Data Generator

class symdet.data.data_generator.DataGenerator[source]

A class to generate data for use in the Symmetry analysis.

domain

Domain values of the function.

Type

tf.Tensor

image

Image values of the function, i.e. f(x) for all x.

Type

tf.Tensor

image_size

Size of the data pool.

Type

int

domain_shape

Shape of the domain points.

Type

tuple

clustered_data

A dictionary of clustered data.

Type

dict

Methods

build_clusters(**kwargs)

Split the raw function data into classes.

load_data(points[, save])

Load some data either from a computation or from a pool into the class state.

plot_clusters([save])

Plot the data clusters.

plot_data([save])

Plot the data.

build_clusters(**kwargs)[source]

Split the raw function data into classes.

Parameters

**kwargs

Return type

Updates the class state.

Notes

In the double well potential we can simply use the range_binning clustering algorithm.

load_data(points: Union[int, ndarray], save: bool = False)[source]

Load some data either from a computation or from a pool into the class state.

Parameters
  • points (Union[int, np.ndarray]) – Points to generate, either an np.ndarray or an integer. If an integer, N points will be generated, if an array, it will either be treated as input to a function to generate values or those indices will be drawn from a pool.

  • save (bool) – If true, save the data after generating it.

plot_clusters(save: bool = False)[source]

Plot the data clusters.

Parameters

save (bool) – If true the figure will be saved.

plot_data(save: bool = False)[source]

Plot the data.

Parameters

save (bool) – If true the figure will be saved.