Tasks

The task module implements a hierarchy of explanation tasks that can be used to explain HPO.

The tasks provide a convenient API to construct surrogate models from different data sources (pretrained estimators, empirical data, or a black box function) and to add domain specific information such as a baseline configuration or an optimizer of interest.

AblationExplanationTask

Bases: BaselineExplanationTask

Defines an ablation explanation task, comparing a configuration of interest to a baseline.

Source code in src/hypershap/task.py

class AblationExplanationTask(BaselineExplanationTask):
    """Defines an ablation explanation task, comparing a configuration of interest to a baseline."""

    def __init__(
        self,
        config_space: ConfigurationSpace,
        surrogate_model: SurrogateModel | list[SurrogateModel],
        baseline_config: Configuration,
        config_of_interest: Configuration,
    ) -> None:
        """Initialize an AblationExplanationTask with a baseline and a configuration of interest.

        Args:
            config_space: The configuration space.
            surrogate_model: The (list of) surrogate model(s) used for the explanation task.
            baseline_config: The baseline configuration.
            config_of_interest: The configuration of interest.

        """
        super().__init__(config_space, surrogate_model, baseline_config)
        self.config_of_interest = config_of_interest

__init__(config_space, surrogate_model, baseline_config, config_of_interest)

Initialize an AblationExplanationTask with a baseline and a configuration of interest.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
surrogate_model	SurrogateModel | list[SurrogateModel]	The (list of) surrogate model(s) used for the explanation task.	required
baseline_config	Configuration	The baseline configuration.	required
config_of_interest	Configuration	The configuration of interest.	required

Source code in src/hypershap/task.py

def __init__(
    self,
    config_space: ConfigurationSpace,
    surrogate_model: SurrogateModel | list[SurrogateModel],
    baseline_config: Configuration,
    config_of_interest: Configuration,
) -> None:
    """Initialize an AblationExplanationTask with a baseline and a configuration of interest.

    Args:
        config_space: The configuration space.
        surrogate_model: The (list of) surrogate model(s) used for the explanation task.
        baseline_config: The baseline configuration.
        config_of_interest: The configuration of interest.

    """
    super().__init__(config_space, surrogate_model, baseline_config)
    self.config_of_interest = config_of_interest

BaselineExplanationTask

Bases: ExplanationTask

Defines an explanation task with a baseline configuration.

Source code in src/hypershap/task.py

class BaselineExplanationTask(ExplanationTask):
    """Defines an explanation task with a baseline configuration."""

    def __init__(
        self,
        config_space: ConfigurationSpace,
        surrogate_model: SurrogateModel | list[SurrogateModel],
        baseline_config: Configuration,
    ) -> None:
        """Initialize a BaselineExplanationTask with a baseline configuration.

        Args:
            config_space: The configuration space.
            surrogate_model: The (list of) surrogate model(s) used for the explanation task.
            baseline_config: The baseline configuration.

        """
        super().__init__(config_space, surrogate_model)
        self.baseline_config = baseline_config

__init__(config_space, surrogate_model, baseline_config)

Initialize a BaselineExplanationTask with a baseline configuration.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
surrogate_model	SurrogateModel | list[SurrogateModel]	The (list of) surrogate model(s) used for the explanation task.	required
baseline_config	Configuration	The baseline configuration.	required

Source code in src/hypershap/task.py

def __init__(
    self,
    config_space: ConfigurationSpace,
    surrogate_model: SurrogateModel | list[SurrogateModel],
    baseline_config: Configuration,
) -> None:
    """Initialize a BaselineExplanationTask with a baseline configuration.

    Args:
        config_space: The configuration space.
        surrogate_model: The (list of) surrogate model(s) used for the explanation task.
        baseline_config: The baseline configuration.

    """
    super().__init__(config_space, surrogate_model)
    self.baseline_config = baseline_config

ExplanationTask

Defines the base class for explanation tasks, providing access to the configuration space and surrogate model.

Source code in src/hypershap/task.py

class ExplanationTask:
    """Defines the base class for explanation tasks, providing access to the configuration space and surrogate model."""

    def __init__(
        self,
        config_space: ConfigurationSpace,
        surrogate_model: SurrogateModel | list[SurrogateModel],
    ) -> None:
        """Initialize an ExplanationTask with a configuration space and surrogate model.

        Args:
            config_space: The configuration space for the explanation task.
            surrogate_model: The (list of) surrogate model(s) used for the explanation task.

        """
        self.config_space = config_space
        self.surrogate_model = surrogate_model

    def is_multi_data(self) -> bool:
        """Return if the explanation task is a multi-data task.

        Returns:
            True if the explanation task is a multi-data task.

        """
        return isinstance(self.surrogate_model, list)

    def get_single_surrogate_model(self) -> SurrogateModel:
        """Return the surrogate model for the explanation task.

        Returns:
            The surrogate model for the explanation task.

        """
        if isinstance(self.surrogate_model, list):
            raise TypeError

        return self.surrogate_model

    def get_surrogate_model_list(self) -> list[SurrogateModel]:
        """Return the list of surrogate models for the explanation task.

        Returns:
            The list of surrogate models for the explanation task.

        """
        if not isinstance(self.surrogate_model, list):
            raise TypeError

        return self.surrogate_model

    def get_num_hyperparameters(self) -> int:
        """Return the number of hyperparameters in the configuration space.

        Returns:
            The number of hyperparameters.

        """
        return len(self.config_space)

    def get_hyperparameter_names(self) -> list[str]:
        """Return the names of the hyperparameters in the configuration space.

        Returns:
            A list of hyperparameter names.

        """
        return list(self.config_space.keys())

    @staticmethod
    def from_base_model(config_space: ConfigurationSpace, base_model: BaseEstimator) -> ExplanationTask:
        """Create an ExplanationTask from a pre-trained base model.

        Args:
            config_space: The configuration space.
            base_model: The pre-trained base model.

        Returns:
            An ExplanationTask instance.

        """
        surrogate_model = ModelBasedSurrogateModel(config_space=config_space, base_model=base_model)
        return ExplanationTask(config_space=config_space, surrogate_model=surrogate_model)

    @staticmethod
    def from_data(
        config_space: ConfigurationSpace,
        data: list[tuple[Configuration, float]],
        base_model: BaseEstimator | None = None,
    ) -> ExplanationTask:
        """Create an ExplanationTask from a dataset of configurations and their performance.

        Args:
            config_space: The configuration space.
            data: A list of tuples, where each tuple contains a configuration and its corresponding performance.
            base_model: The base model to use for training the surrogate model. Defaults to RandomForestRegressor.

        Returns:
            An ExplanationTask instance.

        """
        surrogate_model = DataBasedSurrogateModel(config_space=config_space, data=data, base_model=base_model)
        return ExplanationTask(config_space=config_space, surrogate_model=surrogate_model)

    @staticmethod
    def from_function(
        config_space: ConfigurationSpace,
        function: Callable[[Configuration], float],
        n_samples: int = 1_000,
        base_model: BaseEstimator | None = None,
        seed: int | None = 0,
    ) -> ExplanationTask:
        """Create an ExplanationTask from a function that evaluates configurations.

        Args:
            config_space: The configuration space.
            function: A callable that takes a configuration and returns its performance.
            n_samples: The number of configurations to sample for training the surrogate model. Defaults to 1000.
            base_model: The base model to use for training the surrogate model. Defaults to RandomForestRegressor.
            seed: The seed for the random number generator, it is used to seed a deep copy of the config space.

        Returns:
            An ExplanationTask instance.

        """
        cs = deepcopy(config_space)
        if seed is not None:
            cs.seed(seed)
        samples: list[Configuration] = cs.sample_configuration(n_samples)
        values: list[float] = [function(config) for config in samples]
        data: list[tuple[Configuration, float]] = list(zip(samples, values, strict=False))
        base_model = base_model if base_model is not None else RandomForestRegressor(random_state=seed)

        return ExplanationTask.from_data(config_space=cs, data=data, base_model=base_model)

    @staticmethod
    def from_function_multidata(
        config_space: ConfigurationSpace,
        functions: list[Callable[[Configuration], float]],
        n_samples: int = 1_000,
        base_model: BaseEstimator | None = None,
    ) -> ExplanationTask:
        """Create an ExplanationTask from a list of functions that evaluate configurations.

        Args:
            config_space: The configuration space.
            functions: A list of callables that take a configuration and returns its performance.
            n_samples: The number of configurations to sample for training the surrogate model. Defaults to 1000.
            base_model: The base model to be used for training the surrogate model. Defaults to RandomForestRegressor.

        Returns:
            An ExplanationTask instance.

        """
        surrogate_model_list = [
            ExplanationTask.from_function(config_space, fun, n_samples, base_model).get_single_surrogate_model()
            for fun in functions
        ]
        return ExplanationTask(config_space=config_space, surrogate_model=surrogate_model_list)

    @staticmethod
    def from_data_multidata(
        config_space: ConfigurationSpace,
        data_multidata: list[list[tuple[Configuration, float]]],
        base_model: BaseEstimator | None = None,
    ) -> ExplanationTask:
        """Create an ExplanationTask from a list of datasets of different HPO tasks.

        Args:
            config_space: The configuration space.
            data_multidata: A list of tuples, where each tuple contains a configuration and its corresponding performance.
            base_model: The base model to use for training the surrogate model. Defaults to RandomForestRegressor.

        Returns:
            An ExplanationTask instance.

        """
        surrogate_model_list = [
            ExplanationTask.from_data(config_space, data, base_model).get_single_surrogate_model()
            for data in data_multidata
        ]
        return ExplanationTask(config_space=config_space, surrogate_model=surrogate_model_list)

    @staticmethod
    def from_basemodel_multidata(
        config_space: ConfigurationSpace,
        base_model: list[BaseEstimator],
    ) -> ExplanationTask:
        """Create an ExplanationTask from a list of datasets of different HPO tasks.

        Args:
            config_space: The configuration space.
            base_model: The list of base models to be used as surrogate models.

        Returns:
            An ExplanationTask instance.

        """
        surrogate_model_list: list[SurrogateModel] = [
            ModelBasedSurrogateModel(config_space=config_space, base_model=m) for m in base_model
        ]
        return ExplanationTask(config_space=config_space, surrogate_model=surrogate_model_list)

__init__(config_space, surrogate_model)

Initialize an ExplanationTask with a configuration space and surrogate model.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space for the explanation task.	required
surrogate_model	SurrogateModel | list[SurrogateModel]	The (list of) surrogate model(s) used for the explanation task.	required

Source code in src/hypershap/task.py

def __init__(
    self,
    config_space: ConfigurationSpace,
    surrogate_model: SurrogateModel | list[SurrogateModel],
) -> None:
    """Initialize an ExplanationTask with a configuration space and surrogate model.

    Args:
        config_space: The configuration space for the explanation task.
        surrogate_model: The (list of) surrogate model(s) used for the explanation task.

    """
    self.config_space = config_space
    self.surrogate_model = surrogate_model

from_base_model(config_space, base_model) staticmethod

Create an ExplanationTask from a pre-trained base model.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
base_model	BaseEstimator	The pre-trained base model.	required

Returns:

Type	Description
ExplanationTask	An ExplanationTask instance.

Source code in src/hypershap/task.py

@staticmethod
def from_base_model(config_space: ConfigurationSpace, base_model: BaseEstimator) -> ExplanationTask:
    """Create an ExplanationTask from a pre-trained base model.

    Args:
        config_space: The configuration space.
        base_model: The pre-trained base model.

    Returns:
        An ExplanationTask instance.

    """
    surrogate_model = ModelBasedSurrogateModel(config_space=config_space, base_model=base_model)
    return ExplanationTask(config_space=config_space, surrogate_model=surrogate_model)

from_basemodel_multidata(config_space, base_model) staticmethod

Create an ExplanationTask from a list of datasets of different HPO tasks.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
base_model	list[BaseEstimator]	The list of base models to be used as surrogate models.	required

Returns:

Type	Description
ExplanationTask	An ExplanationTask instance.

Source code in src/hypershap/task.py

@staticmethod
def from_basemodel_multidata(
    config_space: ConfigurationSpace,
    base_model: list[BaseEstimator],
) -> ExplanationTask:
    """Create an ExplanationTask from a list of datasets of different HPO tasks.

    Args:
        config_space: The configuration space.
        base_model: The list of base models to be used as surrogate models.

    Returns:
        An ExplanationTask instance.

    """
    surrogate_model_list: list[SurrogateModel] = [
        ModelBasedSurrogateModel(config_space=config_space, base_model=m) for m in base_model
    ]
    return ExplanationTask(config_space=config_space, surrogate_model=surrogate_model_list)

from_data(config_space, data, base_model=None) staticmethod

Create an ExplanationTask from a dataset of configurations and their performance.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
data	list[tuple[Configuration, float]]	A list of tuples, where each tuple contains a configuration and its corresponding performance.	required
base_model	BaseEstimator | None	The base model to use for training the surrogate model. Defaults to RandomForestRegressor.	None

Returns:

Type	Description
ExplanationTask	An ExplanationTask instance.

Source code in src/hypershap/task.py

@staticmethod
def from_data(
    config_space: ConfigurationSpace,
    data: list[tuple[Configuration, float]],
    base_model: BaseEstimator | None = None,
) -> ExplanationTask:
    """Create an ExplanationTask from a dataset of configurations and their performance.

    Args:
        config_space: The configuration space.
        data: A list of tuples, where each tuple contains a configuration and its corresponding performance.
        base_model: The base model to use for training the surrogate model. Defaults to RandomForestRegressor.

    Returns:
        An ExplanationTask instance.

    """
    surrogate_model = DataBasedSurrogateModel(config_space=config_space, data=data, base_model=base_model)
    return ExplanationTask(config_space=config_space, surrogate_model=surrogate_model)

from_data_multidata(config_space, data_multidata, base_model=None) staticmethod

Create an ExplanationTask from a list of datasets of different HPO tasks.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
data_multidata	list[list[tuple[Configuration, float]]]	A list of tuples, where each tuple contains a configuration and its corresponding performance.	required
base_model	BaseEstimator | None	The base model to use for training the surrogate model. Defaults to RandomForestRegressor.	None

Returns:

Type	Description
ExplanationTask	An ExplanationTask instance.

Source code in src/hypershap/task.py

@staticmethod
def from_data_multidata(
    config_space: ConfigurationSpace,
    data_multidata: list[list[tuple[Configuration, float]]],
    base_model: BaseEstimator | None = None,
) -> ExplanationTask:
    """Create an ExplanationTask from a list of datasets of different HPO tasks.

    Args:
        config_space: The configuration space.
        data_multidata: A list of tuples, where each tuple contains a configuration and its corresponding performance.
        base_model: The base model to use for training the surrogate model. Defaults to RandomForestRegressor.

    Returns:
        An ExplanationTask instance.

    """
    surrogate_model_list = [
        ExplanationTask.from_data(config_space, data, base_model).get_single_surrogate_model()
        for data in data_multidata
    ]
    return ExplanationTask(config_space=config_space, surrogate_model=surrogate_model_list)

from_function(config_space, function, n_samples=1000, base_model=None, seed=0) staticmethod

Create an ExplanationTask from a function that evaluates configurations.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
function	Callable[[Configuration], float]	A callable that takes a configuration and returns its performance.	required
n_samples	int	The number of configurations to sample for training the surrogate model. Defaults to 1000.	1000
base_model	BaseEstimator | None	The base model to use for training the surrogate model. Defaults to RandomForestRegressor.	None
seed	int | None	The seed for the random number generator, it is used to seed a deep copy of the config space.	0

Returns:

Type	Description
ExplanationTask	An ExplanationTask instance.

Source code in src/hypershap/task.py

@staticmethod
def from_function(
    config_space: ConfigurationSpace,
    function: Callable[[Configuration], float],
    n_samples: int = 1_000,
    base_model: BaseEstimator | None = None,
    seed: int | None = 0,
) -> ExplanationTask:
    """Create an ExplanationTask from a function that evaluates configurations.

    Args:
        config_space: The configuration space.
        function: A callable that takes a configuration and returns its performance.
        n_samples: The number of configurations to sample for training the surrogate model. Defaults to 1000.
        base_model: The base model to use for training the surrogate model. Defaults to RandomForestRegressor.
        seed: The seed for the random number generator, it is used to seed a deep copy of the config space.

    Returns:
        An ExplanationTask instance.

    """
    cs = deepcopy(config_space)
    if seed is not None:
        cs.seed(seed)
    samples: list[Configuration] = cs.sample_configuration(n_samples)
    values: list[float] = [function(config) for config in samples]
    data: list[tuple[Configuration, float]] = list(zip(samples, values, strict=False))
    base_model = base_model if base_model is not None else RandomForestRegressor(random_state=seed)

    return ExplanationTask.from_data(config_space=cs, data=data, base_model=base_model)

from_function_multidata(config_space, functions, n_samples=1000, base_model=None) staticmethod

Create an ExplanationTask from a list of functions that evaluate configurations.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
functions	list[Callable[[Configuration], float]]	A list of callables that take a configuration and returns its performance.	required
n_samples	int	The number of configurations to sample for training the surrogate model. Defaults to 1000.	1000
base_model	BaseEstimator | None	The base model to be used for training the surrogate model. Defaults to RandomForestRegressor.	None

Returns:

Type	Description
ExplanationTask	An ExplanationTask instance.

Source code in src/hypershap/task.py

@staticmethod
def from_function_multidata(
    config_space: ConfigurationSpace,
    functions: list[Callable[[Configuration], float]],
    n_samples: int = 1_000,
    base_model: BaseEstimator | None = None,
) -> ExplanationTask:
    """Create an ExplanationTask from a list of functions that evaluate configurations.

    Args:
        config_space: The configuration space.
        functions: A list of callables that take a configuration and returns its performance.
        n_samples: The number of configurations to sample for training the surrogate model. Defaults to 1000.
        base_model: The base model to be used for training the surrogate model. Defaults to RandomForestRegressor.

    Returns:
        An ExplanationTask instance.

    """
    surrogate_model_list = [
        ExplanationTask.from_function(config_space, fun, n_samples, base_model).get_single_surrogate_model()
        for fun in functions
    ]
    return ExplanationTask(config_space=config_space, surrogate_model=surrogate_model_list)

get_hyperparameter_names()

Return the names of the hyperparameters in the configuration space.

Returns:

Type	Description
list[str]	A list of hyperparameter names.

Source code in src/hypershap/task.py

def get_hyperparameter_names(self) -> list[str]:
    """Return the names of the hyperparameters in the configuration space.

    Returns:
        A list of hyperparameter names.

    """
    return list(self.config_space.keys())

get_num_hyperparameters()

Return the number of hyperparameters in the configuration space.

Returns:

Type	Description
int	The number of hyperparameters.

Source code in src/hypershap/task.py

def get_num_hyperparameters(self) -> int:
    """Return the number of hyperparameters in the configuration space.

    Returns:
        The number of hyperparameters.

    """
    return len(self.config_space)

get_single_surrogate_model()

Return the surrogate model for the explanation task.

Returns:

Type	Description
SurrogateModel	The surrogate model for the explanation task.

Source code in src/hypershap/task.py

def get_single_surrogate_model(self) -> SurrogateModel:
    """Return the surrogate model for the explanation task.

    Returns:
        The surrogate model for the explanation task.

    """
    if isinstance(self.surrogate_model, list):
        raise TypeError

    return self.surrogate_model

get_surrogate_model_list()

Return the list of surrogate models for the explanation task.

Returns:

Type	Description
list[SurrogateModel]	The list of surrogate models for the explanation task.

Source code in src/hypershap/task.py

def get_surrogate_model_list(self) -> list[SurrogateModel]:
    """Return the list of surrogate models for the explanation task.

    Returns:
        The list of surrogate models for the explanation task.

    """
    if not isinstance(self.surrogate_model, list):
        raise TypeError

    return self.surrogate_model

is_multi_data()

Return if the explanation task is a multi-data task.

Returns:

Type	Description
bool	True if the explanation task is a multi-data task.

Source code in src/hypershap/task.py

def is_multi_data(self) -> bool:
    """Return if the explanation task is a multi-data task.

    Returns:
        True if the explanation task is a multi-data task.

    """
    return isinstance(self.surrogate_model, list)

MistunabilityExplanationTask

Bases: BaselineExplanationTask

Defines a mistunability explanation task.

Source code in src/hypershap/task.py

class MistunabilityExplanationTask(BaselineExplanationTask):
    """Defines a mistunability explanation task."""

    def __init__(
        self,
        config_space: ConfigurationSpace,
        surrogate_model: SurrogateModel | list[SurrogateModel],
        baseline_config: Configuration,
    ) -> None:
        """Initialize a MistunabilityExplanationTask.

        Args:
            config_space: The configuration space.
            surrogate_model: The (list of) surrogate model(s) used for the explanation task.
            baseline_config: The baseline configuration.

        """
        super().__init__(config_space, surrogate_model, baseline_config)

__init__(config_space, surrogate_model, baseline_config)

Initialize a MistunabilityExplanationTask.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
surrogate_model	SurrogateModel | list[SurrogateModel]	The (list of) surrogate model(s) used for the explanation task.	required
baseline_config	Configuration	The baseline configuration.	required

Source code in src/hypershap/task.py

def __init__(
    self,
    config_space: ConfigurationSpace,
    surrogate_model: SurrogateModel | list[SurrogateModel],
    baseline_config: Configuration,
) -> None:
    """Initialize a MistunabilityExplanationTask.

    Args:
        config_space: The configuration space.
        surrogate_model: The (list of) surrogate model(s) used for the explanation task.
        baseline_config: The baseline configuration.

    """
    super().__init__(config_space, surrogate_model, baseline_config)

MultiBaselineAblationExplanationTask

Bases: MultiBaselineExplanationTask

Defines an ablation explanation task with multiple baseline configurations.

Source code in src/hypershap/task.py

class MultiBaselineAblationExplanationTask(MultiBaselineExplanationTask):
    """Defines an ablation explanation task with multiple baseline configurations."""

    def __init__(
        self,
        config_space: ConfigurationSpace,
        surrogate_model: SurrogateModel | list[SurrogateModel],
        baseline_configs: list[Configuration],
        config_of_interest: Configuration,
    ) -> None:
        """Initialize an MultiBaselineAblationExplanationTask with a list of baseline configurations.

        Args:
            config_space: The configuration space.
            surrogate_model: The (list of) surrogate model(s) used for the explanation task.
            baseline_configs: The baseline configurations.
            config_of_interest: The configuration of interest.

        """
        super().__init__(
            config_space,
            surrogate_model,
            baseline_configs,
        )
        self.config_of_interest = config_of_interest

__init__(config_space, surrogate_model, baseline_configs, config_of_interest)

Initialize an MultiBaselineAblationExplanationTask with a list of baseline configurations.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
surrogate_model	SurrogateModel | list[SurrogateModel]	The (list of) surrogate model(s) used for the explanation task.	required
baseline_configs	list[Configuration]	The baseline configurations.	required
config_of_interest	Configuration	The configuration of interest.	required

Source code in src/hypershap/task.py

def __init__(
    self,
    config_space: ConfigurationSpace,
    surrogate_model: SurrogateModel | list[SurrogateModel],
    baseline_configs: list[Configuration],
    config_of_interest: Configuration,
) -> None:
    """Initialize an MultiBaselineAblationExplanationTask with a list of baseline configurations.

    Args:
        config_space: The configuration space.
        surrogate_model: The (list of) surrogate model(s) used for the explanation task.
        baseline_configs: The baseline configurations.
        config_of_interest: The configuration of interest.

    """
    super().__init__(
        config_space,
        surrogate_model,
        baseline_configs,
    )
    self.config_of_interest = config_of_interest

MultiBaselineExplanationTask

Bases: ExplanationTask

Defines an explanation task with multiple baseline configurations.

Source code in src/hypershap/task.py

class MultiBaselineExplanationTask(ExplanationTask):
    """Defines an explanation task with multiple baseline configurations."""

    def __init__(
        self,
        config_space: ConfigurationSpace,
        surrogate_model: SurrogateModel | list[SurrogateModel],
        baseline_configs: list[Configuration],
    ) -> None:
        """Initialize a MultiBaselineExplanationTask with a list of baseline configurations.

        Args:
            config_space: The configuration space.
            surrogate_model: The (list of) surrogate model(s) used for the explanation task.
            baseline_configs: A list of baseline configurations.

        """
        super().__init__(config_space, surrogate_model)
        self.baseline_configs = baseline_configs

__init__(config_space, surrogate_model, baseline_configs)

Initialize a MultiBaselineExplanationTask with a list of baseline configurations.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
surrogate_model	SurrogateModel | list[SurrogateModel]	The (list of) surrogate model(s) used for the explanation task.	required
baseline_configs	list[Configuration]	A list of baseline configurations.	required

Source code in src/hypershap/task.py

def __init__(
    self,
    config_space: ConfigurationSpace,
    surrogate_model: SurrogateModel | list[SurrogateModel],
    baseline_configs: list[Configuration],
) -> None:
    """Initialize a MultiBaselineExplanationTask with a list of baseline configurations.

    Args:
        config_space: The configuration space.
        surrogate_model: The (list of) surrogate model(s) used for the explanation task.
        baseline_configs: A list of baseline configurations.

    """
    super().__init__(config_space, surrogate_model)
    self.baseline_configs = baseline_configs

OptimizerBiasExplanationTask

Bases: ExplanationTask

Defines an optimizer bias explanation task.

Source code in src/hypershap/task.py

class OptimizerBiasExplanationTask(ExplanationTask):
    """Defines an optimizer bias explanation task."""

    def __init__(
        self,
        config_space: ConfigurationSpace,
        surrogate_model: SurrogateModel | list[SurrogateModel],
        optimizer_of_interest: ConfigSpaceSearcher,
        optimizer_ensemble: list[ConfigSpaceSearcher],
    ) -> None:
        """Initialize an OptimizerBiasExplanationTask.

        Args:
            config_space: The configuration space.
            surrogate_model: The (list of) surrogate model(s) used for the explanation task.
            optimizer_of_interest: The optimizer of interest.
            optimizer_ensemble: The ensemble of optimizers.

        """
        super().__init__(config_space, surrogate_model)
        self.optimizer_of_interest = optimizer_of_interest
        self.optimizer_ensemble = optimizer_ensemble

__init__(config_space, surrogate_model, optimizer_of_interest, optimizer_ensemble)

Initialize an OptimizerBiasExplanationTask.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
surrogate_model	SurrogateModel | list[SurrogateModel]	The (list of) surrogate model(s) used for the explanation task.	required
optimizer_of_interest	ConfigSpaceSearcher	The optimizer of interest.	required
optimizer_ensemble	list[ConfigSpaceSearcher]	The ensemble of optimizers.	required

Source code in src/hypershap/task.py

def __init__(
    self,
    config_space: ConfigurationSpace,
    surrogate_model: SurrogateModel | list[SurrogateModel],
    optimizer_of_interest: ConfigSpaceSearcher,
    optimizer_ensemble: list[ConfigSpaceSearcher],
) -> None:
    """Initialize an OptimizerBiasExplanationTask.

    Args:
        config_space: The configuration space.
        surrogate_model: The (list of) surrogate model(s) used for the explanation task.
        optimizer_of_interest: The optimizer of interest.
        optimizer_ensemble: The ensemble of optimizers.

    """
    super().__init__(config_space, surrogate_model)
    self.optimizer_of_interest = optimizer_of_interest
    self.optimizer_ensemble = optimizer_ensemble

SensitivityExplanationTask

Bases: BaselineExplanationTask

Defines a sensitivity explanation task.

Source code in src/hypershap/task.py

class SensitivityExplanationTask(BaselineExplanationTask):
    """Defines a sensitivity explanation task."""

    def __init__(
        self,
        config_space: ConfigurationSpace,
        surrogate_model: SurrogateModel | list[SurrogateModel],
        baseline_config: Configuration,
    ) -> None:
        """Initialize a SensitivityExplanationTask.

        Args:
            config_space: The configuration space.
            surrogate_model: The (list of) surrogate model(s) used for the explanation task.
            baseline_config: The baseline configuration.

        """
        super().__init__(config_space, surrogate_model, baseline_config)

__init__(config_space, surrogate_model, baseline_config)

Initialize a SensitivityExplanationTask.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
surrogate_model	SurrogateModel | list[SurrogateModel]	The (list of) surrogate model(s) used for the explanation task.	required
baseline_config	Configuration	The baseline configuration.	required

Source code in src/hypershap/task.py

def __init__(
    self,
    config_space: ConfigurationSpace,
    surrogate_model: SurrogateModel | list[SurrogateModel],
    baseline_config: Configuration,
) -> None:
    """Initialize a SensitivityExplanationTask.

    Args:
        config_space: The configuration space.
        surrogate_model: The (list of) surrogate model(s) used for the explanation task.
        baseline_config: The baseline configuration.

    """
    super().__init__(config_space, surrogate_model, baseline_config)

TunabilityExplanationTask

Bases: BaselineExplanationTask

Defines a tunability explanation task.

Source code in src/hypershap/task.py

class TunabilityExplanationTask(BaselineExplanationTask):
    """Defines a tunability explanation task."""

    def __init__(
        self,
        config_space: ConfigurationSpace,
        surrogate_model: SurrogateModel | list[SurrogateModel],
        baseline_config: Configuration,
    ) -> None:
        """Initialize a TunabilityExplanationTask.

        Args:
            config_space: The configuration space.
            surrogate_model: The (list of) surrogate model(s) used for the explanation task.
            baseline_config: The baseline configuration.

        """
        super().__init__(config_space, surrogate_model, baseline_config)

__init__(config_space, surrogate_model, baseline_config)

Initialize a TunabilityExplanationTask.

Parameters:

Name	Type	Description	Default
config_space	ConfigurationSpace	The configuration space.	required
surrogate_model	SurrogateModel | list[SurrogateModel]	The (list of) surrogate model(s) used for the explanation task.	required
baseline_config	Configuration	The baseline configuration.	required

Source code in src/hypershap/task.py

def __init__(
    self,
    config_space: ConfigurationSpace,
    surrogate_model: SurrogateModel | list[SurrogateModel],
    baseline_config: Configuration,
) -> None:
    """Initialize a TunabilityExplanationTask.

    Args:
        config_space: The configuration space.
        surrogate_model: The (list of) surrogate model(s) used for the explanation task.
        baseline_config: The baseline configuration.

    """
    super().__init__(config_space, surrogate_model, baseline_config)