Split bolus doses into bolus and infusion paths (model directive)

splitBolusInfusion() is a model-only directive that rewrites doses aimed at cmt into parallel doses to the target compartments, regardless of whether the dose record is a bolus or an infusion. Each target compartment receives the full original amount; use f() to scale the split.

A plain bolus record is split so the LAST target compartment receives a modeled infusion start/stop pair (it must declare a modeled dur() or rate() property, otherwise its copies stay boluses) while every other target receives a bolus copy. Use splitInfusionBolus() to give the FIRST target the infusion instead. This feeds one bolus dose record into both a bolus path and an infusion path (Monolix-style double absorption with mixed first- and zero-order routes). Steady-state bolus records are not promoted; they are copied as boluses with a warning when the designated infusion target declares dur()/rate().

Infusion records (data RATE/DUR or modeled rate()/dur()) are split preserving their type, exactly like splitInfusion().

This rewrite applies at etTrans() translation time only; doses pushed while solving with evid_() are not split. Only one splitting directive (splitBolus(), splitInfusion(), splitInfusionBolus() or splitBolusInfusion()) is allowed per model.

There is no R function named splitBolusInfusion() exported by rxode2; the directive is parsed from the model text directly, as shown below.

Usage inside a model

model({
  splitBolusInfusion(depot, depot2, central)
  dur(central) <- tk0
  f(depot2) <- 1 - f1
  f(central) <- f1
  d/dt(depot2) <- -ka2 * depot2
  d/dt(central) <- ka2 * depot2 - cl / v * central
  ...
})

A bolus dose recorded against depot is split so depot2 receives the bolus (first-order input scaled by 1 - f1) and central (last target) receives a modeled-duration infusion (zero-order input over tk0, scaled by f1).