sampling

Sampling helpers for generated segments.

pysatl_cpd.data.generator.segments.sampling.feature_names_for_distribution(distribution)[source]

Extract feature names from a distribution specification.

Parameters:: distribution (MultivariateNormalSpec | IndependentColumnsSpec | NormalSpec | UniformSpec | ExponentialSpec | StudentTSpec) – Distribution specification to extract feature names from.
Returns:: Tuple of feature names for the distribution.
Return type:: tuple[str, ...]
Raises:: TypeError – If the distribution type is not supported.

pysatl_cpd.data.generator.segments.sampling.sample_distribution(distribution, length, rng)[source]

Sample data from a distribution specification.

Parameters:

distribution (MultivariateNormalSpec | IndependentColumnsSpec | NormalSpec | UniformSpec | ExponentialSpec | StudentTSpec) – Distribution specification to sample from.
length (int) – Number of samples to generate.
rng (Generator) – Random number generator for reproducible sampling.

Returns:

Array of sampled data with shape (length, num_features).

Return type:

ndarray[tuple[int, ...], dtype[double]]

Raises:

TypeError – If the distribution type is not supported.

pysatl_cpd.data.generator.segments.sampling.sample_univariate_distribution(distribution, length, rng)[source]

Sample from a univariate distribution specification.

Parameters:

distribution (NormalSpec | UniformSpec | ExponentialSpec | StudentTSpec) – Univariate distribution specification.
length (int) – Number of samples to generate.
rng (Generator) – Random number generator for reproducible sampling.

Returns:

Array of sampled values.

Return type:

ndarray[tuple[int, ...], dtype[double]]

Raises:

TypeError – If the distribution type is not supported.

pysatl_cpd.data.generator.segments.sampling.build_covariance_matrix(covariance, dimension)[source]

Construct a covariance matrix from supported covariance inputs.

Parameters:

covariance (ndarray[tuple[int, ...], dtype[double]] | tuple[tuple[float, ...], ...] | tuple[float, ...] | float) – Covariance as scalar, 1D diagonal, or full matrix.
dimension (int) – Expected dimension of the covariance matrix.

Returns:

Properly shaped covariance matrix.

Return type:

ndarray[tuple[int, ...], dtype[double]]

Raises:

ValueError – If the covariance input does not match the expected dimension.