get_metrics

hybrid_learning.experimentation.fuzzy_exp.fuzzy_exp_eval.get_metrics(experiment_root, model_key, split, logic_type, formula_dirs=None, skip_missing=True)

Given a sacred experiment_root, collect all experiment settings and results for the respective filters. Filters: model, formula(s), logic type. Experiment settings: formula formulation (formula_dirs), predicate settings, constant values. Metrics: all result metrics saved by the respective formula verification experiment.

Parameters

• skip_missing (bool) – if unset, raise in case no metrics subdirectory can be found for a matching experiment dir

• experiment_root (str) –

• model_key (str) –

• split (str) –

• logic_type (str) –

• formula_dirs (Optional[list]) –

Returns

list of dictionaries, each entry corresponding to a single experiment (filter/settings & results); can directly be used to create a pd.DataFrame

Return type

List[Dict[str, Any]]