gunz_ml.plots package

Submodules

gunz_ml.plots.categorical_distributions module

Plotting function for visualizing distributions of metrics grouped by categorical hyperparameters.

gunz_ml.plots.categorical_distributions.plot_categorical_distributions(metric_df: DataFrame, param_df: DataFrame, objective_metrics: List[str], output_path: Path, categorical_params_list: list, plot_cfg: omegaconf.DictConfig | Dict) → None

Generates and saves distribution plots for categorical hyperparameters.

gunz_ml.plots.contour module

Plotting function for visualizing 2D contour plots of hyperparameters vs. an objective.

gunz_ml.plots.contour.plot_contour(metric_df: DataFrame, param_df: DataFrame, objective_metrics: List[str], output_path: Path, continuous_params_map: Dict[str, str], plot_cfg: omegaconf.DictConfig | Dict) → None

Generates contour plots of 2 continuous parameters vs. 1 objective.

gunz_ml.plots.enhanced_scatter module

Plotting function for visualizing a continuous hyperparameter against two objective metrics.

gunz_ml.plots.enhanced_scatter.plot_enhanced_scatter(metric_df: DataFrame, param_df: DataFrame, objective_metrics: List[str], output_path: Path, continuous_params_map: Dict[str, str], plot_cfg: omegaconf.DictConfig | Dict) → None

Generates scatter plots of 1 continuous parameter vs. 2 objectives.

gunz_ml.plots.optuna module

gunz_ml.plots.optuna.apply_plot_filters(metric_df: DataFrame, param_df: DataFrame, filter_cfg: List[Dict[str, Any]]) → None

Apply a list of metric-based filters to metric_df and param_df in-place.

gunz_ml.plots.optuna.find_pareto_frontier_fast(df: DataFrame, x_col: str, y_col: str, x_dir: str, y_dir: str) → DataFrame

Find Pareto frontier points efficiently (O(n log n)) using sorting.

gunz_ml.plots.optuna.generate_plots(study: optuna.study.Study, plot_cfg: omegaconf.DictConfig, exp_params: omegaconf.DictConfig, output_path: str | Path)

Main function to generate all Optuna plots. This function encapsulates the logic from the PLOT mode of the run function.

gunz_ml.plots.optuna.plot_categorical_distributions(metric_df: DataFrame, param_df: DataFrame, objective_metrics: List[str], output_path: str | Path, categorical_params_list: list, plot_cfg: omegaconf.DictConfig | Dict) → None

Generates and saves distribution plots for categorical hyperparameters.

gunz_ml.plots.optuna.plot_contour(metric_df: DataFrame, param_df: DataFrame, objective_metrics: List[str], output_path: str | Path, continuous_params_map: Dict[str, str], plot_cfg: omegaconf.DictConfig | Dict) → None

Generates contour plots of 2 continuous parameters vs. 1 objective.

gunz_ml.plots.optuna.plot_enhanced_scatter(metric_df: DataFrame, param_df: DataFrame, objective_metrics: List[str], output_path: str | Path, continuous_params_map: Dict[str, str], plot_cfg: omegaconf.DictConfig | Dict) → None

Generates scatter plots of 1 continuous parameter vs. 2 objectives.

gunz_ml.plots.optuna.plot_pareto_continuous_color(metric_df: DataFrame, param_df: DataFrame, objective_metrics: List[str], study_directions: List[optuna.study.StudyDirection], output_path: str | Path, continuous_params_map: Dict[str, str], plot_cfg: omegaconf.DictConfig | Dict) → None

Generates Pareto fronts of 2 objectives, colored by a continuous parameter.

gunz_ml.plots.optuna.plot_pareto_frontiers_by_category(metric_df: DataFrame, param_df: DataFrame, objective_metrics: List[str], study_directions: List[optuna.study.StudyDirection], output_path: str | Path, categorical_params_list: list, plot_cfg: omegaconf.DictConfig | Dict) → None

Generates Pareto frontier plots for each pair of objectives, grouped by each categorical hyperparameter.

gunz_ml.plots.pareto_by_category module

Plotting function for Pareto frontiers grouped by a categorical hyperparameter.

gunz_ml.plots.pareto_by_category.plot_pareto_frontiers_by_category(metric_df: DataFrame, param_df: DataFrame, objective_metrics: List[str], study_directions: List[optuna.study.StudyDirection], output_path: Path, categorical_params_list: list, plot_cfg: omegaconf.DictConfig | Dict) → None

Generates Pareto plots for objectives, grouped by a categorical parameter.

gunz_ml.plots.pareto_continuous_color module

Plotting function for Pareto frontiers colored by a continuous hyperparameter.

gunz_ml.plots.pareto_continuous_color.plot_pareto_continuous_color(metric_df: DataFrame, param_df: DataFrame, objective_metrics: List[str], study_directions: List[optuna.study.StudyDirection], output_path: Path, continuous_params_map: Dict[str, str], plot_cfg: omegaconf.DictConfig | Dict) → None

Generates Pareto fronts of 2 objectives, colored by a continuous parameter.

gunz_ml.plots.utils module

Shared utilities for generating Optuna analysis plots.