Wave 1 Best Model TE Curve Plots

Plot Wave 1 family-best TE predictions on held-out test curves.

class scripts.reports.analysis.plot_wave1_best_model_te_curves.LoadedFamilyModel(family_name, model_type, run_name, run_instance_id, registry_entry, training_config, model_object)[source]

Bases: object

Loaded family-best model plus registry and config metadata.

Parameters:

  • family_name (str)

  • model_type (str)

  • run_name (str)

  • run_instance_id (str)

  • registry_entry (dict[str, Any])

  • training_config (dict[str, Any])

  • model_object (Any)

family_name: str
model_type: str
run_name: str
run_instance_id: str
registry_entry: dict[str, Any]
training_config: dict[str, Any]
model_object: Any
scripts.reports.analysis.plot_wave1_best_model_te_curves.build_argument_parser()[source]

Build the command-line argument parser.

Return type:

ArgumentParser

scripts.reports.analysis.plot_wave1_best_model_te_curves.parse_command_line_arguments()[source]

Parse command-line arguments.

Return type:

Namespace

scripts.reports.analysis.plot_wave1_best_model_te_curves.load_yaml_dictionary(yaml_path)[source]

Load one YAML dictionary from disk.

Parameters:

yaml_path (Path)

Return type:

dict[str, Any]

scripts.reports.analysis.plot_wave1_best_model_te_curves.resolve_timestamped_output_paths(output_root, report_topic_root)[source]

Resolve the timestamped artifact and report directories.

Parameters:

  • output_root (Path)

  • report_topic_root (Path)

Return type:

tuple[str, Path, Path]

scripts.reports.analysis.plot_wave1_best_model_te_curves.load_family_best_entry(family_registry_root, family_name)[source]

Load the current best registry entry for one model family.

Parameters:

  • family_registry_root (Path)

  • family_name (str)

Return type:

dict[str, Any]

scripts.reports.analysis.plot_wave1_best_model_te_curves.load_family_training_config(best_entry)[source]

Load the immutable training config snapshot for one best entry.

Parameters:

best_entry (dict[str, Any])

Return type:

dict[str, Any]

scripts.reports.analysis.plot_wave1_best_model_te_curves.load_neural_regression_module(best_entry, training_config)[source]

Load one family-best Lightning regression checkpoint.

Parameters:

  • best_entry (dict[str, Any])

  • training_config (dict[str, Any])

Return type:

TransmissionErrorRegressionModule

scripts.reports.analysis.plot_wave1_best_model_te_curves.load_wave1_family_best_models(family_name_list, family_registry_root)[source]

Load the current best model artifact for each requested Wave 1 family.

Parameters:

  • family_name_list (list[str])

  • family_registry_root (Path)

Return type:

list[LoadedFamilyModel]

scripts.reports.analysis.plot_wave1_best_model_te_curves.build_canonical_test_dataset(reference_training_config)[source]

Build the canonical test dataset from the approved Wave 1 split config.

Parameters:

reference_training_config (dict[str, Any])

scripts.reports.analysis.plot_wave1_best_model_te_curves.select_curve_indices(test_curve_count, sample_fraction, max_curves, random_seed)[source]

Select a deterministic subset of test curves.

Parameters:

  • test_curve_count (int)

  • sample_fraction (float)

  • max_curves (int | None)

  • random_seed (int)

Return type:

list[int]

scripts.reports.analysis.plot_wave1_best_model_te_curves.predict_curve_with_loaded_model(loaded_model, input_tensor)[source]

Predict one full TE curve with one loaded family-best model.

Parameters:
Return type:

ndarray

scripts.reports.analysis.plot_wave1_best_model_te_curves.compute_curve_metrics(target_curve, predicted_curve)[source]

Compute curve-level TE prediction metrics.

Parameters:

  • target_curve (ndarray)

  • predicted_curve (ndarray)

Return type:

dict[str, float]

scripts.reports.analysis.plot_wave1_best_model_te_curves.summarize_model_metrics(metric_dictionary_list)[source]

Average curve metrics for one model.

Parameters:

metric_dictionary_list (list[dict[str, float]])

Return type:

dict[str, float]

scripts.reports.analysis.plot_wave1_best_model_te_curves.sanitize_filename_fragment(raw_value)[source]

Sanitize one value for use inside generated artifact filenames.

Parameters:

raw_value (str)

Return type:

str

scripts.reports.analysis.plot_wave1_best_model_te_curves.save_prediction_csv(prediction_csv_path, angular_position_deg, target_curve_deg, prediction_dictionary)[source]

Save one per-point prediction table for a TE curve.

Parameters:

  • prediction_csv_path (Path)

  • angular_position_deg (ndarray)

  • target_curve_deg (ndarray)

  • prediction_dictionary (dict[str, ndarray])

Return type:

None

scripts.reports.analysis.plot_wave1_best_model_te_curves.save_prediction_plot(plot_path, angular_position_deg, target_curve_deg, prediction_dictionary, curve_title)[source]

Save one TE-curve overlay plot.

Parameters:

  • plot_path (Path)

  • angular_position_deg (ndarray)

  • target_curve_deg (ndarray)

  • prediction_dictionary (dict[str, ndarray])

  • curve_title (str)

Return type:

None

scripts.reports.analysis.plot_wave1_best_model_te_curves.evaluate_selected_curves(loaded_model_list, test_dataset, selected_curve_index_list, output_directory, write_plots)[source]

Evaluate selected test curves and persist per-curve artifacts.

Parameters:

  • loaded_model_list (list[LoadedFamilyModel])

  • selected_curve_index_list (list[int])

  • output_directory (Path)

  • write_plots (bool)

Return type:

tuple[list[dict[str, Any]], dict[str, dict[str, float]], list[str], list[str]]

scripts.reports.analysis.plot_wave1_best_model_te_curves.save_per_curve_metrics_csv(output_directory, curve_entry_list)[source]

Save one compact per-curve metrics table.

Parameters:

  • output_directory (Path)

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

Return type:

Path

scripts.reports.analysis.plot_wave1_best_model_te_curves.build_validation_summary(run_instance_id, output_directory, report_path, loaded_model_list, dataset_split_summary, selected_curve_index_list, sample_fraction, random_seed, aggregate_metric_dictionary, curve_entry_list, per_curve_metrics_csv_path, plot_path_list)[source]

Build the machine-readable validation summary.

Parameters:

  • run_instance_id (str)

  • output_directory (Path)

  • report_path (Path)

  • loaded_model_list (list[LoadedFamilyModel])

  • selected_curve_index_list (list[int])

  • sample_fraction (float)

  • random_seed (int)

  • aggregate_metric_dictionary (dict[str, dict[str, float]])

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

  • per_curve_metrics_csv_path (Path)

  • plot_path_list (list[str])

Return type:

dict[str, Any]

scripts.reports.analysis.plot_wave1_best_model_te_curves.build_report_markdown(validation_summary)[source]

Build the Markdown comparison report.

Parameters:

validation_summary (dict[str, Any])

Return type:

str

scripts.reports.analysis.plot_wave1_best_model_te_curves.run_wave1_best_model_te_curve_prediction(arguments)[source]

Run the full Wave 1 family-best TE-curve prediction comparison.

Parameters:

arguments (Namespace)

Return type:

dict[str, Any]

scripts.reports.analysis.plot_wave1_best_model_te_curves.main()[source]

Run the CLI entry point.

Return type:

None