nat.profiler.parameter_optimization.pareto_visualizer

Attributes

logger

Classes

ParetoVisualizer

Functions

load_trials_from_study(→ tuple[pandas.DataFrame, ...)
load_trials_from_csv(→ tuple[pandas.DataFrame, ...)
compute_pareto_optimal_mask(→ numpy.ndarray)
create_pareto_visualization(→ dict[str, ...)

Module Contents

logger

class ParetoVisualizer(

metric_names: list[str],

directions: list[str],

title_prefix: str = 'Optimization Results',

)

metric_names

directions

title_prefix = 'Optimization Results'

plot_pareto_front_2d(

trials_df: pandas.DataFrame,

pareto_trials_df: pandas.DataFrame | None = None,

save_path: pathlib.Path | None = None,

figsize: tuple[int, int] = (10, 8),

show_plot: bool = True,

) → matplotlib.pyplot.Figure

plot_pareto_parallel_coordinates(

trials_df: pandas.DataFrame,

pareto_trials_df: pandas.DataFrame | None = None,

save_path: pathlib.Path | None = None,

figsize: tuple[int, int] = (12, 8),

show_plot: bool = True,

) → matplotlib.pyplot.Figure

plot_pairwise_matrix(

trials_df: pandas.DataFrame,

pareto_trials_df: pandas.DataFrame | None = None,

save_path: pathlib.Path | None = None,

figsize: tuple[int, int] | None = None,

show_plot: bool = True,

) → matplotlib.pyplot.Figure

load_trials_from_study(

study: optuna.Study,

) → tuple[pandas.DataFrame, pandas.DataFrame]

load_trials_from_csv(

csv_path: pathlib.Path,

metric_names: list[str],

directions: list[str],

) → tuple[pandas.DataFrame, pandas.DataFrame]

compute_pareto_optimal_mask(

df: pandas.DataFrame,

value_cols: list[str],

directions: list[str],

) → numpy.ndarray

create_pareto_visualization(

data_source: optuna.Study | pathlib.Path | pandas.DataFrame,

metric_names: list[str],

directions: list[str],

output_dir: pathlib.Path | None = None,

title_prefix: str = 'Optimization Results',

show_plots: bool = True,

) → dict[str, matplotlib.pyplot.Figure]