located on disk. The
    directory argument and file names given without the .xml extension
    were therefore ignored, and macOS reported an Unsupported encoding:
    .xml warning (#36).

  - Removed the blanket stats and utils imports, which caused a
    replacing previous import 'stats::filter' by 'dplyr::filter' warning
    when the package was loaded (#36).

  - calc_derived() and the calc_derived_*() functions now return AUCtau
    and AUCtau_dose_normalized as plain numbers. PKNCA 0.12.1.9000
    attaches a method attribute to the AUC it returns, which propagated
    into the derived-parameter list and broke comparisons against plain
    numeric values (#37).

  - Test snapshots are now committed, the CI workflows have been
    refreshed, and the documentation has been regenerated with
    roxygen2 8.1.0 (#36).

  - calc_derived() and the calc_derived_*() functions now work with
    vector inputs (#29).

                  Changes in version 1.5 (2025-08-25)                   

  - plot_dist has been rewritten to adapt to ggplot2 4.0.0 changes, and
    to make it more intuitive to interpret

  - Dependency gghalves has been removed

  - Many fixes and updates to documentation

                         Changes in version 1.4                         

  - dgr_table() now provides geometric means by default

  - gm() now provides options to strip NA values (na.rm) and
    non-positive values (neg.rm) before computation.

                  Changes in version 1.3 (2023-02-21)                   

  - Added NCA parameter estimation to calc_derived_1cpt(),
    calc_derived_2cpt() and calc_derived_3cpt(), if dose and other
    required information (e.g. tinf, dur, tau) is provided.

  - Added a warning about flip-flop kinetics and their potential effects
    on derived half-lives to calc_derived_1cpt(), calc_derived_2cpt(),
    and calc_derived_3cpt().

  - Added blq_trans() and blq_log_trans() for assistance with
    visualization of measurements below a limit of quantification.

  - Added a test for valid XML to read_nm().

  - Small fixes and enhancements to read_nmtable_single() and
    read_nm_multi_table() to ensure that stray text is properly handled
    while reading in NONMEM output files.

  - count_na() now throws a warning if NaN values are included in the NA
    count, and indicates how many of them there are.

  - Changed the significant digits functionality for 95% confidence
    intervals for get_theta(), get_omega() and get_sigma(), and all
    functions that apply these.

  - Replaced gridExtra with patchwork.

  - Numerous small fixes.

  - Removed some unnecessary dependencies.

  - Added many new unit tests, increasing test coverage considerably.

                        Changes in version 1.2.4                        

  - Fixed some minor documentation issues. Thanks to Julien Grassot for
    spotting these.

                 Changes in version 1.2.3 (2022-04-06)                  

  - Fixed another error in sample_uncert() which was still crashing the
    function.

  - Added helper functions count_na(), dgr_table(), gcv(), fmt_signif().

  - Added distribution plotting function plot_dist().

                        Changes in version 1.2.2                        

  - Fixed an error in sample_uncert() which was crashing the function.

                 Changes in version 1.2.1 (2020-08-26)                  

  - Added vignette describing PK curves.

  - Fixed a systematic error in calc_sd_1cmt_linear_oral_0_lag(),
    calc_ss_1cmt_linear_oral_0_lag(), calc_sd_2cmt_linear_oral_0_lag()
    which resulted in incorrect curves being plotted.

                         Changes in version 1.2                         

  - Rewrote plot_scm function to produce tree diagrams via DiagrammeR.

  - Fixed a rare bug in get_auc in which measurements from different
    individuals could be erroneously mixed.

  - Amended (or added) read_nm(), read_nm_all(), read_nmext(),
    read_nmcov(), read_nmtables() functions to allow reading of multiple
    NONMEM estimation steps.

  - Clarified documentation for calc_derived().

  - Now using xml2 to read NONMEM-generated XML files.

  - Fixed partial argument match warnings encountered during tests.

  - Add function for calculting geometric CV.
