fitpy.dataset module

Datasets: units containing data and metadata.

The dataset is one key concept of the ASpecD framework and hence the FitPy package derived from it, consisting of the data as well as the corresponding metadata. Storing metadata in a structured way is a prerequisite for a semantic understanding within the routines. Furthermore, a history of every processing, analysis and annotation step is recorded as well, aiming at a maximum of reproducibility. This is part of how the ASpecD framework and therefore the FitPy package tries to support good scientific practice.

Therefore, each processing and analysis step of data should always be performed using the respective methods of a dataset, at least as long as it can be performed on a single dataset.

The reason for providing an own class for calculated datasets is to ensure a consistent handling of fit results and provide means for specialised plotters available in the fitpy.plotting module to check for their applicability.

class fitpy.dataset.CalculatedDataset[source]

Bases: CalculatedDataset

Dataset containing results of fitting a model to data.

data

numeric data, residual, and axes

In contrast to other datasets, it contains the residual (difference between fitted model and original data) as well.

Type:: Data

metadata

hierarchical key-value store of metadata

Type:: CalculatedDatasetMetadata

class fitpy.dataset.CalculatedDatasetLHS[source]

Bases: CalculatedDataset

Dataset containing results of fitting a model to data.

data

numeric data, residual, and axes

In contrast to other datasets, it contains the residual (difference between fitted model and original data) as well.

Type:: Data

metadata

hierarchical key-value store of metadata

Type:: CalculatedDatasetLHSMetadata

class fitpy.dataset.Data[source]

Bases: Data

Unit containing numeric data, residual, and corresponding axes.

In contrast to the base class of the ASpecD framework, it contains the residual (difference between fitted model and original data) as well.

property residual

Residual (difference between fitted model and original data).

A residual need always to have the same shape as the corresponding data. If you try to set a residual not conforming to this condition, a ValueError will be raised.

class fitpy.dataset.CalculatedDatasetMetadata[source]

Bases: CalculatedDatasetMetadata

Metadata for a dataset with calculated data.

This class contains the metadata for a dataset consisting of calculated data, i.e., CalculatedDataset.

Metadata can be converted to dict via aspecd.utils.ToDictMixin.to_dict(), e.g., for generating reports using templates and template engines.

calculation

Information on the calculation.

Contain, inter alia, the parameters of the calculation.

Type:: aspecd.metadata.Calculation

model

Details of the model fitted to the data

Type:: Model

data

Details of the data the model has been fitted to

Type:: class:DataMetadata

result

Summary of results of fit

Type:: Result

class fitpy.dataset.CalculatedDatasetLHSMetadata[source]

Bases: CalculatedDatasetMetadata

Metadata for a dataset with calculated data.

This class contains the metadata for a dataset consisting of calculated data, i.e., CalculatedDataset.

Metadata can be converted to dict via aspecd.utils.ToDictMixin.to_dict(), e.g., for generating reports using templates and template engines.

calculation

Information on the calculation.

Contain, inter alia, the parameters of the calculation.

Type:: aspecd.metadata.Calculation

model

Details of the model fitted to the data

Type:: Model

data

Details of the data the model has been fitted to

Type:: class:DataMetadata

result

Summary of results of fit

Type:: Result

lhs

Details of the LHS and its full results for each sampling

Type:: LHS

class fitpy.dataset.Model[source]

Bases: Metadata

Metadata of a model fitted to data of a dataset.

Part of the metadata of a CalculatedDataset containing the data of the model fitted to the data of another (experimental) dataset.

type

Full class name (including package) of the respective model class

Type:: str

parameters

All parameters of the model

Type:: dict

from_model(model)[source]

Set attributes from aspecd.model.Model.

Parameters:: model (aspecd.model.Model) – Model the attributes should be obtained from

class fitpy.dataset.DataMetadata[source]

Bases: Metadata

Metadata of the data(set) a model has been fitted to.

Part of the metadata of a CalculatedDataset containing information of the (experimental) dataset the model has been fitted to.

id

(unique) identifier of the dataset (i.e., path, LOI, or else)

Type:: str

label

Short description of the dataset

Can be set by the user, defaults to the value set as aspecd.dataset.Dataset.id by the importer.

Type:: str

from_dataset(dataset)[source]

Set attributes from aspecd.dataset.Dataset.

Parameters:: dataset (aspecd.dataset.Dataset) – Dataset the attributes should be obtained from

class fitpy.dataset.Result[source]

Bases: Metadata

Metadata of results of fitting a model to data of a dataset.

Part of the metadata of a CalculatedDataset containing the data of the model fitted to the data of another (experimental) dataset.

While resembling the structure of the lmfit.minimizer.MinimizerResult class, this class tries to abstract away from the attributes in terms of their names and introduces more readable (and more lengthily) attribute names.

parameters

The best-fit parameters resulting from the fit.

Type:: lmfit.parameter.Parameters

success

True if the fit succeeded, otherwise False.

Type:: bool

error_bars

True if uncertainties were estimated, otherwise False.

Type:: bool

n_function_evaluations

Number of function evaluations

Type:: int

n_variables

Number of variables of the model

Type:: int

degrees_of_freedom

Degrees of freedom

Type:: int

chi_square

Chi-square value

For this value to be meaningful, the residual function needs to be scaled properly to the uncertainties in the data.

Type:: float

reduced_chi_square

Reduced chi-square value

For this value to be meaningful, the residual function needs to be scaled properly to the uncertainties in the data.

Type:: float

akaike_information_criterion

Akaike Information Criterion statistic

Type:: float

bayesian_information_criterion

Bayesian Information Criterion statistic

Type:: float

variable_names

Ordered list of variable parameter names used in the optimisation.

Type:: list

covariance_matrix

Covariance matrix from minimisation.

Rows and columns correspond to variable_names

Type:: numpy.ndarray

initial_values

List of initial values for variable parameters.

For the corresponding parameter names see variable_names.

Type:: list

message

Message regarding the fit success.

Type:: str

to_dict(remove_empty=False)[source]

Create dictionary containing public attributes of an object.

Parameters:

remove_empty (bool) –

Whether to remove keys with empty values

Default: False

Returns:

public_attributes – Ordered dictionary containing the public attributes of the object

The order of attribute definition is preserved

Return type:

collections.OrderedDict

from_lmfit_minimizer_result(result)[source]

Set attributes from lmfit.minimizer.MinimizerResult.

Parameters:: result (lmfit.minimizer.MinimizerResult) – Result of a minimisation using lmfit

class fitpy.dataset.LHS[source]

Bases: Metadata

Metadata of the LHS and its full results for each sampling.

Part of the metadata of a CalculatedDatasetLHS containing the data of the model fitted to the data of another (experimental) dataset.

samples

Grid of the Latin Hypercube

Type:: numpy.array

discrepancy

Discrepancy of the sample.

The discrepancy is a uniformity criterion used to assess the space filling of a number of samples in a hypercube. A discrepancy quantifies the distance between the continuous uniform distribution on a hypercube and the discrete uniform distribution on distinct sample points. (from scipy.stats.qmc.discrepancy())

Type:: float

results

Results for each sample of the Latin Hypercube.

Each result is an instance of Result.

Type:: list

from_lmfit_minimizer_results(results)[source]

Set attributes from lmfit.minimizer.MinimizerResult.

Parameters:

results (list) –

List of results of a minimisation using lmfit

Each result is an instance of lmfit.minimizer.MinimizerResult and gets transferred to an instance of Result.