rmellipse.uobjects

This module contains objects that represent data with uncertainties.

Data with uncertainties correspond to propagators in rmellipse.propagators

Exceptions

RMEMeasFormatError

Error in formatting of data inside RMEMeas

Classes

UObj

Generic Uncertain Object.

RMEMeas

Class that stores monte-carlo and linear sensitivity uncertainty information.

Package Contents

class rmellipse.uobjects.UObj

Bases: abc.ABC

Generic Uncertain Object.

Defines the interface that all uncertainty objects should inherit from.

abstract property nom

Get the nominal value of the uncertain object.

Return type should be the natural typing for the object being evauluated.

Return type:

Nominal value.

class rmellipse.uobjects.RMEMeas(name: str = None, cov: xarray.DataArray = None, mc: xarray.DataArray = None, covdofs: xarray.DataArray = None, covcats: xarray.DataArray = None, parent: rmellipse.utils.GroupSaveable = None, attrs: dict = None)

Bases: rmellipse.uobjects.UObj, rmellipse.utils.GroupSaveable

Class that stores monte-carlo and linear sensitivity uncertainty information.

Used with the rmellipse.RMEProp propagator.

Initialize a RMEMeas object.

Please refer to xarray’s documentation for information about how to use DataArrays, and how to define coordinates and dimensions.

If covdofs is not provided, distributions of linear uncertainty mechanisms are assumed to have infinite degrees of freedom (Gaussian).

Parameters:

  • name (str, optional) – Name of RMEMeas object. The default is ‘RMEMeas’.

  • cov (xr.DataArray, optional) –

    Covariance data for linear sensitivity analysis. Copies of data set are stored along the first dimension (axis 0) of the DataArray, where the first index of axis 0 is the nominal data set, and the rest of the indexes along that dimension are perturbed by one standard deviation (i.e. 1 standard uncertainty).

    The first dimension must be called ‘umech_id’, and the first label of the ‘parameter_dimensions’ coordinate must be ‘nominal’. The remaining labels for the ‘umech_id’ coordinate should be strings corresponding the the uncertainty mechanism. The default is None.

  • mc (xr.DataArray, optional) –

    Montecarlo trials. Samples of the data sets distribution are stored along the first dimension (axis 0) of the DataArray, where the first index of axis 0 is the nominal data set, and the rest of the indexes are samples of the distribution.

    The first dimension must be called ‘sample_id’, and the first labels of the ‘sample_id’ must start at 0 and count up by 1 (i,e typical integer based indexing). The default is None.

  • covdofs (xr.DataArray, optional) – DataArray that stores the degrees of freedom for each linear uncertainty mechanism in cov. It should be a 1 dimensional DataArray with the dimension called ‘umech_id’ and the coordinate set should be identical to the ‘umech_id’ coordinate in cov. If covariance data is provided and covdofs is not, one will be created that assumes all linear mechanisms have infinite degrees of freedom (i.e. gaussian distributions). The default is None.

  • covcats (xr.DataArray, optional) – DataArray that stores the categories of each uncertainty mechanism. Expected to be have dimensions (‘umech_id’,’categories’), If not provided, all umech_id are assigned a ‘Type’ category of ‘B’. If a mechanism doesn’t have a category, it should be an empty string.

property name

The name of the object.

Type:

str

property cov

Linear uncertainty mechanisms.

Type:

xr.DataArray

property mc

samples of Monte-Carlo distributions.

Type:

xr.DataArray

property covdofs

The degrees of freedom on linear mechanisms.

Type:

xr.DataArray

property covcats

String metadata for linear uncertainty mechanisms.

Type:

xr.DataArray

property dims
property shape
property coords
property dtype
cast_covcats()
cast_umechids()

Cast any umech_id dimensions to the correct type.

classmethod from_nom(name: str, nom: xarray.DataArray) → RMEMeas

Create a RMEMeas object with just a nominal dataset.

Parameters:

  • name (str) – The name of the object to be created.

  • nom (xr.DataArray) – Nominal data set.

Returns:

A RMEMeas object with only nominal values.

Return type:

‘RMEMeas’

classmethod from_dist(name: str, nom: float, std: float, dist='gaussian', mechanism_name: str = None, samples: float = 100, categories: dict[str] = {'Type': 'B'}, use_sample_mean: bool = False) → RMEMeas

Generate a RMEMeas object from a probability distribution.

Will be deprecated in V0.5. Please use RMEmodel instead.

Parameters:

  • name (str) – Name of the object.

  • nom (float) – Expected value of the distribution.

  • std (float) – Standard deviation of the distribution.

  • dist (str, {'gaussian', 'normal', 'uniform', 'rectangular'}) – Name of the distribution to use. Supports ‘gaussian’ or ‘uniform’. The default is ‘gaussian’.

  • mechanism_name (str, optional) – What to name the linear uncertainty mechanisms associated with the distribution. The default is None.

  • samples (float, optional) – How many monte-carlo samples to draw from. The default is 100.

  • categories (dict[str], optional) – What to categorize the linear uncertainty mechanism as Should be {category:value} pairs. The default is {‘Type’:’B’}.

  • use_sample_mean_std (bool, optional) – If true, uses the mean and standard deviation of random samples drawn from the defined distribution for the linear sensitivity analysis and as the nominal and standard uncertainty values. If False, uses the provided nominal and standard deviation of the distribution. The default is True.

Raises:

Exception – Unsupported distribution.

Returns:

RMEMeas based on defined distribution.

Return type:

‘RMEMeas’

classmethod dict_from_group(file: str, group_path: str = None, verbose: bool = False) → dict[RMEMeas]

Read any RMEMeas objects saved in an hdf5 group.

Parameters:

  • file (str) – path to the hdf5 file.

  • group_path (str, optional) – path to group. The default is None.

  • verbose (bool , optional) – If True, prints information about attempts to read files. The default is False.

Returns:

Dictionary where keys are names of the RMEMeas objects and values are the RMEMeas objects themselves.

Return type:

dict[‘RMEMeas’]

classmethod from_h5(group: h5py.Group, nominal_only: bool = False, keep_attrs: bool = True) → RMEMeas

Read a RMEMeas object from HDF5.

Parameters:

  • group – HDF5 Group object

  • nominal_only (bool, optional) – If true, will only load the nominal value of the data set. Saves time and memory if you don’t need that data.

  • keep_attrs (bool, optional) – If true, copies over all metadata in attrs. Otherwise, only keeps metadata required for class instantiation.

Returns:

RMEMeas object.

Return type:

RMEMeas

to_h5(group: h5py.Group, override: bool = False, name: str = None, verbose: bool = False)

Save to an hdf5 object. Wrapper for the group_saveable method ‘save’.

Parameters:

  • group (h5py.Group) – hdf5 group to save under.

  • override (bool,) – if True, will delete the group object being saved to if it already exists. Default is false.

  • name (str,) – If provided will change Name of RMEMeas object prior to saving. Equivalent to calling self.name = name prior to calling save.

Return type:

None.

classmethod from_xml(path: str, from_csv: callable, old_dir: str = None, new_dir: str = None, verbose: bool = False) → RMEMeas

Read a legacy xml MUFmeas object (xml .meas format).

covcats and covdofs metadata will be loaded as the default values.

Parameters:

  • path (str) – Path to xml header file.

  • from_csv (callable) – Read function, takes a path to a copy of the data file and returns an xarray object.

  • old_dir (str, optional) – Name of old path stored in xml file. Will be swapped with new_dir if provided. Old XML format isn’t portable, and the paths need to be manually swapped when the files are moved around. The default is None.

  • new_dir (str, optional) – Path string to replace old_dir with. The default is None.

Raises:

Exception – If old_dir/new_dir are not both provided, but one is.

Returns:

new – MUFmeas object from the legacy format.

Return type:

MUFmeas

to_xml(target_directory: str, to_csv: callable, data_extension: str, header_extension: str = '.meas', header_directory: str = None)

Save to a the Microwave Uncertainty Framework Format.

This does not preserve metadata in covdofs or covcats.

Parameters:

  • target_directory (str) – Directory in which to save the support folder which stores all the copies of the data.

  • to_csv (callable) – Dataformat of the underlying cov-data. Will be inferred if left as None. The default is None.

  • data_extension (str) – What to save the datafile extensions as (e.g. .s1p, .csv, etc)

  • header_extension (str, optional) – What to save the xml header file extension as. The default is ‘.meas’.

  • header_directory (str, optional) – Location in which to store the header file. If None, defaults to the target_directory. Default is None.

Raises:

Exception – If no dataformat provided and one cannot be inferred.

Return type:

None.

copy() → RMEMeas

Make a copy of a RMEMeas object.

Returns:

Copied object.

Return type:

RMEMeas

cull_cov(tolerance: float = 0)

Remove trivial linear uncertainty mechanisms set by tolerance.

Any linear uncertainty mechanisms with a sum of all standard uncertainties < tolerance will be removed from the object.

For example if tolerance is 0.1, the nominal is [0,0,0] and the perturbed data for a mechanisms is [0.0,0.1,-0.2], then the function will evaluate the total standard uncertainty for the mechanism as 0.1 + 0.0 + 0.2 = 0.3, and not remove the mechanism.

This function is called by the auto-cull setting in the RME propagator.

Parameters:

tolerance (float, optional) – Maximum value a sum of standard uncertainties that is deemed trivial, used to determine what mechanisms should be removed. The default is 0.

Return type:

None.

make_umechs_unique(same_uid: bool = False)

Make parameter locations unique by adding a uuid4 string to end of each.

A uiid4 string is added to the end of each parameter_location string. Useful when reading data from different sources that have the same names of uncertainty files. (E.G, multiple device definitions come from calibration service formats with generic ‘ua,ub’ naming)

Parameters:

same_uid (bool, optional) – If true, all parameter locations will have the same uid added to the end. Default is False.

Return type:

None.

add_mc_sample(sample: xarray.DataArray)

Add a Monte Carlo sample to the distribution.

Parameters:

sample (xr.DataArray) – Single sample of the probability distribution that represents the data set. Expected to be the same shape, dimensions, and coordinates of the nominal.

Return type:

None.

add_umech(name: str, value: xarray.DataArray, dof: float = np.inf, category: dict = {'Type': 'B'}, add_uid=None)

Add a linear mechanisms to covariance data.

Note:

Parameters:

  • name (str) – Name of new mechanism, must be unique.

  • value (xr.DataArray) – Array of nominal+1 standard uncertainty of new mechanism. If value has the same size as the nominal data, the function will insert a new dimension at axis zero before concatenating to the covariance data.

  • dof (float,optional) – Degrees of freedom associated with the uncertainty mechanism. Default is infinite.

  • category (dict,) – Dictionary of key-value string pairs categorizing the uncertainty mechanisms. E.g. {‘Type’:’B’,’Origin’:’Datasheet’}. The default is {‘Type’:’B’}.

  • add_uid (bool, optional) – Append a UID to name to guarantee uniqueness.

Return type:

None.

property nom

Get the nominal values of the RMEMeas object.

Raises:

Exception – if no data is store.

Returns:

Nominal values in an xr.DataArray

Return type:

xr.DataArray

property umech_id

Get the names of uncertainty mechanisms, excluding the nominal.

Returns:

List of uncertainty mechanism strings.

Return type:

List

confint(percent: float, deg: bool = False, rad: bool = False)

Generate the lower, upper confidence intervals for a fractional percent confidence.

Confidence intervals are generate using the linear sensitivity analysis data (.cov) using a student-t distribution with means of .nominal, and a scale of 1 standard uncertainty.

Parameters:

  • percent (float) – Percent confidence to calculate intervals on, centered about the nominal. For example, a value of 0.5 will calculate intervals such that 0.25 of samples are between the mean and the lower, and 0.25 of expected samples are between the mean and the upper.

  • deg (bool, optional) – If you are calculated confidence intervals on degrees. The default is False.

  • rad (bool, optional) – If you are calculating confidence intervals on radians. The default is False.

Returns:

  • lower (xr.DataArray) – Lower bounds on the confidence interval.

  • upper (xr.DataArray) – Upper bounds on the confidence interval.

dof(deg: bool = False, rad: bool = False)

Get the dof of of the RMEMeas object’s covariance data.

Returns:

  • float – degrees of freedom of standard uncertainty from linear sensitivity.

  • float – degrees of freedom of standard uncertainty from monte-carlo

stdunc(k: float = 1, deg: bool = False, rad: bool = False)

Get the standard uncertainty with expansion factor k.

Supports uncertainties on angles via the deg/rad key word arguments.

Parameters:

  • k (float, optional) – Expansion factor. The default is 1.

  • deg (bool, optional) – If true, treats the values as angles and finds the minimum distance between the perturbed and un-perturbed data sets in degrees. The default is False

  • rad (bool, optional) – If true, treats the values as angles and finds the minimum distance between the perturbed and un-perturbed data sets in radians. The default is False

Returns:

  • covunc (xr.DataArray) – xr.DataArray of std unc from covariance data.

  • mcunc (xr.DataArray) – standard uncertainty of montecarlo data.

uncbounds(k: float = 1, deg: bool = False, rad: bool = False)

Get uncertainty bounds (nominal + k*stdunc).

Supports uncertainties on angles via the deg/rad key word arguments.

Parameters:

  • k (float, optional) – Expansion factor. The default is 1.

  • deg (bool, optional) – If true, treats the values as angles and finds the minimum distance between the perturbed and un-perturbed data sets in degrees. The default is False

  • rad (bool, optional) – If true, treats the values as angles and finds the minimum distance between the perturbed and un-perturbed data sets in radians. The default is False

Returns:

  • xr.DataArray – Uncertainty bounds from cov data.

  • xr.DataArray – Uncertainty bounds from montecarlo data.

assign_categories(mechanisms: list[str], categories: list[str], designation: list[str])

Assign categories to mechanisms.

Mechanisms, categories, and designation should all be the same shape.

categories and designation correspond to key,

Parameters:

  • mechanisms (list[str]) – 1-d List of mechanisms to be assigned a category

  • categories (list[str]) – 1-d list of categories being assigned. Can be new categories that don’t already exist. Index corresponds to index of mechanisms.

  • designation (list[str]) – 1-d list of designation to assign each mechanism. Index corresponds to index of mechanisms.

Return type:

None.

assign_categories_to_all(**categories: dict)

Assign categories to all the linear uncertainty mechanisms.

Parameters:

**categories (list[str]) – Keyword argument pairs of category names : designation to assign to all linear uncertainty mechanisms in a variable.

Return type:

None.

create_empty_categories(categories: list[str])

Add empty categories to the covcats array.

If an element in categories already exists, it is ignored and not added.

Parameters:

category (list[str]) – List of categories to create. If they already exist, they are not added.

Return type:

None.

group_combine_mechanisms(deg: bool = False, rad: bool = False) → RMEMeas

Group linear uncertainty mechanism originating from the same combine call.

Preserves degrees of freedom does NOT preserve categorical information, used by the self.dof() function prior to running the welch-stat equation.

Returns:

New RMEMeas object with the linear uncertainty mechanisms originating from the same combine call (Type A Analysis) grouped together.

Return type:

RMEMeas

get_unique_categories()

Get list of unique categories.

print_categories()

Print out the unique categories.

interp(coords: dict = None, method: str = 'linear', assume_sorted: bool = False, kwargs: dict = None, **coords_kwargs)

Interpolate RMEMeas object.

See xarray interp documentation for more details.

Parameters:

  • coords (dict, optional) – Mapping from dimension names to the new coordinates. New coordinate can be a scalar, array-like or DataArray. If DataArrays are passed as new coordinates, their dimensions are used for the broadcasting. Missing values are skipped. The default is None.

  • method (str, optional) –

    {“linear”, “nearest”, “zero”, “slinear”, “quadratic”, “cubic”,

    ”quintic”, “polynomial”, “pchip”, “barycentric”, “krogh”, “akima”, “makima”}) – Interpolation method. The default is ‘linear’.

  • assume_sorted (bool, optional) – If False, values of x can be in any order and they are sorted first. If True, x has to be an array of monotonically increasing values. The default is False.

  • kwargs (dict, optional) – Additional keyword arguments passed to scipy’s interpolator. Valid options and their behavior depend whether interp1d or interpn is used.. The default is None.

  • **coords_kwargs (dict like) – The keyword arguments form of coords. One of coords or coords_kwargs must be provided..

Return type:

None.

usel(umech_id: collections.abc.Iterable[str] = None, sample_id: collections.abc.Iterable[int] = None) → RMEMeas

Get a view into specific uncertainty mechanisms or Monte Carlo samples.

Parameters:

  • umech_id (iter[str], optional) – Linear uncertainty mechanisms to look at. The default is None.

  • sample_id (iter[int], optional) – Monte Carlo samples to look at. The default is None.

Raises:

ValueError – If ‘nominal’ is passed to umech_id, or 0 is passed to the sample_id. Those represent the nominal values and are always included by default, since an uncertainty object should always have a nominal value.

Returns:

View into RMEMeas object with only the selected linear uncertainty mechanisms and Monte Carlo samples.

Return type:

‘RMEMeas’

property loc

Get view into underlying data with label based indexing.

Ignores the umech_id dimension, it cannot be indexed into with this function.Index into array assuming the umech_id dimensions doesn’t exist in cov (i.e. as if indexing into only the nominal)

Generates a view on the underlying cov,mc, covcats and covdofs array. Use .copy() to create an unconnected version.

sel(indexers: dict = None, method: str = None, tolerance: float = None, **indexers_kwargs) → RMEMeas

Get view into underlying data with label based selection.

Ignores the umech_id dimension, it cannot be indexed into with this function.Index into array assuming the umech_id dimensions doesn’t exist in cov (i.e. as if indexing into only the nominal)

Generates a view on the underlying cov,mc, covcats and covdofs array. Use .copy() to create an unconnected version.

See documentation on DataArray.sel in the xarray package for details.

Parameters:

  • indexers (TYPE, optional) – A dict with keys matching dimensions and values given by scalars, slices or arrays of tick labels. For dimensions with multi-index, the indexer may also be a dict-like object with keys matching index level names. umech_id can’t be provided.

  • method (str, optional) – Method to use for inexact matches: * None (default): only exact matches * pad / ffill: propagate last valid index value forward * backfill / bfill: propagate next valid index value backward * nearest: use nearest valid index value The default is None.

  • tolerance (float, optional) – Maximum distance between original and new labels for inexact matches. The values of the index at the matching locations must satisfy the equation abs(index[indexer] - target) <= tolerance.

  • **indexers_kwargs (dict) – The keyword arguments form of indexers. One of indexers or indexers_kwargs must be provided.

Raises:

ValueError – DESCRIPTION.

Returns:

out – View into indexed RMEMeas object.

Return type:

RMEMeas

isel(indexers: dict = None, drop: bool = False, missing_dims: str = 'raise', **indexers_kwargs) → RMEMeas

Get view into underlying data with integer based selection.

Ignores the umech_id dimension, it cannot be indexed into with this function.Index into array assuming the umech_id dimensions doesn’t exist in cov (i.e. as if indexing into only the nominal)

Generates a view on the underlying cov,mc, covcats and covdofs array. Use .copy() to create an unconnected version.

Parameters:

  • indexers (dict, optional) – A dict with keys matching dimensions and values given by integers, slice objects or arrays. indexer can be a integer, slice, array-like or DataArray. If DataArrays are passed as indexers, xarray-style indexing will be carried out.One of indexers or indexers_kwargs must be provided.. The default is None.

  • drop (bool, optional) – drop coordinates variables indexed by integers instead of making them scalar. The default is False.

  • missing_dims (str, optional) – What to do if dimensions that should be selected from are not present in the DataArray: - “raise”: raise an exception - “warn”: raise a warning, and ignore the missing dimensions - “ignore”: ignore the missing dimensions. The default is ‘raise’.

  • **indexers_kwargs (TYPE) – The keyword arguments form of indexers..

Raises:

ValueError – DESCRIPTION.

Returns:

out – View into indexed RMEMeas object.

Return type:

RMEMeas

polyfit(dim: str, deg: int, apply_cov: bool = True) → RMEMeas

Apply a polynominal fit along dim.

Parameters:

  • dim (str) – Name of dimension to fit along.

  • deg (int) – Degree of fit

  • aoo

Returns:

Coefficients of data.

Return type:

RMEMeas

polyval(coord: xarray.DataArray, degree_dim: str) → RMEMeas

Evaluate a polynomial fit along coord.

Assumes self is a measurement of fit-parameters, with the dimension of polynomial fits along degree_dim.

Parameters:

  • coord (xr.DataArray) – Coordinate to fit along.

  • degree_dim (str) – Polynomial degree dimension.

Returns:

Coefficients of data.

Return type:

RMEMeas

curvefit(coords: str | xarray.DataArray | Sequence[str] | Sequence[xarray.DataArray], func: callable, reduce_dims: str | collections.abc.Iterable | None = None, skipna: bool = True, p0: dict = None, bounds: dict = None, param_names: Sequence | None = None, errors: str = 'raise', **kwargs)

Simple binding to curvefit on xarray.

The nominal value is solved for first, than used as an initial guess for the perturbed datasets.

See https://docs.xarray.dev/en/stable/generated/xarray.DataArray.curvefit.html for more details.

Parameters:

  • coords (str | xr.DataArray | Sequence[str] | Sequence[xr.DataArray], DataArray) – Independent coordinate(s) over which to perform the curve fitting. Must share at least one dimension with the calling object. When fitting multi-dimensional functions, supply coords as a sequence in the same order as arguments in func. To fit along existing dimensions of the calling object, coords can also be specified as a str or sequence of strs.

  • func (callable) – User specified function in the form f(x, *params) which returns a numpy array of length len(x). Params are the fittable parameters which are optimized by scipy curve_fit. x can also be specified as a sequence containing multiple coordinates, e.g. f((x0, x1), *params).

  • reduce_dims (str | Iterable | None, optional) – Additional dimension(s) over which to aggregate while fitting. For example, calling ds.curvefit(coords=’time’, reduce_dims=[‘lat’, ‘lon’], …) will aggregate all lat and lon points and fit the specified function along the time dimension.

  • skipna (bool, optional) – Whether to skip missing values when fitting. Default is True.

  • p0 (dict, optional) – Optional dictionary of parameter names to initial guesses passed to the curve_fit p0 arg. If the values are DataArrays, they will be appropriately broadcast to the coordinates of the array. If none or only some parameters are passed, the rest will be assigned initial values following the default scipy behavior

  • bounds (dict, optional) – Optional dictionary of parameter names to tuples of bounding values passed to the curve_fit bounds arg. If any of the bounds are DataArrays, they will be appropriately broadcast to the coordinates of the array. If none or only some parameters are passed, the rest will be unbounded following the default scipy behavior.

  • param_names (Sequence | None, optional) – Sequence of names for the fittable parameters of func. If not supplied, this will be automatically determined by arguments of func. param_names should be manually supplied when fitting a function that takes a variable number of parameters.

  • errors (str, optional) – If ‘raise’, any errors from the scipy.optimize_curve_fit optimization will raise an exception. If ‘ignore’, the coefficients and covariances for the coordinates where the fitting failed will be NaN.

  • **kwargs (optional) – Additional keyword arguments passed to scipy curve_fit.

curveval(func: callable, coords: xarray.DataArray)

Evaluate the output of curvefit.

Assumes there is a dimension called ‘param’ that coresponds. to the fit coefficients of func. Assumes function is nonlinear and iterates over each function evaluation, which can be slow.

Parameters:

  • func (callable) – callable function

  • coords (xr.DataArray) – coordinates.

exception rmellipse.uobjects.RMEMeasFormatError(message)

Bases: Exception

Error in formatting of data inside RMEMeas

Parameters:

  • Exception (_type_) – _description_

  • signature. (Initialize self. See help(type(self)) for accurate)

message