auto_uncertainties.uncertainty.uncertainty_containers

Attributes

`UType`	Type alias for the underlying data types supporting `Uncertainty` objects.
`VectorUncertainty`	Alias for `Uncertainty` to maintain backward compatibility.
`ScalarUncertainty`	Alias for `Uncertainty` to maintain backward compatibility.

Classes

Uncertainty

Representation of a central value and its associated uncertainty.

Functions

`nominal_values`(…)	Return the central value of an `Uncertainty` object if it is one, otherwise returns the object.
`std_devs`(…)	Return the uncertainty of an `Uncertainty` object if it is one, otherwise returns zero.

Module Contents

class auto_uncertainties.uncertainty.uncertainty_containers.UType[source]

Type alias for the underlying data types supporting Uncertainty objects.

TYPE ALIAS:

float
| floating
| NDArray[floating]

class auto_uncertainties.uncertainty.uncertainty_containers.Uncertainty(value: int, error: ErrT | None = ...)[source]

class auto_uncertainties.uncertainty.uncertainty_containers.Uncertainty(value, error=...)

class auto_uncertainties.uncertainty.uncertainty_containers.Uncertainty(value)

class auto_uncertainties.uncertainty.uncertainty_containers.Uncertainty(value, error=...)

class auto_uncertainties.uncertainty.uncertainty_containers.Uncertainty(value, error=..., skip=...)

Bases: Generic[T], auto_uncertainties.UncertaintyDisplay

Inheritance diagram of auto_uncertainties.uncertainty.uncertainty_containers.Uncertainty

Representation of a central value and its associated uncertainty.

Parameters can be numbers, sequences, numpy arrays, pint.Quantity objects, other Uncertainty objects, or lists / tuples of Uncertainty objects.

The Uncertainty class automatically determines which methods should be implemented based on whether it represents a vector uncertainty, or a scalar uncertainty. When instantiated with sequences or numpy arrays, vector-based operations are enabled; when instantiated with scalars, only scalar operations are permitted.

Uncertainty objects only support float-based data types. If integers or integer arrays are passed as parameters to the Uncertainty constructor, they will be cast to float (or numpy.float64 if a numpy.integer subclass is detected).

Parameters:

value – The central value(s)
error – The uncertainty value(s). Zero if not provided.

Raises:

NegativeStdDevError – If err is negative, or contains negative values
TypeError – If the parameters are of incompatible types
ValueError – If the parameters have incomaptible values (e.g., misaligned array sizes)

Returns:

An initialized Uncertainty object

Example

>>> u1 = Uncertainty(1.25, 0.25)
>>> u2 = Uncertainty([1.4, 2.8, 0.09], [0.1, 0.14, 0.12])
>>> u3 = Uncertainty([1.4, 2.8, 0.09], 0.1)
>>> u4 = Uncertainty(u1)
>>> u5 = Uncertainty(np.array([1.4, 2.8]), np.array([0.1, 0.14]))
>>> u6 = Uncertainty(np.array([1.4, 2.8]), 0.1)

>>> u3.value
array([1.4 , 2.8 , 0.09])

>>> u3.error
array([0.1, 0.1, 0.1])

>>> np.cos(u1)
0.315322 +/- 0.237246

>>> u4 == u1
True

Pint Quantity Example

>>> from pint import Quantity
>>> val = Quantity(2.24, 'kg')
>>> err = Quantity(0.208, 'kg')
>>> new_quantity = Uncertainty(val, err)
>>> new_quantity
<Quantity(2.24 +/- 0.208, 'kilogram')>

Note

If sequences (not NumPy arrays) are supplied for value and error, their numeric values will always be converted to numpy.float64.
If pint.Quantity objects are supplied for either parameter, the behavior is exactly as described in the from_quantities method.
If an Uncertainty is supplied for value, its error attribute will always override any error argument (if it is supplied).
If the error parameter is not finite, the resulting Uncertainty object will have its error attribute set to zero.