Run the complete validation-release programme

Usage

run_eyeprocess_validation_program(
  corpus,
  output_dir,
  model_jobs = list(),
  sbc_jobs = list(),
  engine_jobs = list(),
  reproduction_jobs = list(),
  grouped_jobs = list(),
  leakage_jobs = list(),
  multiverse_jobs = list(),
  benchmark_jobs = list(),
  reporting_dataset = NULL,
  public_benchmark_dataset = NULL,
  public_benchmark_include_samples = FALSE,
  advanced_evidence = list(),
  evidence_spec = advanced_model_evidence_spec(),
  overwrite = FALSE
)

Arguments

corpus: Validation corpus or manifest.
output_dir: Output directory.
model_jobs: Named list of model-validation job specifications. Each job supplies `simulator`, `fitter`, `extractor`, `truth_extractor`, and optional `grid` and `spec`.
sbc_jobs: Named list of simulation-based-calibration job specifications.
engine_jobs: Named list of equivalent-engine comparison jobs.
reproduction_jobs: Named list of licensed empirical-reproduction jobs.
grouped_jobs: Named list of grouped-validation jobs. Set `crossed = TRUE` in a job to call `crossed_grouped_cv()`.
leakage_jobs: Named list of leakage-quantification jobs.
multiverse_jobs: Named list of preprocessing-multiverse jobs.
benchmark_jobs: Named list of zero-argument functions or benchmark argument lists.
reporting_dataset: Optional `eye_dataset` for reporting-guideline coverage.
public_benchmark_dataset: Optional `eye_dataset` from which to write a de-identified public benchmark bundle.
public_benchmark_include_samples: Whether the public benchmark retains sample-level tables.
advanced_evidence: Optional named evidence records keyed by model function.
evidence_spec: Advanced-model promotion-gate specification.
overwrite: Whether to replace the output directory.

Value

An `eye_validation_program` object.