Machine-readable diagnostics for an agent experiment
Source:R/methods_llmr.R
diagnostics.agent_experiment.RdReturns the cell-level health numbers behind an agent_experiment() result:
how many cells ran, how many failed, the failure rate, and the total
wall-clock seconds. When the frame carries a rep column, the number of
distinct conditions and the replication count are reported too. Robust to a
frame missing the error, duration, or rep columns.
Usage
# S3 method for class 'agent_experiment'
diagnostics(x, ...)Arguments
- x
An
agent_experiment()result (a tibble of classagent_experiment).- ...
Unused.
Value
A one-row tibble with n_cells, n_failed, n_ok,
failure_rate, total_duration_s, and (when a rep column is present)
n_conditions and reps.