Structure of a Chart Class

The first line of a chart class specifies the matlab.graphics.chartcontainer.ChartContainer class as the superclass. For example, the first line of a class called ConfidenceChart looks like this:

classdef ConfidenceChart < matlab.graphics.chartcontainer.ChartContainer

In addition to specifying the superclass, include the following components in your class definition.

Implicit Constructor Method

You do not have to write a constructor method for your class because a constructor is inherited from the ChartContainer base class. The constructor accepts optional input arguments: a parent container and any number of name-value pair arguments for setting properties on the chart. For example, if you define a class called ConfidenceChart that has the public properties XData and YData, you can create an instance of your class using either of these commands:

c = ConfidenceChart(gcf,'XData',[1 2 3],'YData',[4 5 6])
c = ConfidenceChart('XData',[1 2 3],'YData',[4 5 6])

If you want to provide an interface that accepts input arguments in the same way as a typical function does, you can define a custom constructor method. See Write Constructors for Chart Classes for more information.

Public and Private Property Blocks

Divide your class properties between at least two blocks:

The properties that go in the public block store the input values provided by the user. For example, a chart that displays a line might store the x- and y-coordinate vectors in two public properties. Since the property name-value pair arguments are optional inputs to the implicit constructor method, the recommended approach is to initialize the public properties to default values. If you define public properties that store coordinate values, initializing them to NaN values or empty arrays constructs an empty chart if the user calls the constructor without any inputs.

The properties that go in the private block store the underlying graphics objects that make up your chart, in addition to any calculated values you want to store. Eventually, your class will use the data in the public properties to configure the underlying objects. By including the Transient and NonCopyable attributes for the private block, you avoid storing redundant information if the user copies or saves an instance of the chart.

For example, here are the property blocks for a chart that displays a Line object and a Patch object. The public property block stores values that the user can control: the x- and y-coordinates of the line, a confidence margin value, a marker symbol, and a color value. The private property block stores the Line and Patch objects.

properties
    XData = NaN
    YData = NaN
    ConfidenceMargin = 0.15
    MarkerSymbol = 'o'
    Color = [1 0 0]    
end
    
properties(Access = private,Transient,NonCopyable)
    LineObject
    PatchObject
end  

Setup Method

The setup method executes once when MATLAB constructs the chart object. Any property values passed as name-value pair arguments to the constructor method are assigned after this method executes.

Use the setup method to:

Many graphics functions have an optional input argument for specifying the target axes object. These functions include plotting functions (such as plot, scatter, and bar) and functions that modify the axes (such as hold, grid, and title). When you call these types of functions from within a class method, you must specify the target axes object. You can access the axes object by calling the getAxes method. This method returns the axes object, or it creates a Cartesian axes object if the chart does not already contain an axes object.

When you call plotting functions in the setup method, specify temporary values (such as NaN) for the coordinate data. Also, specify temporary values for other arguments that correspond to public properties of your class. Doing so avoids setting the same property values in both the setup and the update methods.

If you want to display multiple primitive objects in the axes, call the hold function between plotting commands. Set the hold state back to 'off' after your last plotting command.

For example, consider a chart that displays a line and a patch. It has these properties:

The setup method gets the axes object by calling the getAxes method. Then it calls the patch function and stores the output in the PatchObject property. The next line of code set the hold state of the axes to 'on' before calling the plot function to create the LineObject property. The last line of code sets the axes hold state back to 'off'.

function setup(obj)
    % Get the axes
    ax = getAxes(obj);
            
    % Create Patch and Line objects
    obj.PatchObject = patch(ax,NaN,NaN,'r','FaceAlpha',0.2,...
        'EdgeColor','none');
    hold(ax,'on')
    obj.LineObject = plot(ax,NaN,NaN);
            
    % Turn hold state off
    hold(ax,'off')
end

Update Method

When the user changes one or more property values on the chart object, MATLAB marks the chart object for updating. The update method runs for the first time after the setup method runs. Then it runs the next time drawnow executes. The drawnow function automatically executes periodically, based on the state of the graphics environment in the user's MATLAB session. Thus, there might be a delay between changing property values and seeing the results of those changes.

Use the update method to reconfigure the underlying graphics objects in your chart based on the new values of the properties. Typically, this method does not distinguish which of the properties changed. It reconfigures all aspects of the underlying graphics objects that depend on the properties.

For example, consider a chart that has these properties:

The update method updates the XData and Color properties of the Line and Patch objects.

function update(obj)
            
    % Update XData of line object
    obj.LineObject.XData = obj.XData;
            
    % Update patch XData 
    x = obj.XData;
    obj.PatchObject.XData = [x x(end:-1:1)];
            
    % Update line object colors
    obj.LineObject.Color = obj.Color;
    obj.PatchObject.FaceColor = obj.Color;
           
end

Example: Confidence Bounds Chart

This example shows how to create a chart for plotting a line with confidence bounds. Create a class definition file named ConfidenceChart.m in a folder that is on the MATLAB path. Define the class by following these steps.

classdef ConfidenceChart < matlab.graphics.chartcontainer.ChartContainer

    properties
        XData = NaN
        YData = NaN
        ConfidenceMargin = 0.15
        MarkerSymbol = 'o'
        Color = [1 0 0]
    end

    properties(Access = private,Transient,NonCopyable)
        LineObject
        PatchObject
    end

    methods(Access = protected)
        function setup(obj)
            % get the axes
            ax = getAxes(obj);

            % Create Patch and Line objects
            obj.PatchObject = patch(ax,NaN,NaN,'r','FaceAlpha',0.2,...
                'EdgeColor','none');
            hold(ax,'on')
            obj.LineObject = plot(ax,NaN,NaN);

            % Turn hold state off
            hold(ax,'off')
        end

        function update(obj)
            % Update XData and YData of Line
            obj.LineObject.XData = obj.XData;
            obj.LineObject.YData = obj.YData;

            % Update patch XData and YData
            x = obj.XData;
            obj.PatchObject.XData = [x x(end:-1:1)];
            y = obj.YData;
            c = obj.ConfidenceMargin;
            obj.PatchObject.YData = [y+c y(end:-1:1)-c];

            % Update colors
            obj.LineObject.Color = obj.Color;
            obj.PatchObject.FaceColor = obj.Color;

            % Update markers
            obj.LineObject.Marker = obj.MarkerSymbol;
        end
    end
end

Next, create an instance of the chart by calling the implicit constructor method with a few of the public properties:

x = 0:0.2:10;
y = besselj(0,x);
c = ConfidenceChart('XData',x,'YData',y,'ConfidenceMargin',0.15);

Change the color.

c.Color = [0 0 1];

Support Common Graphics Features

By default, instances of your charts support much of the functionality that is common to all MATLAB charts. For example, the gca and findobj functions can return instances of your chart. You can also pass instances of your chart to the set and get functions, and you can configure the properties of the chart in the Property Inspector.

The features described in this table are supported only if you enable them for your chart.