Skip to contents

Discovers all timestamped run directories, loads their exported artifacts, and performs seven comparison analyses covering sample overlap, sentiment drift, code stability, theme evolution, entry migration, correlation stability, and a run summary dashboard.

Usage

compare_runs(current_dir, results_base, config = NULL, run_dirs = NULL)

Arguments

current_dir

Path to the current run's output directory

results_base

Path to the parent directory containing all run folders

config

ThematicConfig object (or NULL)

run_dirs

Optional explicit, ordered character vector of run directories to compare instead of discovering all runs under results_base. The LAST element is treated as the "current" run. compare_models uses this to compare a specific cross-model pair rather than the two chronologically-latest runs.

Value

A ComparisonResult S3 object, or NULL if fewer than 2 runs exist