petab_select.plot

Visualization routines for model selection.

Plotting methods generally take a PlotData object as input, which can be used to specify information used by multiple plotting methods.

Module Attributes

`LABEL_FONTSIZE`	The font size of axis labels.
`TICK_LABEL_FONTSIZE`	The font size of axis tick labels.
`DEFAULT_NODE_COLOR`	The default color of nodes in graph plots.

Functions

`bar_criterion_vs_models`(plot_data[, ax, ...])	Plot all calibrated models and their criterion value.
`graph_history`(plot_data[, ...])	Plot all calibrated models in the model space, as a directed graph.
`graph_iteration_layers`(plot_data[, ax, ...])	Graph the models of each iteration of model selection.
`line_best_by_iteration`(plot_data[, ax])	Plot the improvement in criterion across iterations.
`scatter_criterion_vs_n_estimated`(plot_data)	Plot criterion values against number of estimated parameters.
`upset`(plot_data)	Plot an UpSet plot of estimated parameters and criterion.

Classes

PlotData(models[, criterion, ...])

Manage data used in plots.

petab_select.plot.DEFAULT_NODE_COLOR = 'darkgrey': The default color of nodes in graph plots.

petab_select.plot.LABEL_FONTSIZE = 16: The font size of axis labels.

class petab_select.plot.PlotData(models, criterion=None, relative_criterion=False, labels=None, parameter_labels=None, colors=None)[source]

Manage data used in plots.

models: The models.

criterion: The criterion.

labels: Model labels. Defaults to petab_select.Model.model_label or petab_select.Model.model_id or petab_select.Model.hash. Keys are model hashes, values are the labels.

relative_criterion: Whether to display relative criterion values.

parameter_labels: Labels for parameters. Keys are parameter IDs, values are labels. To use the parameterName column of your PEtab parameters table, provide petab_problem.parameter_df["parameterName"].to_dict().

colors: Colors for each model. Keys are model hashes, values are matplotlib color specifiers ( https://matplotlib.org/stable/users/explain/colors/colors.html ).

__init__(models, criterion=None, relative_criterion=False, labels=None, parameter_labels=None, colors=None)[source]

augment_labels(criterion=False, added_parameters=False, removed_parameters=False)[source]

Add information to the plotted model labels.

N.B.: this resets any previous augmentation.

Parameters:

criterion (bool) – Whether to include the criterion values.
added_parameters (bool) – Whether to include the added parameters, compared to the predecessor model.
removed_parameters (bool) – Whether to include the removed parameters, compared to the predecessor model.

Return type:

None

get_default_labels()[source]

Get default model labels.

Return type:: dict[ModelHash, str]

petab_select.plot.TICK_LABEL_FONTSIZE = 12: The font size of axis tick labels.

petab_select.plot.bar_criterion_vs_models(plot_data, ax=None, bar_kwargs=None)[source]

Plot all calibrated models and their criterion value.

Parameters:

plot_data (PlotData) – The plot data.
ax (Axes) – The axis to use for plotting.
bar_kwargs (dict[str, Any]) – Passed to the matplotlib ax.bar call.

Return type:

Axes

Returns:

The plot axes.

petab_select.plot.graph_history(plot_data, draw_networkx_kwargs=None, spring_layout_kwargs=None, ax=None)[source]

Plot all calibrated models in the model space, as a directed graph.

Parameters:

plot_data (PlotData) – The plot data.
draw_networkx_kwargs (dict[str, Any]) – Forwarded to networkx.draw_networkx.
spring_layout_kwargs (dict[str, Any]) – Forwarded to networkx.spring_layout.
ax (Axes) – The axis to use for plotting.

Return type:

Axes

Returns:

The plot axes.

petab_select.plot.graph_iteration_layers(plot_data, ax=None, draw_networkx_kwargs=None, colorbar_mappable=None)[source]

Graph the models of each iteration of model selection.

Parameters:

plot_data (PlotData) – The plot data.
ax (Axes) – The axis to use for plotting.
draw_networkx_kwargs (dict[str, Any] | None) – Passed to the networkx.draw_networkx call.
colorbar_mappable (_ScalarMappable) – Customize the colors. See documentation for the mappable argument of matplotlib.pyplot.colorbar.

Return type:

Axes

Returns:

The plot axes.

petab_select.plot.line_best_by_iteration(plot_data, ax=None)[source]

Plot the improvement in criterion across iterations.

Parameters:

plot_data (PlotData) – The plot data.
ax (Axes) – The axis to use for plotting.

Return type:

Axes

Returns:

The plot axes.

petab_select.plot.scatter_criterion_vs_n_estimated(plot_data, ax=None, scatter_kwargs=None, max_jitter=0.2)[source]

Plot criterion values against number of estimated parameters.

Parameters:

plot_data (PlotData) – The plot data.
ax (Axes) – The axis to use for plotting.
scatter_kwargs (dict[str, str]) – Forwarded to matplotlib.axes.Axes.scatter.
max_jitter (float) – Add noise to distinguish models with the same number of parameters and similar criterion values. This is a positive value that is the maximal difference to the original value.

Return type:

Axes

Returns:

The plot axes.

petab_select.plot.upset(plot_data)[source]

Plot an UpSet plot of estimated parameters and criterion.

Parameters:: plot_data (PlotData) – The plot data.
Return type:: dict[str, Axes | None]
Returns:: The plot axes (see documentation from the upsetplot package).