plot_overview_for_concept

hybrid_learning.experimentation.ca_exp_eval.plot_overview_for_concept(merged_stats_map, concepts, max_val=None, save_as=None, model_name=None, metric='set_iou', one_plot=True, legend_in=- 1, fig_size=None, axes=None, **plot_args)[source]

Plot and save the output of merge_to_overview using plot_best_ious_wt_std.

Parameters

  • merged_stats_map (Dict[str, DataFrame]) – dict with values the output of merge_to_overview; each value will get its own graph

  • concepts (Union[str, Iterable[str]]) – the concept or concepts to plot

  • max_val (Optional[float]) – the common x-axis limit for all created plots; defaults to the maximum metric value for a concept

  • save_as (Optional[str]) – an optional save location; must contain {concept} if one_plot is not True; if a directory is given, the file name is auto-inferred from the other arguments

  • model_name (Optional[str]) – for file name auto-inference

  • metric (str) – for axis label and file name auto-inference

  • one_plot (bool) – whether to put all into one common plot (save_as must not contain formatting string for concept)

  • legend_in (int) – if one_plot, index of the axis the legend should be placed in; if set to None, legend is placed outside of plot

  • axes (Optional[Sequence[Axes]]) – plot into given axes

  • fig_size (Optional[Tuple[float, float]]) – if one_plot, used fig_size