mlr_assertions: Assertion for mlr3 Objects

Description

Functions intended to be used in packages extending mlr3. Most assertion functions ensure the right class attrbiture, and optionally additional properties. Additionally, the following compound assertions are implemented:

assert_learnable(task, learner) (Task, Learner) -> NULL Checks if the learner is applicable to the task. This includes type checks on the type, the feature types, and properties.

If an assertion fails, an exception is raised. Otherwise, the input object is returned invisibly.

Usage

assert_backend(b, .var.name = vname(b))
assert_task(
  task,
  task_type = NULL,
  feature_types = NULL,
  task_properties = NULL,
  .var.name = vname(task)
)
assert_tasks(
  tasks,
  task_type = NULL,
  feature_types = NULL,
  task_properties = NULL,
  .var.name = vname(tasks)
)
assert_learner(
  learner,
  task = NULL,
  properties = character(),
  .var.name = vname(learner)
)
assert_learners(
  learners,
  task = NULL,
  properties = character(),
  .var.name = vname(learners)
)
assert_learnable(task, learner)
assert_measure(
  measure,
  task = NULL,
  learner = NULL,
  .var.name = vname(measure)
)
assert_measures(
  measures,
  task = NULL,
  learner = NULL,
  .var.name = vname(measures)
)
assert_resampling(
  resampling,
  instantiated = NULL,
  .var.name = vname(resampling)
)
assert_resamplings(
  resamplings,
  instantiated = NULL,
  .var.name = vname(resamplings)
)
assert_prediction(prediction, .var.name = vname(prediction))
assert_resample_result(rr, .var.name = vname(rr))
assert_benchmark_result(bmr, .var.name = vname(bmr))
assert_row_ids(row_ids, null.ok = FALSE, .var.name = vname(row_ids))