Continuous

ContinuousParameter (BaseParameter)

Continuous parameter represented by a float in a given range.

Source code in parameterspace/parameters/continuous.py

class ContinuousParameter(BaseParameter):
    """Continuous parameter represented by a float in a given range."""

    @store_init_arguments
    def __init__(
        self,
        name: str,
        bounds: Tuple[float, float],
        *,
        prior: Optional[BasePrior] = None,
        transformation: Union[BaseTransformation, str, None] = None,
        inactive_numerical_value: Optional[float] = np.nan,
    ):
        """
        Initialize with options for continuous parameter.

        Args:
            name: Name of the parameter.
            bounds: Lower and upper bound for the parameter.
            prior: Default prior is the Uniform prior which will fail if at least one
                of the bounds is infinite!
            transformation: Default transformation is the affine-linear transformation
                of the bounds to `[0,1]`. If a string is given, the transform is created
                automatically for supported transforms.
            inactive_numerical_value:  Placeholder value for this parameter in case
                it is not active.
        """
        if not np.isfinite(bounds).all():
            raise ValueError(f"Bounds need to be finite, but are {bounds}")

        self.bounds = np.array(bounds)

        if not isinstance(transformation, BaseTransformation):
            if transformation is None:
                transformation = ZeroOneFloat(self.bounds)
            elif transformation == "log":
                transformation = LogZeroOneFloat(self.bounds)

        prior = Uniform() if prior is None else prior
        super().__init__(
            name,
            prior,
            transformation,
            is_continuous=True,
            is_ordered=True,
            num_values=np.inf,
            inactive_numerical_value=inactive_numerical_value,
        )

    def __repr__(self):
        """Add bounds to the string representation."""
        string = super().__repr__()
        string += f"Bounds: ({self.bounds[0]}, {self.bounds[1]})\n"
        return string

    def check_value(self, value):
        """Check if value is valid."""
        return self.bounds[0] <= value <= self.bounds[1]

__init__(self, name, bounds, *, prior=None, transformation=None, inactive_numerical_value=nan) special

Initialize with options for continuous parameter.

Parameters:

Name	Type	Description	Default
name	str	Name of the parameter.	required
bounds	Tuple[float, float]	Lower and upper bound for the parameter.	required
prior	Optional[parameterspace.priors.base.BasePrior]	Default prior is the Uniform prior which will fail if at least one of the bounds is infinite!	None
transformation	Union[parameterspace.transformations.base.BaseTransformation, str]	Default transformation is the affine-linear transformation of the bounds to [0,1]. If a string is given, the transform is created automatically for supported transforms.	None
inactive_numerical_value	Optional[float]	Placeholder value for this parameter in case it is not active.	nan

Source code in parameterspace/parameters/continuous.py

@store_init_arguments
def __init__(
    self,
    name: str,
    bounds: Tuple[float, float],
    *,
    prior: Optional[BasePrior] = None,
    transformation: Union[BaseTransformation, str, None] = None,
    inactive_numerical_value: Optional[float] = np.nan,
):
    """
    Initialize with options for continuous parameter.

    Args:
        name: Name of the parameter.
        bounds: Lower and upper bound for the parameter.
        prior: Default prior is the Uniform prior which will fail if at least one
            of the bounds is infinite!
        transformation: Default transformation is the affine-linear transformation
            of the bounds to `[0,1]`. If a string is given, the transform is created
            automatically for supported transforms.
        inactive_numerical_value:  Placeholder value for this parameter in case
            it is not active.
    """
    if not np.isfinite(bounds).all():
        raise ValueError(f"Bounds need to be finite, but are {bounds}")

    self.bounds = np.array(bounds)

    if not isinstance(transformation, BaseTransformation):
        if transformation is None:
            transformation = ZeroOneFloat(self.bounds)
        elif transformation == "log":
            transformation = LogZeroOneFloat(self.bounds)

    prior = Uniform() if prior is None else prior
    super().__init__(
        name,
        prior,
        transformation,
        is_continuous=True,
        is_ordered=True,
        num_values=np.inf,
        inactive_numerical_value=inactive_numerical_value,
    )

__repr__(self) special

Add bounds to the string representation.

Source code in parameterspace/parameters/continuous.py

def __repr__(self):
    """Add bounds to the string representation."""
    string = super().__repr__()
    string += f"Bounds: ({self.bounds[0]}, {self.bounds[1]})\n"
    return string

check_value(self, value)

Check if value is valid.

Source code in parameterspace/parameters/continuous.py

def check_value(self, value):
    """Check if value is valid."""
    return self.bounds[0] <= value <= self.bounds[1]