src.plots.vvuq_plots module

This module is responsible for the plotting the results from the VV&UQ methodology.

It includes several functions that plot different aspects. See details in their own documentations. Generally, a index dictionary can be used to select only a subset of the data arrays that shall be plotted.

Examples:: idx_dict = {‘quantities’: ‘Car.ax’, ‘space_samples’: 1, ‘timesteps’: slice(1000, 3000)} idx_dict = {‘quantities’: ‘Car.ax’, ‘timesteps’: slice(1000, 3000)} idx_dict = {‘quantities’: ‘Car.ax’, ‘space_samples’: 1} idx_dict = {‘quantities’: ‘Car.ax’}

Contact person: Stefan Riedmaier Creation date: 20.08.2020 Python version: 3.8

src.plots.vvuq_plots.plot_area_metrics(config, cfgpl, pbox_y_model, pbox_y_system, pbox_x_model_list, pbox_x_system_list, metric_da, scenarios_da, save_path='')

This function plots the area between CDFs or pboxes (see area validation metrics).

Parameters:

config (dict) – user configuration
cfgpl (dict) – user configuration of this plot
pbox_y_model (np.ndarray) – y vector of the model
pbox_y_system (np.ndarray) – y vector of the system
pbox_x_model_list (list[np.ndarray, np.ndarray]) – list with left and right x vector of the model
pbox_x_system_list (list[np.ndarray, np.ndarray]) – list with left and right x vector of the system
metric_da (xr.DataArray) – array with validation metric results
scenarios_da (xr.DataArray) – array with scenarios
save_path (str) – path where the plots shall be saved (otherwise they will be shown)

Returns:

src.plots.vvuq_plots.plot_cdf(config, cfgpl, qois_kpi_raw_da, fill_flag=False, qois_kpi_da=None, save_path='')

This function plots an Empirical (step function) Cumulative Distribution Function (ECDF).

Parameters:

config (dict) – user configuration
cfgpl (dict) – user configuration of this plot
qois_kpi_raw_da (xr.DataArray) – array with KPIs, in the raw form without CDFs
fill_flag (bool) – flag to fill the area between CDFs
qois_kpi_da (xr.DataArray) – array with KPIs, in the form of a pbox
save_path (str) – path where the plots shall be saved (otherwise they will be shown)

Returns:

src.plots.vvuq_plots.plot_decision_space(config, cfgpl, scenarios_application_da=<xarray.DataArray ()> array(None, dtype=object), decision_application_system_estimated_da=<xarray.DataArray ()> array(None, dtype=object), decision_application_system_da=<xarray.DataArray ()> array(None, dtype=object), decision_application_model_da=<xarray.DataArray ()> array(None, dtype=object), scenarios_validation_da=<xarray.DataArray ()> array(None, dtype=object), decision_validation_da=<xarray.DataArray ()> array(None, dtype=object), decision_validation_system_da=<xarray.DataArray ()> array(None, dtype=object), decision_validation_model_da=<xarray.DataArray ()> array(None, dtype=object), save_path='')

This function plots the binary decisions across the scenario space by coloring the points.

The idx_dict must contain a ‘qois’ and a ‘parameters’ key. The ‘qois’-key must contain exactly one string value: qoi name from the decision arrays. The ‘parameters’-key must contain a list of either two or three string values: parameter names from scenario arrays. Depending on the ‘parameters’-key, a two or three dimensional scatter plot will be created. The order of the parameter names in the list determines the x, y and z axes in the plot. The idx_dict can contain a ‘space_samples’ key to index only a subset of the whole data, e.g. slice(None, None, 10). This is important for high-dimensional spaces to only visualize one layer in the 2D scatter plot.

The function arguments are not fully optional. Please provide either application or validation data, so that something can be plotted. Please provide the scenario and decision data together, since they cannot be plotted standalone.

The decision_system_da and decision_model_da arguments are fully optional. However, please provide not both at the same time.

Parameters:

config (dict) – user configuration
cfgpl (dict) – user configuration of this plot
scenarios_application_da (xr.DataArray) – (optional) application scenario array
decision_application_system_estimated_da (xr.DataArray) – (optional) decisions of the system estimator in appl.
decision_application_system_da (xr.DataArray) – (optional) decisions of the actual system in the appl. domain
decision_application_model_da (xr.DataArray) – (optional) decisions of the model in the application domain
scenarios_validation_da (xr.DataArray) – (optional) validation scenario array
decision_validation_da (xr.DataArray) – (optional) model accuracy decisions in the validation domain
decision_validation_system_da (xr.DataArray) – (optional) safety decisions of the system in validation domain
decision_validation_model_da (xr.DataArray) – (optional) safety decisions of the model in the validation domain
save_path (str) – (optional) path where the plots shall be saved (otherwise they will be shown)

src.plots.vvuq_plots.plot_error_integration_deterministic(config, cfgpl, qois_kpi_model_da, qois_kpi_system_estimated_da, scenarios_model_da, qois_kpi_system_da=<xarray.DataArray ()> array(None, dtype=object), save_path='')

This function plots the deterministic model kpis with uncertainty, the true system kpis and the regulation.

Parameters:

config (dict) – user configuration
cfgpl (dict) – user configuration of this plot
qois_kpi_model_da (xr.DataArray) – array with model kpis in the application domain
qois_kpi_system_estimated_da (xr.DataArray) – array with estimated system values in the application domain
scenarios_model_da (xr.DataArray) – array with application scenarios
qois_kpi_system_da (xr.DataArray) – (optional) array with the true system kpis in the application domain
save_path (str) – (optional) path where the plots shall be saved (otherwise they will be shown)

Returns:

src.plots.vvuq_plots.plot_extrapolation_surface(config, cfgpl, scenarios_validation_da, metric_validation_da, scenarios_application_da, error_validation_da, save_path='')

This function plots the response surface of the error model including prediction interval and the metric results.

The idx_dict must contain a ‘qois’ and a ‘parameters’ key. The ‘qois’-key must contain exactly one string value: qoi name from the decision arrays. The ‘parameters’-key must contain a list of two string values: parameter names from scenario arrays. The order of the parameter names in the list determines the x, y and z axes in the plot. The idx_dict can contain a ‘space_samples’ key to index only a subset of the whole data, e.g. slice(None, None, 10). This is important for high-dimensional spaces to only visualize one layer in the 2D scatter plot.

In case of non-deterministic data, the idx_dict should contain an ‘interval’-key. Then, either the left or the right metric results are plotted (scatter points) and three surfaces for the lower and upper prediction interval and the regression estimate. In case of deterministic data, the idx_dict should not contain an ‘interval’-key. Then, the metric results are plotted (scatter points) and the error estimation with both bounds (three surfaces).

Parameters:

config (dict) – user configuration
cfgpl (dict) – user configuration of this plot
scenarios_validation_da (xr.DataArray) – array with validation scenarios
metric_validation_da (xr.DataArray) – array with validation metric results
scenarios_application_da (xr.DataArray) – array with application scenarios
error_validation_da (xr.DataArray) – array with inferred errors across the application domain
save_path (str) – (optional) path where the plots shall be saved (otherwise they will be shown)

src.plots.vvuq_plots.plot_kpi_surface(config, cfgpl, scenarios_da, qois_kpi_raw_da, plot_type='', save_path='')

This function plots a response surface of KPIs across the scenario space.

The idx_dict must contain a ‘qois’ and a ‘parameters’ key. The ‘qois’-key must contain exactly one string value: qoi name from the decision arrays. The ‘parameters’-key must contain a list of either two or three string values: parameter names from scenario arrays. The order of the parameter names in the list determines the x and y (and z) axes in the plot. The idx_dict can contain a ‘space_samples’ key to index only a subset of the whole data, e.g. slice(None, None, 10). This is important for high-dimensional spaces to only visualize one layer in the 2D scatter plot. The idx_dict can contain the dimensions ‘repetitions’, ‘aleatory_samples’ or ‘epistemic_samples’ to index a subset.

It creates a surface plot if the selected data is one-dimensional. Otherwise it creates a scatter plot. Depending on the ‘parameters’-key, the scatter plot is two or three dimensional.

Parameters:

config (dict) – user configuration
cfgpl (dict) – user configuration of this plot
scenarios_da (xr.DataArray) – array with scenarios
qois_kpi_raw_da (xr.DataArray) – array with KPIs
plot_type (str) – (optional) type of the plot: surface, scatter or stem (depending on dimensionality!)
save_path (str) – (optional) path where the plots shall be saved (otherwise they will be shown)

src.plots.vvuq_plots.plot_scenario_space(config, cfgpl, scenarios_verification_da=<xarray.DataArray ()> array(None, dtype=object), scenarios_validation_model_da=<xarray.DataArray ()> array(None, dtype=object), scenarios_validation_system_da=<xarray.DataArray ()> array(None, dtype=object), scenarios_application_da=<xarray.DataArray ()> array(None, dtype=object), scenarios_space_validation_da=<xarray.DataArray ()> array(None, dtype=object), scenarios_space_application_da=<xarray.DataArray ()> array(None, dtype=object), save_path='')

This function plots scatter points over the coordinates given by scenario points of multiple domains.

The idx_dict must contain a ‘parameters’ key. The ‘parameters’-key must contain a list of either two or three string values: parameter names from scenario arrays. Depending on the ‘parameters’-key, a two or three dimensional scatter plot will be created. The order of the parameter names in the list determines the x, y and z axes in the plot.

The function arguments are not fully optional. Please provide at least one array, so that something can be plotted.

Parameters:

config (dict) – user configuration
cfgpl (dict) – user configuration of this plot
scenarios_verification_da (xr.DataArray) – (optional) verification scenario array
scenarios_validation_model_da (xr.DataArray) – (optional) validation scenario array of the simulator
scenarios_validation_system_da (xr.DataArray) – (optional) validation scenario array of the experiment
scenarios_application_da (xr.DataArray) – (optional) application scenario array
scenarios_space_validation_da (xr.DataArray) – (optional) nominal validation scenario points
scenarios_space_application_da (xr.DataArray) – (optional) nominal appl. scenario points
save_path (str) – (optional) path where the plots shall be saved (otherwise they will be shown)

src.plots.vvuq_plots.plot_timeseries(config, cfgpl, qois_ts_da, qois_kpi_da=<xarray.DataArray ()> array(None, dtype=object), save_path='')

This function plots possibly multiple time series.

The idx_dict should contain the qoi to be plotted. It can include a timesteps key to slice the time dimension.

Parameters:

config (dict) – user configuration
cfgpl (dict) – user configuration of this plot
qois_ts_da (xr.DataArray) – array of qoi time series
qois_kpi_da (xr.DataArray) – (optional) array of model kpis
save_path (str) – (optional) path where the plots shall be saved (otherwise they will be shown)

src.plots.vvuq_plots.plot_timeseries_unecer79(cfgpl, qois_ts_da, quantities_ts_da, qois_ts_untrimmed_da, qois_kpi_da=<xarray.DataArray ()> array(None, dtype=object), start_idx=None, stop_idx=None, ay_lower_bound=None, ay_upper_bound=None, save_path='')

This function plots a specific time series of lateral acceleration and distance to line for the UNECE-R79.

In the upper subplot, it visualizes the check against the stationary ay-conditions to extract the valid events. In the lower subplot, it visualizes the actual distance to line qoi signal during the valid event. The idx_dict should not contain the actual qoi.

If the start and stop indices are provided, it visualizes vertical lines as event borders. If the ay bounds are provided, it visualizes horizontal bounds as stationary conditions.

If the KPI is provided, it will be highlighted as point in the time series of the distance to line.

Parameters:

cfgpl (dict) – user configuration of this plot
qois_ts_da (xr.DataArray) – array of qoi time series
quantities_ts_da (xr.DataArray) – array of quantity time series
qois_ts_untrimmed_da (xr.DataArray) – array of untrimmed qoi time series
start_idx (tuple(np.ndarray)) – (optional) tuple of integer index arrays with the event start indices
stop_idx (tuple(np.ndarray)) – (optional) tuple of integer index arrays with the event stop indices
ay_lower_bound (np.ndarray) – (optional) lower ay bound to be plotted as horizontal line
ay_upper_bound (np.ndarray) – (optional) upper ay bound to be plotted as horizontal line
qois_kpi_da (xr.DataArray) – (optional) array of model kpis
save_path (str) – (optional) path where the plots shall be saved (otherwise they will be shown)

src.plots.vvuq_plots.plot_uncertainty_expansion_nondeterministic(config, cfgpl, qois_kpi_model_da, qois_kpi_system_estimated_da, qois_kpi_system_estimated_validation_da, scenarios_model_da, qois_kpi_system_da=<xarray.DataArray ()> array(None, dtype=object), save_path='')

This function plots the three non-deterministic model uncertainties, the true system uncertainty and the regulation.

Parameters:

config (dict) – user configuration
cfgpl (dict) – user configuration of this plot
qois_kpi_model_da (xr.DataArray) – array with model kpis in the application domain
qois_kpi_system_estimated_da (xr.DataArray) – array with estimated system responses
qois_kpi_system_estimated_validation_da (xr.DataArray) – array with estimated system responses, only validation
scenarios_model_da (xr.DataArray) – array with application scenarios
qois_kpi_system_da (xr.DataArray) – (optional) array with the true system kpis in the application domain
save_path (str) – (optional) path where the plots shall be saved (otherwise they will be shown)

Returns: