Plot model convergence status table — plot_model_convergence

Creates a visual status table showing convergence diagnostics with color-coded pass/warn/fail indicators for each metric. Works with both likelihood-based and loss-based calibration methods by auto-detecting the appropriate diagnostics file.

Usage

plot_model_convergence_status(
  results_dir,
  plots_dir = NULL,
  status = NULL,
  verbose = TRUE
)

Arguments

results_dir: Path to results directory containing convergence diagnostics (expects either "convergence_diagnostics.json" or "convergence_diagnostics_loss.json")
plots_dir: Path to plots directory (default: "../plots" relative to results_dir)
status: Optional precomputed result list from calc_model_convergence_status. When supplied the recomputation and CSV write are skipped.
verbose: Logical indicating whether to print messages

Value

Invisible NULL. Creates a PDF file with the convergence status table.

Details

The status-table assembly and the convergence_status.csv write live in calc_model_convergence_status; this function renders the PDF table from that result. Pass a precomputed status object to skip recomputation (and the CSV write).

The function automatically detects whether to use likelihood-based or loss-based diagnostics by checking for the presence of diagnostics files in this order:

convergence_diagnostics.json (likelihood-based, e.g., calibration_test_10.R)
convergence_diagnostics_loss.json (loss-based, older methods)

Examples

if (FALSE) { # \dontrun{
# Works with likelihood-based calibration (calibration_test_10.R)
plot_model_convergence_status(
  results_dir = "./local/calibration/calibration_test_10/results"
)

# Also works with loss-based calibration (older methods)
plot_model_convergence_status(
  results_dir = "./local/calibration/calibration_test_9/results"
)
} # }