NEWS

invivoPKfit 2.0.2 (2025-08-26)

Minor bug fix for adjusting rate constants after fitting to standard "1/hour".
Improved pk object organization and CLI message outputs.
Externalized parallel processing for fitting with mirai and purrr 1.1.0. Now, users should set the number of processes through those interfaces if parallel or distributed computing is desired.
Setting data, error, NCA, sd, and LOQ groups now uses unquoted data variable names, the latter two have their own separate pkproto object setups: stat_sd_group() and stat_loq_group().
Parameter starts and and estimates are stored as singular values in a list-column, which allows character vector constant parameters for calls to more complex PBTK models, see model_httk_gas_pbtk.
Users may now set specific parameters to optimize and specific start values using the new functions set_params_optimize() and set_params_starts().
Updated cvt object to align with latest CvTdb release.

Ensured only optimized parameters will be used when calculating AIC. This is really only a concern when using models that hold many of the total parameters constant.
Added a new feature allowing users to calculate log-likelihood from predictions generated outside invivoPKfit.
Updated documentation and cleaned up vignettes.

Initial CRAN submission.
Redesigned package to an S3 object-oriented approach with some ggplot2-like syntax.
Re-factored function/method code to ensure the code base is more legible to future contributors.
Implemented messages that save and update data processing "status".
Parallel processing parameter optimization was added as an option to take advantage of multi-processing cores.
Included more chemical records than previously published in the cvt object, represented a processed portion of the CvTdb database.