Skip to contents

Extract all parameter comments from a model as ModelComments object

Source: R/comments-parsing.R
get_model_parameter_info.Rd

Parses parameter comments from a completed NONMEM model run and returns structured metadata for theta, omega, and sigma parameters.

Usage

get_model_parameter_info(mod, lookup_path = NULL)

Arguments

mod

A hyperion_nonmem_model object or path to a run output directory containing an .lst file.

lookup_path

Optional path to a TOML lookup file. If provided, fills NULL fields (display, description, unit, parameterization) from the lookup.

Value

A ModelComments object containing theta, omega, and sigma comments.

Comment Parsing Modes

The parsing behavior is controlled by the pharos.toml configuration file. In the [nonmem.comments] section, set type = "type1" to enable structured type1 comment parsing. If this setting is absent or set to any other value, raw comment parsing is used (the default).

type1 mode: Expects comments in a structured format with explicit field delimiters. This mode provides more precise extraction but requires comments to follow the type1 specification.

raw mode (default): Flexibly parses parameter names, units, and descriptions from free-form comment text. More forgiving but may be less precise for complex comment structures.

Raw Comment Formats (default)

Applies to text after ; on lines within $THETA, $OMEGA, and $SIGMA blocks.

Parsing pipeline in raw mode: extract transform -> strip prefix -> extract unit -> extract name.

THETA/SIGMA

General form: [PREFIX] NAME [(UNIT)] [TRANSFORM_SEP TRANSFORM]

Common accepted examples:

  • CL

  • CL (L/HR)

  • CL [L/HR]

  • CL ;exp

  • CL :LOG

  • CL (L/HR) ;exp

  • THETA1 CL (L/HR) ;exp

  • 1: CL (L/HR) ;exp

Notes:

  • Prefixes are case-insensitive; colon after prefix is optional.

  • Numeric prefixes like 1, 1:, 1-, 1. are accepted.

  • Units are read from () or [], and can appear anywhere in the string.

  • Colon transform form requires leading whitespace (e.g., CL :EXP).

  • THETA strips trailing punctuation from extracted name tokens; SIGMA currently does not.

  • SIGMA also supports unit in transform segment, e.g. Name ;Transform (unit).

OMEGA

General form: [PREFIX] NAME_PART [THETA_REF] [TRANSFORM_SEP TRANSFORM]

Common accepted examples:

  • IIV-CL :EXP

  • IIV CL ;exp

  • IIV on CL ;exp

  • Corr CL-V ;normal

  • 11: IIV CL ;exp

  • OMEGA(1,1): IIV CL ;exp

Notes:

  • THETA_REF may split on -, /, :, , into multiple associated thetas.

  • If full THETA_REF matches a known theta name (case-insensitive), it is kept whole (for example, CL/F).

  • Linking words on, for, of are skipped in space-separated forms.

Transform keyword mapping (case-insensitive):

  • exp, log, lognormal -> LogNormal

  • logit -> Logit

  • add, adderr, additive -> AddErr

  • logadd, logadderr, logerr -> LogAddErr

  • prop, proportional -> Proportional

  • identity, normal, none -> Identity

See also

get_parameter_transform(), get_theta_names(), get_comment()