NDUncertainty#
- class astropy.nddata.NDUncertainty(array=None, copy=True, unit=None)[source]#
Bases:
object
This is the metaclass for uncertainty classes used with
NDData
.- Parameters:
- array
any
type, optional The array or value (the parameter name is due to historical reasons) of the uncertainty.
numpy.ndarray
,Quantity
orNDUncertainty
subclasses are recommended. If thearray
islist
-like ornumpy.ndarray
-like it will be cast to a plainnumpy.ndarray
. Default isNone
.- unitastropy:unit-like, optional
Unit for the uncertainty
array
. Strings that can be converted to aUnit
are allowed. Default isNone
.- copy
bool
, optional Indicates whether to save the
array
as a copy.True
copies it before saving, whileFalse
tries to save every parameter as reference. Note however that it is not always possible to save the input as reference. Default isTrue
.
- array
- Raises:
IncompatibleUncertaintiesException
If given another
NDUncertainty
-like class asarray
if theiruncertainty_type
is different.
Attributes Summary
numpy.ndarray
: the uncertainty's value.NDData
: reference toNDData
instance with this uncertainty.This uncertainty as an
Quantity
object.bool
: Supports uncertainty propagation with correlated uncertainties?str
: Short description of the type of uncertainty.Unit
: The unit of the uncertainty, if any.Methods Summary
propagate
(operation, other_nddata, ...[, axis])Calculate the resulting uncertainty given an operation on the data.
represent_as
(other_uncert)Convert this uncertainty to a different uncertainty type.
Attributes Documentation
- array#
numpy.ndarray
: the uncertainty’s value.
- parent_nddata#
NDData
: reference toNDData
instance with this uncertainty.In case the reference is not set uncertainty propagation will not be possible since propagation might need the uncertain data besides the uncertainty.
bool
: Supports uncertainty propagation with correlated uncertainties?Added in version 1.2.
- uncertainty_type#
str
: Short description of the type of uncertainty.Defined as abstract property so subclasses have to override this.
Methods Documentation
- propagate(operation, other_nddata, result_data, correlation, axis=None)[source]#
Calculate the resulting uncertainty given an operation on the data.
Added in version 1.2.
- Parameters:
- operation
callable()
The operation that is performed on the
NDData
. Supported arenumpy.add
,numpy.subtract
,numpy.multiply
andnumpy.true_divide
(ornumpy.divide
).- other_nddata
NDData
instance The second operand in the arithmetic operation.
- result_data
Quantity
orndarray
The result of the arithmetic operations on the data.
- correlation
numpy.ndarray
or number The correlation (rho) is defined between the uncertainties in sigma_AB = sigma_A * sigma_B * rho. A value of
0
means uncorrelated operands.- axis
int
ortuple
ofint
, optional Axis over which to perform a collapsing operation.
- operation
- Returns:
- resulting_uncertainty
NDUncertainty
instance Another instance of the same
NDUncertainty
subclass containing the uncertainty of the result.
- resulting_uncertainty
- Raises:
ValueError
If the
operation
is not supported or if correlation is not zero but the subclass does not support correlated uncertainties.
Notes
First this method checks if a correlation is given and the subclass implements propagation with correlated uncertainties. Then the second uncertainty is converted (or an Exception is raised) to the same class in order to do the propagation. Then the appropriate propagation method is invoked and the result is returned.
- represent_as(other_uncert)[source]#
Convert this uncertainty to a different uncertainty type.
- Parameters:
- other_uncert
NDUncertainty
subclass The
NDUncertainty
subclass to convert to.
- other_uncert
- Returns:
- resulting_uncertainty
NDUncertainty
instance An instance of
other_uncert
subclass containing the uncertainty converted to the new uncertainty type.
- resulting_uncertainty
- Raises: