Original Dataset Exact Model Bank Support

This page documents the support utilities for the faithful RCIM Model-Bank Reproduction original-dataset exact-model-bank reimplementation. This is the bidirectional campaign surface that promotes accepted models into models/paper_reference/rcim_track1 and updates the RCIM Tables 2-5 benchmark.

The full-dataset RCIM Model-Bank Reproduction surface is closed after completed forward and backward paper-faithful grid-search campaigns. Future restricted-dataset reruns should use separate campaign namespaces and comparison reports.

Support utilities for the original-dataset exact RCIM model-bank branch.

class scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.original_dataset_exact_model_bank_support.OriginalDatasetExactModelBankBundle(exact_dataset_bundle, validation_feature_matrix, validation_target_matrix, dataset_root, dataset_config_path, direction_label, direction_prefix, selected_harmonic_list, decomposition_point_stride, split_row_count_dictionary, split_file_count_dictionary, smoke_settings)[source]

Bases: object

Prepared original-dataset bundle for one directional exact-model run.

Parameters:

  • exact_dataset_bundle (ExactPaperDatasetBundle)

  • validation_feature_matrix (DataFrame)

  • validation_target_matrix (DataFrame)

  • dataset_root (Path)

  • dataset_config_path (Path)

  • direction_label (str)

  • direction_prefix (str)

  • selected_harmonic_list (list[int])

  • decomposition_point_stride (int)

  • split_row_count_dictionary (dict[str, int])

  • split_file_count_dictionary (dict[str, int])

  • smoke_settings (dict[str, Any])

exact_dataset_bundle: ExactPaperDatasetBundle
validation_feature_matrix: DataFrame
validation_target_matrix: DataFrame
dataset_root: Path
dataset_config_path: Path
direction_label: str
direction_prefix: str
selected_harmonic_list: list[int]
decomposition_point_stride: int
split_row_count_dictionary: dict[str, int]
split_file_count_dictionary: dict[str, int]
smoke_settings: dict[str, Any]
scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.original_dataset_exact_model_bank_support.load_original_dataset_exact_model_bank_config(config_path)[source]

Load one original-dataset exact-model-bank configuration.

Parameters:

config_path (str | Path)

Return type:

dict[str, Any]

scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.original_dataset_exact_model_bank_support.resolve_original_dataset_direction_settings(training_config)[source]

Resolve the configured playback direction and its target-name prefix.

Parameters:

training_config (dict[str, Any])

Return type:

tuple[str, str]

scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.original_dataset_exact_model_bank_support.resolve_original_dataset_smoke_settings(training_config)[source]

Resolve optional smoke-validation limits for the original-dataset branch.

Parameters:

training_config (dict[str, Any] | None)

Return type:

dict[str, Any]

scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.original_dataset_exact_model_bank_support.build_original_dataset_exact_model_bank_bundle(training_config)[source]

Build the original-dataset exact-model-bank split bundle.

Parameters:

training_config (dict[str, Any])

Return type:

OriginalDatasetExactModelBankBundle

scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.original_dataset_exact_model_bank_support.build_original_dataset_validation_summary(resolved_config_path, output_directory, training_config, original_dataset_bundle, family_summary_list, family_search_summary_dictionary, per_target_ranking_dictionary, onnx_export_summary, model_bundle_path)[source]

Build the canonical validation summary for the original-dataset branch.

Parameters:

  • resolved_config_path (Path)

  • output_directory (Path)

  • training_config (dict[str, Any])

  • original_dataset_bundle (OriginalDatasetExactModelBankBundle)

  • family_summary_list (list[dict[str, Any]])

  • family_search_summary_dictionary (dict[str, dict[str, Any]])

  • per_target_ranking_dictionary (dict[str, list[dict[str, Any]]])

  • onnx_export_summary (dict[str, Any])

  • model_bundle_path (Path)

Return type:

dict[str, Any]

scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.original_dataset_exact_model_bank_support.build_original_dataset_validation_report_path(training_config)[source]

Build the Markdown report path for the original-dataset branch.

Parameters:

training_config (dict[str, Any])

Return type:

Path

scripts.paper_reimplementation.rcim_ml_compensation.original_dataset_exact_model_bank.original_dataset_exact_model_bank_support.build_original_dataset_validation_report_markdown(validation_summary)[source]

Build the human-readable Markdown report for the original-dataset branch.

Parameters:

validation_summary (dict[str, Any])

Return type:

str