Skip to contents

Changelog

Source: NEWS.md

lstparsR 0.1.2

  • Preserve scientific exponents and missing matrix diagonals, and read printed ETA shrinkage with or without an equals sign.
  • Select a consistent final estimation step across parameter and scalar results.
  • Handle truncated objective output, grouped eigenvalues and singular matrices.
  • Preserve duplicate uploads, expose incomplete results, and improve table layout and download isolation.
  • Clarify supported output layouts and first-submission metadata.

lstparsR 0.1.1

Bug fixes

  • fetch_ofv(): No longer warns “OFV not found” for $SIMULATION-only runs (e.g. VPC simulation outputs). Such runs legitimately produce no objective function value; the function now returns NA_real_ silently in that case. The warning is still emitted when an estimation step ran but its OFV cannot be parsed.
  • fetch_condn(): No longer warns “Eigenvalue section not found” when the user set EIGENVLS. PRINTED: NO in $COV. In that case the covariance step ran but the eigenvalue block is intentionally absent; the function now returns NA_real_ silently. The warning is still emitted when eigenvalues are expected but cannot be located.
  • Reported by Christiane on multiple unrelated projects.

Internal

  • New internal helpers .has_estimation_step() and .eigenvalues_suppressed() in R/utils-internal.R.

lstparsR 0.1.0

  • Initial development release.
  • read_lst_file(): reads NONMEM .lst files; returns S3 class "lst".
  • fetch_thetas(): THETA estimates with SE and RSE.
  • fetch_etas(): OMEGA diagonal with SE, RSE, and ETA shrinkage.
  • fetch_sigmas(): SIGMA diagonal with SE and RSE.
  • fetch_ofv(): objective function value with fallback to workflow footer.
  • fetch_condn(): condition number from eigenvalues.
  • fetch_all(): convenience wrapper returning all of the above.
  • run_app(): interactive Shiny application for uploading, viewing, and downloading parsed results.
  • Recognizes FOCE-I, FOCE, FO, SAEM, IMP, IMPMAP, and Bayesian method headers.
  • Handles multi-line parameter blocks (>12 THETAs spanning multiple rows).
  • Missing optional covariance quantities are returned as NA.
  • Exercised against more than 7,600 listing files, predominantly FOCE-I.