mlos_bench.storage.base_tunable_config_data module

Base interface for accessing the stored benchmark (tunable) config data.

class mlos_bench.storage.base_tunable_config_data.TunableConfigData(*, tunable_config_id: int)

Bases: object

Base interface for accessing the stored experiment benchmark (tunable) config data.

A configuration in this context is the set of tunable parameter values.

Attributes:

config_df

Retrieve the trials’ tunable configuration from the storage.

config_dict

Retrieve the trials’ tunable configuration from the storage as a dict.

tunable_config_id

Unique ID of the (tunable) configuration.

abstract property config_df: DataFrame

Retrieve the trials’ tunable configuration from the storage.

Note: this corresponds to the Trial object’s “tunables” property.

Returns:

configpandas.DataFrame

A dataframe with the tunable configuration of the trial. It has two str columns, “parameter” and “value”.

property config_dict: Dict[str, int | float | str | None]

Retrieve the trials’ tunable configuration from the storage as a dict.

Note: this corresponds to the Trial object’s “tunables” property.

Returns:

configdict

property tunable_config_id: int

Unique ID of the (tunable) configuration.