Run Original Dataset Exact Model Bank Validation

This page documents the runner used by the faithful RCIM Model-Bank Reproduction original-dataset exact-model-bank campaigns. The accepted forward/backward campaign outputs are archived under models/paper_reference/rcim_track1 and reported in doc/reports/analysis/rcim_paper_reference/RCIM Paper Reference Benchmark.md.

The current full-dataset RCIM Model-Bank Reproduction closure uses this runner as the faithful forward/backward model-bank surface. Yellow and red benchmark cells are kept as evidence, not as blockers for the closed faithful reproduction.

Run the original-dataset exact RCIM model-bank validation workflow.

scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.run_original_dataset_exact_model_bank_validation.resolve_stage_execution_flags(workflow_stage, no_eval, no_export)[source]

Resolve whether evaluation and export stages should run.

Parameters:
  • workflow_stage (str)

  • no_eval (bool)

  • no_export (bool)

Return type:

tuple[bool, bool]

scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.run_original_dataset_exact_model_bank_validation.resolve_best_parameter_summary_payload(workflow_stage, training_config, dataset_bundle, enabled_family_list, workflow_variant, best_parameter_summary_path, best_parameter_registry_path)[source]

Resolve the summary payload used for original-dataset exact LoadBest runs.

Parameters:
  • workflow_stage (str)

  • training_config (dict[str, object])

  • dataset_bundle (ExactPaperDatasetBundle)

  • enabled_family_list (list[str])

  • workflow_variant (str)

  • best_parameter_summary_path (Path | None)

  • best_parameter_registry_path (Path | None)

Return type:

tuple[dict[str, object], str]

scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.run_original_dataset_exact_model_bank_validation.run_original_dataset_exact_model_bank_validation(config_path, output_suffix='original_dataset_exact_validation', workflow_stage='search', best_parameter_summary_path=None, best_parameter_registry_path=None, no_eval=False, no_export=False, grid_search_verbose_override=None, historical_cross_validate_verbose_override=None)[source]

Run the direction-specific original-dataset exact-model workflow.

Parameters:
  • config_path (Path)

  • output_suffix (str)

  • workflow_stage (str)

  • best_parameter_summary_path (Path | None)

  • best_parameter_registry_path (Path | None)

  • no_eval (bool)

  • no_export (bool)

  • grid_search_verbose_override (int | None)

  • historical_cross_validate_verbose_override (int | None)

Return type:

tuple[Path | None, Path | None]

scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.run_original_dataset_exact_model_bank_validation.parse_command_line_arguments()[source]

Parse command-line arguments for the original-dataset exact workflow.

Return type:

Namespace

scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.run_original_dataset_exact_model_bank_validation.main()[source]

Run the original-dataset exact validation entry point.

Return type:

None