Skip to contents

Build a model-configuration slot for a data list

Source: R/0-model_config.R
model_config.Rd

Bundles the model-structure arguments of fit_mod(), predation mode, initialization, harvest control rule, and the build_*() process specifications, into a single validated object that can be stored on a data list (data_list$model_config, e.g. via build_data(..., model_config = model_config(...))). A configuration then travels with the data instead of living only in the fit_mod() call.

Usage

model_config(
  msmMode = 0,
  initMode = "NonEquilibrium",
  avgnMode = 0,
  suitMode = 0,
  niter = 3,
  HCR = build_hcr(),
  recFun = build_srr(),
  M1Fun = build_M1(),
  growthFun = build_growth(),
  qFun = build_catchability(),
  selFun = build_selectivity(),
  compFun = build_composition()
)

Arguments

msmMode

Predation-mortality mode (see fit_mod()). Default 0.

initMode

Population initialization (see fit_mod()). Default "NonEquilibrium".

avgnMode

Average-N mode. Default 0.

suitMode

Suitability mode. Default 0.

niter

Number of predation iterations. Default 3.

HCR

A harvest-control-rule specification from build_hcr().

recFun

A stock-recruit specification from build_srr().

M1Fun

A natural-mortality specification from build_M1().

growthFun

A growth specification from build_growth().

qFun

A catchability specification from build_catchability().

selFun

A selectivity specification from build_selectivity().

compFun

A composition specification from build_composition().

Value

An object of class "Rceattle_model_config" (a named list) to be stored as data_list$model_config.

Details

The defaults are exactly fit_mod()'s own argument defaults, so a data list holding model_config() fits identically to one with no slot at all. When a data list has a model_config, fit_mod() reads each field only for arguments the caller did not pass explicitly, an argument passed to fit_mod() always wins, even when passed at its default. Omit an argument to let the stored configuration take effect for that field.

Persistence

The configuration is code-side model structure, not one of the workbook data sheets, so it is not written by write_data() and does not survive an xlsx round-trip, build_data(base = x, model_config = cfg) piped through write_data() then read_data() returns without the slot. Re-attach it in code, store it alongside the data, or persist it as a documented, git-diffable YAML with save_config() / load_config() and apply it with fit_mod(data_list, config = load_config("run.yaml")).

See also

fit_mod(), build_data(), save_config(), load_config().

Examples

cfg <- model_config(msmMode = 1, initMode = "NonEquilibrium")
dat <- build_data(base = BS2017MS, model_config = cfg)
# fit_mod(dat) would then fit as multispecies without passing msmMode.