Function dataset

continuiti.data.function.function_dataset

Function data set implementation.

FunctionOperatorDataset(input_function_set, x_sampler, n_sensors, output_function_set, y_sampler, n_evaluations, parameter_sampler, n_observations, x_transform=None, u_transform=None, y_transform=None, v_transform=None)

Bases: OperatorDataset

A dataset class for generating samples from function sets.

This class extends OperatorDataset to specifically handle scenarios where both inputs and outputs are known functions. It utilizes samplers to generate discrete representations of function spaces (input and output function sets) and physical spaces.

PARAMETER	DESCRIPTION
input_function_set	A function set representing the input space. TYPE: FunctionSet
x_sampler	A sampler for generating discrete representations of the domain. TYPE: Sampler
n_sensors	The number of sensors to sample in the input physical space. TYPE: int
output_function_set	A function set representing the output set of the underlying operator. TYPE: FunctionSet
y_sampler	A sampler for generating discrete representations of the codomain. TYPE: Sampler
n_evaluations	The number of evaluation points to sample in the output physical space. TYPE: int
parameter_sampler	A sampler for sampling parameters to instantiate functions from the function sets. TYPE: Sampler
n_observations	The number of observations (=different sets of parameters) to generate. TYPE: int
x_transform,	Optional transformation functions applied to the sampled physical spaces and function evaluations, respectively. TYPE: (u_transform, y_transform, v_transform)

Note

The input_function_set and the output_function_set are evaluated on the same set of parameters. Therefore, the order of parameters need to be taken into consideration when defining both function sets.

Source code in src/continuiti/data/function/function_dataset.py

def __init__(
    self,
    input_function_set: FunctionSet,
    x_sampler: Sampler,
    n_sensors: int,
    output_function_set: FunctionSet,
    y_sampler: Sampler,
    n_evaluations: int,
    parameter_sampler: Sampler,
    n_observations: int,
    x_transform: Optional[Transform] = None,
    u_transform: Optional[Transform] = None,
    y_transform: Optional[Transform] = None,
    v_transform: Optional[Transform] = None,
):
    self.input_function_set = input_function_set
    self.x_sampler = x_sampler
    self.output_function_set = output_function_set
    self.y_sampler = y_sampler
    self.p_sampler = parameter_sampler

    x, u, y, v = self._generate_observations(
        n_sensors=n_sensors,
        n_evaluations=n_evaluations,
        n_observations=n_observations,
    )

    super().__init__(
        x=x,
        u=u,
        y=y,
        v=v,
        x_transform=x_transform,
        u_transform=u_transform,
        y_transform=y_transform,
        v_transform=v_transform,
    )

---

Last update: 2024-08-20
Created: 2024-08-20