Parse a NONMEM report file (.lst/.res) — parse

Extracts the embedded control stream and the final parameter estimates (THETA vector, OMEGA and SIGMA covariance matrices, and the objective function value) from a NONMEM report/output file. This is useful when only the report file is available and the .ext file (which pharmpy normally reads final estimates from) is missing.

Usage

parse_output(output_file = NULL, code = NULL)

Arguments

output_file: path to a NONMEM report file (.lst, .res, ...).
code: character string with the report file contents (alternative to output_file).

Value

a list with elements control_stream (character), theta (numeric vector), omega (numeric matrix or NULL), sigma (numeric matrix or NULL), and ofv (numeric scalar or NA).

Details

Precision: the FINAL PARAMETER ESTIMATE block in a .lst is rounded to roughly three significant figures, so estimates recovered this way are lower precision than those in the corresponding .ext file. Use the .ext file (via create_model_from_file()) when full precision is required.

Standard errors and covariance-step output are not parsed. OMEGA/SIGMA parsing handles the single column block NONMEM prints for up to six random effects; wider matrices, which NONMEM splits across multiple column blocks, are not yet supported.