NumericalTarget¶

class baybe.targets.numerical.NumericalTarget[source]¶

Bases: Target, SerialMixin

Class for numerical targets.

Public methods

`__init__`(name, *args[, ...])	Translate legacy target specifications.
`abs`()	Apply an absolute transformation to the target.
`clamp`([min, max])	Clamp the target to a given range.
`exp`()	Apply an exponential transformation to the target.
`from_dict`(dictionary)	Create an object from its dictionary representation.
`from_json`(string)	Create an object from its JSON representation.
`from_legacy_interface`(name, mode[, bounds, ...])	A deprecation helper for creating targets using the legacy interface.
`from_modern_interface`(name[, ...])	A deprecation helper for creating targets using the modern interface.
`get_codomain`([interval])	Get the codomain of an interval (assuming transformation continuity).
`get_image`([interval])	Get the image of an interval (assuming transformation continuity).
`log`()	Apply a logarithmic transformation to the target.
`match_absolute`(name, match_value, *[, metadata])	Create a target to match a given value using an absolute transformation.
`match_bell`(name, match_value, sigma, *[, ...])	Create a target to match a given value using a bell transformation.
`match_power`(name, match_value, exponent, *)	Create a target to match a given value using a power transformation.
`match_quadratic`(name, match_value, *[, metadata])	Create a target to match a given value using a quadratic transformation.
`match_triangular`(name[, match_value, ...])	Create a target to match a given value using a triangular transformation.
`negate`()	Apply a negation transformation to the target.
`normalize`()	Normalize the target to the unit interval using an affine transformation.
`normalized_ramp`(name, cutoffs, *[, ...])	Create a target that is affine in a given range and clamped to 0/1 outside.
`normalized_sigmoid`(name, anchors, *[, metadata])	Create a sigmoid-transformed target.
`power`(exponent)	Apply a power transformation to the target.
`summary`()	Return a custom summarization of the target.
`to_dict`()	Create an object's dictionary representation.
`to_json`()	Create an object's JSON representation.
`to_objective`()	Create a single-task objective from the target.
`transform`([series, data])	Transform target measurements to computational representation.

Public attributes and properties

`transformation`	An optional target transformation.
`minimize`	Boolean flag indicating if the target is to be minimized.
`description`	The description of the target.
`is_normalized`	Boolean flag indicating if the target is normalized to the unit interval.
`metadata`	Optional metadata containing description, unit, and other information.
`name`	The name of the target.
`unit`	The unit of measurement for the target.

__init__(name: str, *args, _enforce_modern_interface: bool = False, **kwargs)[source]¶

Translate legacy target specifications.

For details on the parameters, see Public attributes and properties.

abs()[source]¶

Apply an absolute transformation to the target.

Return type:: NumericalTarget
Returns:: The target with applied absolute transformation.

clamp(min: float | None = None, max: float | None = None)[source]¶

Clamp the target to a given range.

Parameters:

min (Optional[float]) – The minimum value of the clamping range.
max (Optional[float]) – The maximum value of the clamping range.

Return type:

NumericalTarget

Returns:

The clamped target.

exp()[source]¶

Apply an exponential transformation to the target.

Return type:: NumericalTarget
Returns:: The target with applied exponential transformation.

classmethod from_dict(dictionary: dict)¶

Create an object from its dictionary representation.

Parameters:: dictionary (dict) – The dictionary representation.
Return type:: TypeVar(_T, bound= SerialMixin)
Returns:: The reconstructed object.

classmethod from_json(string: str)¶

Create an object from its JSON representation.

Parameters:: string (str) – The JSON representation of the object.
Return type:: TypeVar(_T, bound= SerialMixin)
Returns:: The reconstructed object.

A deprecation helper for creating targets using the legacy interface.

Parameters:

name (str) – The name of the target.
mode (TargetMode) – The target mode (MAX, MIN, MATCH).
bounds (Union[Interval, Iterable[float], None]) – Optional target bounds.
transformation (Optional[TargetTransformation]) – An optional target transformation.
metadata (Union[MeasurableMetadata, dict[str, Any], None]) – Optional metadata containing description, unit, and other information.

Return type:

NumericalTarget

Returns:

The created target object.

classmethod from_modern_interface(name: str, transformation: Transformation | None = None, *, minimize: bool = False, metadata: MeasurableMetadata | dict[str, Any] | None = None)[source]¶

A deprecation helper for creating targets using the modern interface.

Parameters:

name (str) – The name of the target.
transformation (Optional[Transformation]) – An optional transformation.
minimize (bool) – Boolean flag indicating if the target should be minimized.
metadata (Union[MeasurableMetadata, dict[str, Any], None]) – Optional metadata containing description, unit, and other information.

Return type:

NumericalTarget

Returns:

The created target object.

get_codomain(interval: Interval | None = None, /)[source]¶

Get the codomain of an interval (assuming transformation continuity).

Return type:: Interval

get_image(interval: Interval | None = None, /)[source]¶

Get the image of an interval (assuming transformation continuity).

Return type:: Interval

log()[source]¶

Apply a logarithmic transformation to the target.

Return type:: NumericalTarget
Returns:: The target with applied logarithmic transformation.

classmethod match_absolute(name: str, match_value: float, *, metadata: MeasurableMetadata | dict[str, Any] | None = None)[source]¶

Create a target to match a given value using an absolute transformation.

Parameters:

name (str) – The name of the target.
match_value (float) – The value to be matched.
metadata (Union[MeasurableMetadata, dict[str, Any], None]) – See baybe.targets.numerical.NumericalTarget.

Return type:

NumericalTarget

Returns:

The target with applied absolute matching transformation.

classmethod match_bell(name: str, match_value: float, sigma: float, *, metadata: MeasurableMetadata | dict[str, Any] | None = None)[source]¶

Create a target to match a given value using a bell transformation.

Parameters:

name (str) – The name of the target.
match_value (float) – The value to be matched.
sigma (float) – The scale parameter controlling the width of the bell curve. For more details, see baybe.transformations.basic.BellTransformation.
metadata (Union[MeasurableMetadata, dict[str, Any], None]) – See baybe.targets.numerical.NumericalTarget.

Return type:

NumericalTarget

Returns:

The target with applied bell matching transformation.

classmethod match_power(name: str, match_value: float, exponent: int, *, metadata: MeasurableMetadata | dict[str, Any] | None = None)[source]¶

Create a target to match a given value using a power transformation.

Parameters:

name (str) – The name of the target.
match_value (float) – The value to be matched.
exponent (int) – The exponent of applied the power transformation.
metadata (Union[MeasurableMetadata, dict[str, Any], None]) – See baybe.targets.numerical.NumericalTarget.

Return type:

NumericalTarget

Returns:

The target with applied power matching transformation.

classmethod match_quadratic(name: str, match_value: float, *, metadata: MeasurableMetadata | dict[str, Any] | None = None)[source]¶

Create a target to match a given value using a quadratic transformation.

Parameters:

name (str) – The name of the target.
match_value (float) – The value to be matched.
metadata (Union[MeasurableMetadata, dict[str, Any], None]) – See baybe.targets.numerical.NumericalTarget.

Return type:

NumericalTarget

Returns:

The target with applied quadratic matching transformation.

Create a target to match a given value using a triangular transformation.

Parameters:

name (str) – The name of the target.
match_value (Optional[float]) – The value to be matched. Can be omitted when cutoffs are provided, in which case it defaults to the midpoint.
cutoffs (Union[Interval, Iterable[float], None]) – The cutoff values where the output of the transformation reaches zero.
width (Optional[float]) – The width of the (symmetric) triangular transformation.
margins (Optional[Sequence[float]]) – The margins defining how far the triangle extends in both directions.
metadata (Union[MeasurableMetadata, dict[str, Any], None]) – See baybe.targets.numerical.NumericalTarget.

Raises:

ValueError – If more than one of cutoffs, width, or margins is provided.

Return type:

NumericalTarget

Returns:

The target with applied triangular matching transformation.

negate()[source]¶

Apply a negation transformation to the target.

Return type:: NumericalTarget
Returns:: The target with applied negation transformation.

normalize()[source]¶

Normalize the target to the unit interval using an affine transformation.

Raises:: IncompatibilityError – If the target image is not bounded.
Return type:: NumericalTarget
Returns:: The normalized target.

classmethod normalized_ramp(name: str, cutoffs: Interval | Iterable[float] | None, *, descending: bool = False, metadata: MeasurableMetadata | dict[str, Any] | None = None)[source]¶

Create a target that is affine in a given range and clamped to 0/1 outside.

Parameters:

name (str) – The name of the target.
cutoffs (Union[Interval, Iterable[float], None]) – The cutoff values defining the affine region.
descending (bool) – Boolean flag indicating if the transformation is ascending or descending in the affine region.
metadata (Union[MeasurableMetadata, dict[str, Any], None]) – See baybe.targets.numerical.NumericalTarget.

Return type:

NumericalTarget

Returns:

The target with applied clamped affine transformation.

classmethod normalized_sigmoid(name: str, anchors: Sequence[Sequence[float]], *, metadata: MeasurableMetadata | dict[str, Any] | None = None)[source]¶

Create a sigmoid-transformed target.

Parameters:

name (str) – The name of the target.
anchors (Sequence[Sequence[float]]) – See baybe.transformations.basic.SigmoidTransformation.
metadata (Union[MeasurableMetadata, dict[str, Any], None]) – See baybe.targets.numerical.NumericalTarget.

Return type:

NumericalTarget

Returns:

The target with applied sigmoid transformation.

power(exponent: int)[source]¶

Apply a power transformation to the target.

Parameters:: exponent (int) – The exponent of the power transformation.
Return type:: NumericalTarget
Returns:: The target with applied power transformation.

summary()[source]¶: Return a custom summarization of the target.

to_dict()¶

Create an object’s dictionary representation.

Return type:: dict
Returns:: The dictionary representation of the object.

to_json()¶

Create an object’s JSON representation.

Return type:: str
Returns:: The JSON representation as a string.

to_objective()¶

Create a single-task objective from the target.

Return type:: SingleTargetObjective

transform(series: Series | None = None, /, *, data: DataFrame | None = None)[source]¶

Transform target measurements to computational representation.

Parameters:: series (Optional[Series]) – The target measurements in experimental representation to be transformed.
Return type:: Series
Returns:: A series containing the transformed measurements. The series name matches that of the input.

property description: str | None¶: The description of the target.

property is_normalized: UncertainBool¶: Boolean flag indicating if the target is normalized to the unit interval.

metadata: MeasurableMetadata¶: Optional metadata containing description, unit, and other information.

minimize: bool¶: Boolean flag indicating if the target is to be minimized.

name: str¶: The name of the target.

transformation: Transformation¶: An optional target transformation.

property unit: str | None¶: The unit of measurement for the target.