Sandbox

systematica.tuners.sandbox

Deprecated: No need to test

OptunaSandbox

OptunaSandbox()

Method generated by attrs for class OptunaSandbox.

Static methods

get_feature_output

get_feature_output(
    study: optuna.study.study.Study,
) ‑> systematica.tuners.sandbox.FeatureOutput

Get feature output from the optuna study.

For multiobjective, it selects a single solution from the Pareto front, using the scalarization method.

Parameters:

Name	Type	Default	Description
study	optuna.Study	--	Sutdy run through Sandbox.

Returns:

Type	Description
FeatureOutput	Results.

Methods

run

run(
    self,
    search: str = None,
) ‑> Dict[str, Dict[optuna.study.study.Study, Tuple[float, ...]]]

Calculate all registered features.

Parameters:

Name	Type	Default	Description
search	str	--	Filter config by timeframe or study_name. Key should match timeframe. Default to None.

Returns:

Type	Description
tp.Dict[str, tp.NamedTuple]	Dictionary mapping feature names to computed features.

get_features

get_features(
    self,
    search: str | List[str],
) ‑> List[~FeatureConfig]

Filter filter_config with specific key-word or list of key-words.

This method searches within the id, study_name, and timeframe attributes of each FeatureConfig in the config. The search is performed in a case-sensitive manner on id while study_name and timeframe are compared case-insensitively.

It supports standard substring matching as well as prefix matching using a trailing ’*’ in the search term.

Parameters:

Name	Type	Default	Description
search	str	--	or list of str The search term or list of search terms. Each term is compared against the attributes id, study_name, and timeframe of FeatureConfig. For example, a term like “my-model*” will match any entry where the attribute (with the aforementioned case rules) starts with “my-model”.

Returns:

Type	Description
tp.List[FeatureConfigT]	A list of FeatureConfig objects whose attributes match any of the search terms.