utils

plot_utils module

cbr_fox.utils.plot_utils.visualize_combined_pyplot(cbr_fox_instance, **kwargs)

Visualize the combined data and best cases’ components using Matplotlib.

This method generates plots that display the forecasted window, prediction, and a combined data representation for each component in the dataset. The function helps in visually analyzing how the CBR system’s predictions align with the combined data and best matching cases. Users can customize the plot appearance and behavior through Matplotlib configurations passed via kwargs.

Parameters:

cbr_fox_instance (object) –

An instance of the CBR system that contains the necessary data for plotting, including forecasted windows, predictions, and combined records.

kwargskeyword arguments

Additional arguments for customizing the plot appearance and behavior. Supported options: - ‘forecast_label’ : str, optional, default=”Forecasted Window”

The label for the forecasted window line.

• ’prediction_label’str, optional, default=”Prediction”

  The label for the prediction point.

• ’combined_label’str, optional, default=”Combined Data”

  The label for the combined data plot.

• ’combined_target_label’str, optional, default=”Combined Target”

  The label for the scatter points representing the combined target values.

• ’xlim’tuple, optional

  The limits for the x-axis (min, max).

• ’ylim’tuple, optional

  The limits for the y-axis (min, max).

• ’xtick_rotation’int, optional, default=0

  The rotation angle for x-axis tick labels.

• ’xtick_ha’str, optional, default=’right’

  Horizontal alignment of the x-axis tick labels (‘left’, ‘center’, ‘right’).

• ’title’str, optional, default=”Combined Plot {i + 1}”

  The title for the plot.

• ’xlabel’str, optional, default=”Axis X”

  The label for the x-axis.

• ’ylabel’str, optional, default=”Axis Y”

  The label for the y-axis.

• ’legend’bool, optional, default=True

  Whether to display the legend in the plot.

Returns:

A list of tuples where each tuple contains a figure and axis object for each plot generated, allowing further customization or saving.

Return type:

list of tuples

Notes

• The function generates a plot for each component in the dataset based on the number of available target training windows.

• It overlays forecasted data with combined records to facilitate a direct comparison.

• This function requires a valid cbr_fox_instance containing precomputed records.

cbr_fox.utils.plot_utils.visualize_correlation_per_window(cbr_fox_instance, **kwargs)

Plot correlation values per window using Matplotlib.

Produce a simple line plot of the CBR system’s correlation values for each window to help assess how well the forecasted window matches training windows.

Parameters:

• cbr_fox_instance (object) – CBR system instance containing an array-like attribute correlation_per_window with correlation values for each window.

• kwargs (keyword arguments) –

  Optional plotting parameters: - fig_size : tuple, optional

  Figure size in inches (width, height). Default used elsewhere is (20, 12).

  • correlation_labelstr, optional

    Label for the primary correlation line.

  • fmtstr, optional

    Matplotlib format string to pass to the second plot invocation.

  • plot_paramsdict, optional

    Additional keyword arguments forwarded to ax.plot for the second plot.

  • xlim, ylimtuple, optional

    Axis limits passed to ax.set_xlim / ax.set_ylim.

  • xtick_rotationint, optional

    Rotation angle for x-axis tick labels.

  • title, xlabel, ylabelstr, optional

    Axis and title text.

  • legendbool, optional

    Whether to enable the legend.

Returns:

(fig, ax) Matplotlib figure and axis objects for further customization or saving.

Return type:

tuple

cbr_fox.utils.plot_utils.visualize_pyplot(cbr_fox_instance, **kwargs)

Visualize the best cases’ components using Matplotlib.

This method generates multiple plots to visualize the components of the best cases found by the Case-Based Reasoning (CBR) system. It visualizes the forecasted window, the prediction, and the best matching windows with the target windows. Each plot is customized based on parameters passed through kwargs for flexibility.

Parameters:

cbr_fox_instance (object) –

An instance of the CBR system that contains the necessary data for plotting, such as the forecasted window, predictions, and best windows.

kwargskeyword arguments

Additional arguments for customizing the plot appearance and behavior. The following options are supported: - ‘forecast_label’ : str, optional, default=”Forecasted Window”

The label for the forecasted window line in the plot.

• ’prediction_label’str, optional, default=”Prediction”

  The label for the prediction point in the plot.

• ’fmt’str, optional

  The format string for plotting the best windows.

• ’plot_params’dict, optional

  Additional keyword arguments for customizing the best window plot.

• ’scatter_params’dict, optional

  Additional parameters for customizing the scatter plot for target windows.

• ’xlim’tuple, optional

  The limits for the x-axis (min, max).

• ’ylim’tuple, optional

  The limits for the y-axis (min, max).

• ’xtick_rotation’int, optional, default=0

  The rotation angle for x-axis tick labels.

• ’xtick_ha’str, optional, default=’right’

  Horizontal alignment of the x-axis tick labels (‘left’, ‘center’, ‘right’).

• ’title’str, optional, default=”Plot {i + 1}”

  The title for the plot.

• ’xlabel’str, optional, default=”Axis X”

  The label for the x-axis.

• ’ylabel’str, optional, default=”Axis Y”

  The label for the y-axis.

• ’legend’bool, optional, default=True

  Whether to display the legend in the plot.

Returns:

A list of tuples where each tuple contains a figure and axis object for each plot generated, which can be used for further customization or saving.

Return type:

list of tuples

Notes

• The function will create a plot for each component in the target training windows based on the number of components available in the data.

• This function requires a working instance of the CBR system, which holds the data for the best windows and predictions.

cbr_fox.utils.plot_utils.visualize_smoothed_correlation(cbr_fox_instance, **kwargs)

Visualize a smoothed correlation series with annotated peak and valley points. :param cbr_fox_instance:

An object that must provide the following attributes:

• smoothed_correlation1D array-like (numpy array, list, pandas Series)

  Smoothed correlation values to plot.

• peak_indexint or array-like of int

  Index or indices of peak point(s) in smoothed_correlation to mark.

• valley_indexint or array-like of int

  Index or indices of valley point(s) in smoothed_correlation to mark.

Parameters:

kwargs (dict, optional) –

Optional plotting parameters (all keys are optional):

• fig_sizetuple(float, float)

  Figure size in inches. Default: (20, 12).

• smoothed_labelstr

  Label used for the primary line plot of smoothed_correlation. Default: “Smoothed Correlation per Window”.

• fmtstr

  Matplotlib format string for the second line plot (e.g. ‘r–‘).

• plot_paramsdict

  Additional keyword arguments forwarded to ax.plot for the second plot.

• labelstr

  Label for the second plot (default falls back to smoothed_label).

• peak_paramsdict

  Keyword arguments forwarded to ax.scatter when marking peak(s).

• valley_paramsdict

  Keyword arguments forwarded to ax.scatter when marking valley(s).

Returns:

• tuple – (fig, ax) where fig is the matplotlib.figure.Figure and ax is the matplotlib.axes.Axes created by the function.

• Behavior

• ——–

• • Creates a new matplotlib figure and axis, applies fig_size if provided, – plots the smoothed correlation as a line, optionally re-plots it with a

  format string and additional plot parameters, and marks the peak and valley points using scatter.

• - Calls plt.show() before returning the (fig, ax) pair.

• • Expects that smoothed_correlation supports len() and indexing with – peak_index / valley_index. If those indices are lists/arrays, multiple

  points are marked.

Notes

• This function assumes matplotlib.pyplot is available as plt and numpy as np in the module scope.

• No validation is performed on the provided indices or arrays; callers should ensure indices are within bounds.