Skip to contents

Create a new bayesim_task_result object

Source: R/results.R
new_task_result.Rd

Constructs a result object from a single simulation task execution.

Usage

new_task_result(
  task_id = character(1),
  status = "success",
  metrics = NULL,
  diagnostics = NULL,
  timing = list(total = 0),
  error = NULL,
  warnings = character()
)

Arguments

task_id

Character scalar identifying the task

status

Character scalar: one of "success", "failed", or "skipped"

metrics

Named list of computed metrics (NULL if task failed or skipped)

diagnostics

Named list of diagnostic values (NULL if task failed)

timing

List containing timing information, must include total

error

NULL, or a list with error_class and error_message if failed

warnings

Character vector of warning messages

Value

A validated bayesim_task_result object

Details

The bayesim_task_result class captures the outcome of a single simulation task, including its computed metrics, any diagnostics, timing information, and errors or warnings.

Validation rules:

  • status must be one of "success", "failed", or "skipped"

  • If status is "success", metrics must not be NULL

  • If status is "failed", error must not be NULL

  • timing$total must be non-negative

Examples

# Successful task
result <- new_task_result(
  task_id = "task_001",
  status = "success",
  metrics = list(rmse = 0.05, bias = 0.01),
  diagnostics = list(n_eff = 500, rhat = 1.01),
  timing = list(total = 5.2)
)

# Failed task
result <- new_task_result(
  task_id = "task_002",
  status = "failed",
  error = list(error_class = "convergence_error", error_message = "R-hat > 1.1"),
  timing = list(total = 2.0)
)