OCDocker.OCScore.Analysis package

Subpackages

• OCDocker.OCScore.Analysis.Metrics package
  • Submodules
    • OCDocker.OCScore.Analysis.Metrics.Bootstrap module
      • bootstrap_ci()
    • OCDocker.OCScore.Analysis.Metrics.Calibration module
      • LOGIT_CLIP
      • logits_to_probabilities()
      • clip_probabilities()
      • expected_calibration_error()
      • evaluate_calibration_metrics()
      • reliability_curve_points()
      • ProbabilityCalibrator
      • is_calibration_metric_key()
      • apply_calibration_report_mode_to_metrics()
      • collect_calibration_report_issues()
      • validate_calibration_report_mode()
      • extract_calibration_metrics()
      • build_calibration_report_section()
      • merge_calibration_metrics()
      • enrich_dudez_export_metrics()
      • calibration_metric_names()
    • OCDocker.OCScore.Analysis.Metrics.Ranking module
      • DEFAULT_SCREENING_COMPARISON_METRICS
      • orient_scores()
      • is_valid_ranking_scores()
      • score_ranking_diagnostics()
      • bedroc()
      • rie()
      • aggregate_group_metric()
      • evaluate_screening_metrics_by_group()
      • evaluate_scoring_functions_by_group()
      • enrichment_factor()
      • ndcg_at_fraction()
      • groupwise()
      • pr_auc()
      • riep()
      • roc_auc()
      • threshold_at_precision()
      • top_fraction_precision()
      • top_k_precision()
      • binary_threshold_youden()
      • confusion_counts()
      • classification_metrics_at_threshold()
      • aggregate_group_classification_metrics()
      • evaluate_screening_metrics()
  • Module contents
    • Modules
• OCDocker.OCScore.Analysis.Plotting package
  • Modules
  • plot_combined_metric_scatter()
  • plot_boxplots()
  • plot_barplots()
  • plot_scatterplot()
  • plot_bar_with_significance()
  • plot_heatmap()
  • plot_normality_and_variance_diagnostics()
  • plot_pca_importance_barplot()
  • plot_pca_importance_histogram()
  • save_pca_importance_groups()
  • save_pca_importance_bins()
  • set_color_mapping()
  • Submodules
    • OCDocker.OCScore.Analysis.Plotting.Colouring module
      • LOGGER
      • set_color_mapping()
    • OCDocker.OCScore.Analysis.Plotting.ArchitecturePlots module
      • ArchitectureBlock
      • ArchitectureDiagram
      • load_architecture_document()
      • normalize_architecture()
      • plot_architecture_diagram()
      • save_architecture_figures()
    • OCDocker.OCScore.Analysis.Plotting.Core module
      • apply_basic_style()
      • new_fig()
    • OCDocker.OCScore.Analysis.Plotting.CrossValidationPlots module
      • aggregate_cv_per_target_metrics()
      • load_cross_validation_artifacts()
      • plot_fold_metric_heatmap()
      • plot_fold_metric_bars()
      • plot_fold_metric_lines()
      • plot_mean_std_bars()
      • plot_ocscore_wins()
      • plot_per_target_boxplot()
      • plot_per_target_heatmap()
      • plot_per_target_ocscore_wins()
      • resolve_cross_validation_dir()
      • save_baseline_comparison_figures()
      • save_calibration_reliability_figures()
      • save_cross_validation_figures()
      • save_per_target_figures()
    • OCDocker.OCScore.Analysis.Plotting.MetricsPlots module
      • enrichment_plot()
      • pr_plot()
      • roc_plot()
      • reliability_plot()
    • OCDocker.OCScore.Analysis.Plotting.Stats module
      • plot_ablation_bedroc_significance_bars()
      • classify_policies_by_shortcut_rule()
      • classify_policies_by_eligibility_and_shortcut_risk()
      • plot_bedroc_vs_shortcut_risk_scatter()
      • plot_bar_with_significance()
      • plot_barplots()
      • plot_boxplots()
      • plot_combined_metric_scatter()
      • plot_heatmap()
      • plot_normality_and_variance_diagnostics()
      • plot_pca_importance_barplot()
      • plot_pca_importance_histogram()
      • plot_scatterplot()
      • save_pca_importance_bins()
      • save_pca_importance_groups()
• OCDocker.OCScore.Analysis.SHAP package
  • Submodules
    • OCDocker.OCScore.Analysis.SHAP.ExportRunner module
      • run_export_shap_analysis()
    • OCDocker.OCScore.Analysis.SHAP.Explain module
      • compute_shap_values()
    • OCDocker.OCScore.Analysis.SHAP.Plots module
      • assign_feature_families()
      • beeswarm()
      • compute_feature_importance_table()
      • compute_family_importance_table()
      • compute_label_family_distribution_table()
      • compute_target_family_shap_table()
      • feature_importance_barh()
      • load_family_spec()
      • save_beeswarm_plot()
      • save_dependence_plots()
      • save_family_importance_plot()
      • save_global_feature_importance_plot()
      • save_label_family_distribution_plot()
      • save_shap_plot_suite()
      • save_shap_plot_suite_from_paths()
      • save_target_family_heatmap()
      • shap_correlation_heatmap()
    • OCDocker.OCScore.Analysis.SHAP.Paths module
      • OutputPaths
  • Module contents

Submodules

• OCDocker.OCScore.Analysis.ProductionBaselines module
  • PRODUCTION_BASELINE_RANK_METRICS
  • ProductionBaselineConfig
  • TrainOnlyFitError
  • aggregate_baseline_rows()
  • build_baseline_rank_table()
  • evaluate_individual_sf_baselines()
  • evaluate_sf_consensus_baselines()
  • evaluate_descriptor_aggregate_baselines()
  • evaluate_learned_sf_baselines()
  • run_and_write_production_baselines()
  • run_production_baselines_for_replica()
  • validate_fit_uses_train_only()
  • write_production_baseline_reports()

Module contents

Unified exports for the OCScore Analysis package.

Usage:

import OCDocker.OCScore.Analysis as ocanalysis

Modules (current / staged pipeline)

• Metrics: Screening and regression metric helpers.

• Plotting: Cross-validation, baseline, and metrics plots.

• SHAP: Pipeline-native SHAP for exported model bundles.

OCDocker.OCScore.Analysis.evaluate_screening_metrics(y_true, y_score, groups=None, higher_is_better=True, bedroc_alpha=20.0)

Evaluate DUDEz classification and early-recognition metrics.

Classifier logits and probabilities should use higher_is_better=True. Lower-is-better docking scores must set higher_is_better=False.

When groups is provided, BEDROC, EF, and NDCG are averaged across targets/receptors with both actives and decoys present.

Parameters:

• bedroc_alpha (float, optional) – Exponential BEDROC weighting factor, by default 20.0.

• y_true (ndarray)

• y_score (ndarray)

• groups (ndarray | None)

• higher_is_better (bool)

Return type:

dict[str, float]

Staged-pipeline analysis surface (metrics, plotting, export SHAP).