Compute all metrics for a task — compute_all_metrics • bayesim

Iterates through all metrics and computes their values, handling errors gracefully based on each metric's required property.

Usage

compute_all_metrics(
  fit_result,
  data_bundle,
  context,
  metrics,
  task_ctx,
  result_path = NULL
)

Arguments

fit_result: A bayesim_fit_result object from a successful fit
data_bundle: A validated data bundle list
context: A list with precomputed values (from build_metric_context)
metrics: List of S7 Metric objects
task_ctx: Task context with identification information

Value

A named list with:

metrics: Named list of computed and flattened metric values
warnings: Character vector of any warning messages

Details

For each metric:

If the metric has required = TRUE and fails, the error is re-thrown
If the metric has required = FALSE and fails, NA values are returned with error information stored in <metric_name>__error_class and <metric_name>__error_message
All metric outputs are flattened using flatten_metric_output()