lightcurvelynx.math_nodes.scipy_random

Wrapper classes for some of scipy’s sampling functions.

Classes

`NumericalInversePolynomialFunc`	A class for sampling from scipy's NumericalInversePolynomial
`PDFFunctionWrapper`	A class that just wraps a given PDF function.
`SamplePDF`	A node for sampling from a given PDF function.
`LogPDFFunctionWrapper`	A class that just wraps a given Log PDF function.
`SampleLogPDF`	A node for sampling from a given Log PDF function.
`ScipyRandomDist`	The base class sampling from scipy.stats distributions.

Module Contents

class NumericalInversePolynomialFunc(dist=None, *, domain=None, seed=None, **kwargs)[source]

Bases: lightcurvelynx.base_models.FunctionNode

A class for sampling from scipy’s NumericalInversePolynomial given a distribution function, an object with a pdf function, or a class from which to create such an object.

Note

If a class is provided, then the sampling function will create a new object (with the sampled parameters) for each sampling. This is very expensive.

Parameters:

dist (object or class) – An object or class with either a pdf() or logpdf() method that defines the distribution from which to sample.
domain (tuple, optional) – A tuple of (min, max) values to use as bounds for the sampling. If not provided, scipy will try to infer the domain.
seed (int, optional) – The seed to use.

set_seed(new_seed)[source]

Update the random number generator’s seed to a given value.

Parameters:: new_seed (int) – The given seed

compute(graph_state, rng_info=None, **kwargs)[source]

Execute the wrapped function.

The input arguments are taken from the current graph_state and the outputs are written to graph_state.

Parameters:

graph_state (GraphState) – An object mapping graph parameters to their values. This object is modified in place as it is sampled.
rng_info (numpy.random._generator.Generator, optional) – A given numpy random number generator to use for this computation. If not provided, the function uses the node’s random number generator.
**kwargs (dict, optional) – Additional function arguments.

Returns:

results – The result of the computation. This return value is provided so that testing functions can easily access the results.

Return type:

any

class PDFFunctionWrapper(func)[source]

A class that just wraps a given PDF function.

_pdf[source]

The PDF function.

Type:: function

pdf[source]

class SamplePDF(dist, *, domain=None, **kwargs)[source]

Bases: NumericalInversePolynomialFunc

A node for sampling from a given PDF function.

Parameters:

dist (function, class, or object) – The pdf function from which to sample or a class/object with that function.
domain (tuple, optional) – A tuple of (min, max) values to use as bounds for the sampling. If not provided, scipy will try to infer the domain.

class LogPDFFunctionWrapper(func)[source]

A class that just wraps a given Log PDF function.

_log_pdf[source]

The log PDF function.

Type:: function

logpdf[source]

class SampleLogPDF(dist, **kwargs)[source]

Bases: NumericalInversePolynomialFunc

A node for sampling from a given Log PDF function.

Parameters:: dist (function, class, or object) – The pdf function from which to sample or a class/object with that function.

class ScipyRandomDist(dist_name, seed=None, node_label=None, **kwargs)[source]

Bases: lightcurvelynx.base_models.FunctionNode

The base class sampling from scipy.stats distributions.

dist_class[source]

The distribution from which to sample.

Type:: scipy.stats distribution object

_rng[source]

This object’s random number generator.

Type:: numpy.random._generator.Generator

sample_size

The shape of the array to generate for each sample. The actual returned value will be (num_samples, *size). If an empty tuple will generate a single value per sample.

Type:: tuple

params_names[source]

The names of the parameters used to create the distribution object.

Type:: list of str

Parameters:

dist_name (str) – The name of the distribution from which to sample.
seed (int, optional) – The seed to use.
node_label (str, optional) – An optional label for the node. If not provided, a default label will be created based on the function name and parameters.
**kwargs (dict, optional) – The parameters to use to create the distribution object.

dist_class[source]

params_names = [][source]

set_seed(new_seed)[source]

Update the random number generator’s seed to a given value.

Parameters:: new_seed (int) – The given seed

compute(graph_state, rng_info=None, **kwargs)[source]

Execute the wrapped function.

The input arguments are taken from the current graph_state and the outputs are written to graph_state.

Parameters:

graph_state (GraphState) – An object mapping graph parameters to their values. This object is modified in place as it is sampled.
rng_info (numpy.random._generator.Generator, optional) – A given numpy random number generator to use for this computation. If not provided, the function uses the node’s random number generator.
**kwargs (dict, optional) – Additional function arguments.

Returns:

results – The result of the computation. This return value is provided so that testing functions can easily access the results.

Return type:

any

Raises:

ValueError – is func attribute is None.