reemission.salib.distributions

class reemission.salib.distributions.CategoricalConversionMixin

Bases: object

Mixin for categorical conversion methods.

cat_to_num(cat_values: Sequence[str] | str) → Sequence[int] | int

Convert categorical values to numerical indices

static is_iterable(value) → bool

Check if the value is an iterable. :param value: The value to check.

Returns: bool: True if the value is an iterable and not a scalar, False otherwise.

num_to_cat(num_values: Sequence[int] | int) → Sequence[str] | str

Convert numerical indices to categorical values

class reemission.salib.distributions.CategoricalHistogramDistribution(categories: ~typing.Tuple[str, ...] = <factory>, probabilities: ~typing.Tuple[float, ...] = <factory>, nominal_as_index: bool = True, nominal_category: str | None = None)

Bases: CategoricalConversionMixin, Distribution

Categorical histogram distribution for categorical variables

__post_init__() → None

Ensure categories and probabilities are valid for categorical histogram distribution

categories: Tuple[str, ...]

collapse() → Distribution

Collapse the distribution to a fixed value

nominal_as_index: bool = True

nominal_category: str | None = None

property nominal_value: str

Returns the category with the highest probability

probabilities: Tuple[float, ...]

sample(n_samples: int) → ndarray[tuple[Any, ...], dtype[_ScalarT]]

Generate N samples from the categorical histogram distribution

to_dict() → dict

Convert the distribution to a dictionary representation

property type: str

Returns the type of the parameter as a string.

Returns:

The type of the parameter.

Return type:

str

class reemission.salib.distributions.CategoricalNormalDistribution(categories: ~typing.Tuple[str, ...] = <factory>, mean: float = 0.0, std_dev: float = 1.0, nominal_as_index: bool = True, nominal_category: str | None = None)

Bases: CategoricalConversionMixin, Distribution

Normal distribution for categorical variables This distribution is defined by a mean and standard deviation, and samples are drawn from a normal distribution with respect to the categories. .. attribute:: categories

Categories for the categorical variable.

type:

Tuple[str, …]

mean

Mean of the normal distribution.

Type:

float

std_dev

Standard deviation of the normal distribution.

Type:

float

nominal_as_index

If True, nominal_value is the middle category index, else it’s the middle category itself.

Type:

bool

nominal_category

If provided, overrides nominal_as_index to use this category as the nominal value.

Type:

Optional[str]

__post_init__() → None

Ensure categories are valid for categorical normal distribution

categories: Tuple[str, ...]

collapse() → Distribution

Collapse the distribution to a fixed value

mean: float = 0.0

nominal_as_index: bool = True

nominal_category: str | None = None

property nominal_value: str

Return the category whose index is closest to the mean or return the nominal catergory if provided as attribute.

sample(n_samples: int) → ndarray[tuple[Any, ...], dtype[_ScalarT]]

Generate N samples from the normal distribution for categorical variables

std_dev: float = 1.0

to_dict() → dict

Convert the distribution to a dictionary representation

property type: str

Returns the type of the parameter as a string.

Returns:

The type of the parameter.

Return type:

str

class reemission.salib.distributions.CategoricalUniformDistribution(categories: ~typing.Tuple[str, ...] = <factory>, nominal_as_index: bool = True, nominal_category: str | None = None)

Bases: CategoricalConversionMixin, Distribution

Uniform distribution for categorical variables This distribution is defined by a set of categories and samples are drawn uniformly from these categories. .. attribute:: categories

Categories for the categorical variable.

type:

Tuple[str, …]

nominal_as_index

If True, nominal_value is the middle category index, else it’s the middle category itself.

Type:

bool

nominal_category

If provided, overrides nominal_as_index to use this category as the nominal value.

Type:

Optional[str]

categories: Tuple[str, ...]

collapse() → Distribution

Collapse the distribution to a fixed value

nominal_as_index: bool = True

nominal_category: str | None = None

property nominal_value: str

Return the middle category as nominal value

sample(n_samples: int) → ndarray[tuple[Any, ...], dtype[_ScalarT]]

Generate N samples from the discrete uniform distribution

to_dict() → dict

Convert the distribution to a dictionary representation

property type: str

Returns the type of the parameter as a string.

Returns:

The type of the parameter.

Return type:

str

class reemission.salib.distributions.DiscreteHistogramDistribution(values: ~typing.Tuple[int, ...] = <factory>, probabilities: ~typing.Tuple[float, ...] = <factory>)

Bases: Distribution

Discrete histogram distribution for numerical variables. .. attribute:: values

Values for the discrete histogram distribution.

type:

Tuple[int, …]

probabilities

Probabilities for each value in the discrete histogram distribution.

Type:

Tuple[float, …]

__post_init__() → None

Ensure that values and probabilities are valid for discrete histogram distribution

collapse() → Distribution

Collapse the distribution to a fixed value

property nominal_value: int

Returns the category with the highest probability, or the middle category if all probabilities are equal.

probabilities: Tuple[float, ...]

sample(n_samples: int) → ndarray[tuple[Any, ...], dtype[_ScalarT]]

Generate N samples from the discrete histogram distribution

to_dict() → dict

Convert the distribution to a dictionary representation

property type: str

Returns the type of the parameter as a string.

Returns:

The type of the parameter.

Return type:

str

values: Tuple[int, ...]

class reemission.salib.distributions.DiscreteNormalDistribution(mean: int, std_dev: float)

Bases: Distribution

Discrete normal distribution for numerical variables .. attribute:: mean

Mean of the discrete normal distribution (must be an integer).

type:

int

std_dev

Standard deviation of the discrete normal distribution (must be positive).

Type:

float

__post_init__() → None

Ensure mean is an integer for discrete normal distribution

collapse() → Distribution

Collapse the distribution to a fixed value

mean: int

property nominal_value: int

Returns the nominal value of the distribution”

Returns:

The nominal value or range of values for the distribution.

Return type:

int | float | List[int]

sample(n_samples: int) → ndarray[tuple[Any, ...], dtype[_ScalarT]]

Generate N samples from the discrete normal distribution

std_dev: float

to_dict() → dict

Convert the distribution to a dictionary representation

property type: str

Returns the type of the parameter as a string.

Returns:

The type of the parameter.

Return type:

str

class reemission.salib.distributions.DiscreteUniformDistribution(bounds: Tuple[int, int])

Bases: Distribution

Discrete uniform distribution for numerical variables .. attribute:: bounds

Lower and upper bounds of the discrete uniform distribution (must be integers).

type:

Tuple[int, int]

__post_init__() → None

Ensure bounds are integers for discrete uniform distribution

bounds: Tuple[int, int]

collapse() → Distribution

Collapse the distribution to a fixed value

property nominal_value: int

Returns the nominal value of the distribution”

Returns:

The nominal value or range of values for the distribution.

Return type:

int | float | List[int]

sample(n_samples: int) → ndarray[tuple[Any, ...], dtype[_ScalarT]]

Generate N samples from the discrete uniform distribution

to_dict() → dict

Convert the distribution to a dictionary representation

property type: str

Returns the type of the parameter as a string.

Returns:

The type of the parameter.

Return type:

str

class reemission.salib.distributions.Distribution

Bases: ABC

Abstract base for parameterizations (std-dev or min/max).

cat_to_num(cat_values: Sequence[str]) → Sequence[int]

Default implementation for converting categories to numbers

abstract collapse() → Distribution

Collapse the distribution to a fixed value, if applicable.

Returns:

A fixed distribution with bounds set to the nominal value.

Return type:

Distribution

has_bounds() → bool

Check if the distribution has bounds

is_categorical() → bool

Check if the distribution is categorical

is_discrete() → bool

Check if the distribution is discrete

is_normal() → bool

Check if the distribution is normal

is_uniform() → bool

Check if the distribution is uniform or discrete uniform

abstract property nominal_value: int | float | None

Returns the nominal value of the distribution”

Returns:

The nominal value or range of values for the distribution.

Return type:

int | float | List[int]

num_to_cat(num_values: Sequence[int]) → Sequence[str]

Default implementation for converting numbers to categories

abstract sample(n_samples: int) → ndarray[tuple[Any, ...], dtype[_ScalarT]]

Generate N samples from the distribution

Returns:

An array of samples drawn from the distribution.

Return type:

NDArray

abstract to_dict() → dict

Convert the distribution to a dictionary representation

abstract property type: str

Returns the type of the parameter as a string.

Returns:

The type of the parameter.

Return type:

str

class reemission.salib.distributions.DistributionType(value, names=None, *, module=None, qualname=None, type=None, start=1, boundary=None)

Bases: Enum

Enum for distribution types

CATEGORICALHISTOGRAM = 'categorical-histogram'

CATEGORICALNORMAL = 'categorical-normal'

CATEGORICALUNIFORM = 'categorical-uniform'

DISCRETEHISTOGRAM = 'discrete-histogram'

DISCRETENORMAL = 'discrete-normal'

DISCRETEUNIFORM = 'discrete-uniform'

NORMAL = 'real-normal'

UNIFORM = 'real-uniform'

class reemission.salib.distributions.NormalDistribution(mean: int | float, std_dev: int | float)

Bases: Distribution

Standard (Gaussian) distribution .. attribute:: mean

Mean of the normal distribution.

type:

int | float

std_dev

Standard deviation of the normal distribution.

Type:

int | float

__post_init__() → None

Ensure mean and std_dev are valid for normal distribution

collapse() → Distribution

Collapse the distribution to a fixed value

mean: int | float

property nominal_value: int | float

Returns the nominal value of the distribution”

Returns:

The nominal value or range of values for the distribution.

Return type:

int | float | List[int]

sample(n_samples: int) → ndarray[tuple[Any, ...], dtype[_ScalarT]]

Generate N samples from the normal distribution

std_dev: int | float

to_dict() → dict

Convert the distribution to a dictionary representation

property type: str

Returns the type of the parameter as a string.

Returns:

The type of the parameter.

Return type:

str

class reemission.salib.distributions.UniformDistribution(bounds: Tuple[float, float])

Bases: Distribution

Uniform distribution supported by SALib and Sobol sampling .. attribute:: bounds

Lower and upper bounds of the uniform distribution.

type:

Tuple[float, float]

__post_init__() → None

Ensure bounds are valid for uniform distribution

bounds: Tuple[float, float]

collapse() → Distribution

Collapse the distribution to a fixed value

property nominal_value: int | float

Returns the nominal value of the distribution”

Returns:

The nominal value or range of values for the distribution.

Return type:

int | float | List[int]

sample(n_samples: int) → ndarray[tuple[Any, ...], dtype[_ScalarT]]

Generate N samples from the uniform distribution

to_dict() → dict

Convert the distribution to a dictionary representation

property type: str

Returns the type of the parameter as a string.

Returns:

The type of the parameter.

Return type:

str

class reemission.salib.distributions.Variable(name: str, distribution: ~reemission.salib.distributions.Distribution, fixed: bool = False, group: str | None = None, metadata: dict[str, ~typing.Any] = <factory>)

Bases: object

Represents a variable with a distribution, name, and optional metadata. This class is used to define model parameters and input parameters used in sensitivity analysis.

name

Unique name of the variable.

Type:

str

distribution

Probability distribution associated with the variable.

Type:

Distribution

group

Optional logical group (e.g. “inputs”, “parameters”).

Type:

Optional[str]

metadata

Arbitrary extra metadata for downstream use.

Type:

dict[str, Any]

cat_to_num(cat_values: Sequence[str]) → Sequence[int]

Convert categories to numbers if supported by the distribution, else return input. Catches NotImplementedError if the distribution does not support this conversion.

distribution: Distribution

fixed: bool = False

group: str | None = None

is_categorical() → bool

Check if the variable is categorical based on its distribution

is_discrete() → bool

Check if the variable’s distribution is discrete

is_fixed() → bool

Check if the variable is fixed (not sampled)

is_uniform() → bool

Check if the variable’s distribution is uniform or discrete uniform

metadata: dict[str, Any]

name: str

num_to_cat(num_values: Sequence[int]) → Sequence[str]

Convert numbers to categories if supported by the distribution, else return input. Catches NotImplementedError if the distribution does not support this conversion.

to_dict() → dict

Present the variable to a dictionary representation.

Classes

CategoricalConversionMixin()	Mixin for categorical conversion methods.
CategoricalHistogramDistribution(categories, ...)	Categorical histogram distribution for categorical variables
CategoricalNormalDistribution(categories, ...)	Normal distribution for categorical variables This distribution is defined by a mean and standard deviation, and samples are drawn from a normal distribution with respect to the categories.
CategoricalUniformDistribution(categories, ...)	Uniform distribution for categorical variables This distribution is defined by a set of categories and samples are drawn uniformly from these categories.
DiscreteHistogramDistribution(values, ...)	Discrete histogram distribution for numerical variables.
DiscreteNormalDistribution(mean, std_dev)	Discrete normal distribution for numerical variables .
DiscreteUniformDistribution(bounds)	Discrete uniform distribution for numerical variables .
Distribution()	Abstract base for parameterizations (std-dev or min/max).
DistributionType(value[, names, module, ...])	Enum for distribution types
NormalDistribution(mean, std_dev)	Standard (Gaussian) distribution .
UniformDistribution(bounds)	Uniform distribution supported by SALib and Sobol sampling .
Variable(name, distribution, fixed, group, ...)	Represents a variable with a distribution, name, and optional metadata.