Run a pharmpy nlmixr-format model with nlmixr2 — run_nlme

Internal companion to run_nlme(); called when the input model is in pharmpy's nlmixr backend. The function turns the pharmpy nlmixr model into an rxode2/nlmixr2 function, fits it with nlmixr2::nlmixr2(), and wraps the result in a list that mimics the shape of a pharmpy ModelfitResults so downstream helpers (attach_fit_info(), compare_nlme_fit(), etc.) work without engine-specific branches.

Usage

run_nlme_nlmixr(
  model,
  data = NULL,
  id,
  path = getwd(),
  estimation_method = NULL,
  control = NULL,
  force = NULL,
  save_fit = TRUE,
  save_summary = TRUE,
  save_final = TRUE,
  clean = TRUE,
  mu_reference = "auto",
  verbose = TRUE
)

Arguments

model: pharmpy model object or NONMEM model code (character) or path to NONMEM model file.
data: filename of dataset or data.frame as input to NONMEM / nlmixr. Optional, can also be included in model object (if specified as pharmpy model object).
id: run id, e.g. run1. This will be the folder in which the NONMEM model is run. If no folder is specified, it will create a folder run1 in the current working directory, and will increment the run number for each subsequent run.
path: path to nonmem model. If not specified, will assume current working directory.
estimation_method: Optional. Character vector of estimation method(s) to apply to model. Will remove all existing estimation steps in the model and update with methods specified in argument.
control: nlmixr2-only. Optional control list passed verbatim to nlmixr2::nlmixr2() (e.g. nlmixr2est::foceiControl() or nlmixr2est::saemControl()). Ignored for NONMEM models.
force: if run folder (id) exists, should existing results be removed before rerunning NONMEM? Default FALSE.
save_fit: save fit object. If TRUE, will save as <id>.rds inside path. Can also specify a filename (rds) to save to; relative filenames are resolved against path, absolute ones are used as-is. FALSE writes nothing.
save_summary: save fit summary and parameter estimates to file? Default is TRUE. Files are written to path as <id>_fit_summary.txt and <id>_fit_parameters.csv.
save_final: after running the model, should a file final.mod be created with the final estimates from the run.
clean: clean up run folder after NONMEM execution?
mu_reference: Controls mu-referencing for SAEM models. "auto" (default) automatically applies pharmr::mu_reference_model() when SAEM is used and the model is not already mu-referenced. TRUE always applies mu-referencing. FALSE never applies mu-referencing (old behaviour: warns when SAEM is used without mu-referencing).
verbose: verbose output?

Details

Pharmpy can in principle drive nlmixr2 itself, but that path requires the Python pyreadr module which is not part of the standard install; calling nlmixr2 directly is also faster (no R→Python→R round-trip).