param_space

A collection of parameters jointly defining an operating or optimization space

Classes

ParamSpace(params)

Class designed to define the parameter space in which an optimization can be conducted.

class obsidian.parameters.param_space.ParamSpace(params: list[Parameter])

Bases: IParamSpace

Class designed to define the parameter space in which an optimization can be conducted.

params

A tuple of Parameter types defining the parameter space.

Type:

tuple[Parameter]

X_names

A tuple of the names of the parameters.

Type:

tuple[str]

X_cont

A list of the names of the continuous parameters.

Type:

list[str]

X_obs

A list of the names of the observational parameters.

Type:

list[str]

X_discrete

A dictionary mapping the names of the discrete parameters to their categories.

Type:

dict[str, list[str]]

X_task

A dictionary mapping the names of the task parameters to their categories.

Type:

dict[str, list[str]]

X_min

A tensor containing the minimum values of the parameters.

Type:

torch.Tensor

X_max

A tensor containing the maximum values of the parameters.

Type:

torch.Tensor

X_range

A tensor containing the range of the parameters.

Type:

torch.Tensor

X_static

A list of the names of the static parameters.

Type:

list[str]

X_vary

A list of the names of the varying parameters.

Type:

list[str]

n_dim

The number of dimensions in the parameter space.

Type:

int

n_tdim

The total number of dimensions in the parameter space after transformation.

Type:

int

t_map

A dictionary mapping parameter indices to transformed indices.

Type:

dict

tinv_map

A dictionary mapping transformed indices to parameter indices.

Type:

dict

X_discrete_idx

A list of the indices of the discrete parameters.

Type:

list[int]

X_t_discrete_idx

A list of the indices of the transformed discrete parameters.

Type:

list[int]

X_t_cat_idx

A list of the indices of the transformed categorical parameters.

Type:

list[int]

X_t_task_idx

The index of the transformed task parameter.

Type:

int

search_space

The allowable search space for future optimization.

Type:

pd.DataFrame

Raises:

• ValueError – If the X_names are not unique.

• UnsupportedError – If there is more than one task parameter.

clear_constraints() → None

Clears all constraints from the input space.

constrain_inputs(constraint: Input_Constraint) → None

Constrains the input space based on the specified equality, inequality, or nonlinear constraint.

Parameters:

constraint (Input_Constraint) – The constraint to be applied to the input space.

decode(X)

Decode parameter from transformed space

encode(X)

Encode parameter to a format that can be used for training

classmethod load_state(obj_dict: dict)

Loads the state of the ParamSpace object from a dictionary.

Parameters:

obj_dict (dict) – A dictionary containing the state of the ParamSpace object.

Returns:

A new ParamSpace object with the loaded state.

Return type:

ParamSpace

map_inv_transform() → dict

Maps the inverse of the transformed dictionary.

Returns:

A dictionary where the keys are the original values and the values are the original keys.

Return type:

dict

map_transform() → dict

Maps the parameter indices to transformed indices based on the parameter types.

Returns:

A dictionary mapping parameter indices to transformed indices.

Return type:

dict

mean() → DataFrame

Calculates the mean values for each parameter in the parameter space.

Returns:

A DataFrame containing the mean values for each parameter.

Return type:

pd.DataFrame

open_search() → None

Set the search space to the parameter space

save_state() → dict

Saves the state of the ParamSpace object.

Returns:

A dictionary containing the state of the ParamSpace object.

Return type:

dict

property search_space: DataFrame

Returns the search space for the parameter space.

Returns:

A dataframe containing the search space for the parameter space.

Return type:

pd.DataFrame

unit_demap(X)

Map from 0,1 space to measured space

unit_map(X)

Map from measured to 0,1 space