Save and Load Part of Diagram Data to/from JSON Files

In this blog post we will demonstrate how you can implement methods that save and load part of the information that a diagram renders into JSON format. The Diagram class has built-in toJson and fromJson methods. They serialize completely the diagram with every detail so you can restore it completely, exactly the way you see it. What we are going to do is write custom save and load methods that use only part of the information – namely the size and location of nodes and links as well as their text.

Continue reading

One-way Links in Diagramming for JavaScript

In this blog post we demonstrate how you can use the JavaScript Diagram library to allow users to draw an org chart or one-way graph. By default, users are allowed to draw links between any two nodes and the count of links is unlimited. We will use event and properties, available in JS Diagram to allow users to:
– draw only one link between two nodes
– draw links only in one direction.

This behavior mirrors the hierarchy of an organization: in general , each employee, represented by a node, should have only one direct boss.


Continue reading

JavaScript Diagram with Chart Nodes

In this blog post we will create DiagramNode instances that render a chart. The charts get resized according to the size of the node. We will use two MindFusion libraries for JavaScript – Charting and Diagramming. You can run the sample online from this link:

I. Project Setup

We will use a web page that will hold the HTML Canvas element that we need in order to render the diagram:

<div style="width: 100%; height: 100%; overflow: scroll;">
    <canvas id="diagram" width="2100" height="2100" style="background:#f0f0f0;">
        This page requires a browser that supports HTML 5 Canvas element.
    </canvas>
</div>

We give the diagram’s Canavs big width and height and we also provide an id. The id is important because we will need to access the Canvas from code.

We add the scripts that we need in order to use the Charting and Diagramming libraries:

<script src="Scripts/MindFusion.Common.js" type="text/javascript"></script>
<script src="Scripts/MindFusion.Diagramming.js" type="text/javascript"></script>
<script src="Scripts/MindFusion.Charting.js" type="text/javascript"></script>

We add the references at the end of the web page, just before the closing BODY tag. We also need to add a reference to a JavaScript file that will hold the code for our sample. We name it “PieNode.js”.

II. The Diagram

In the code-behind file we create a diagram instance. We use the DOMContentLoaded event to initialize our diagram:

document.addEventListener("DOMContentLoaded", function ()
{
// create a Diagram component that wraps the "diagram" canvas
diagram = MindFusion.Diagramming.Diagram.create(document.getElementById("diagram"));
diagram.setBehavior(MindFusion.Diagramming.Behavior.LinkShapes);
diagram.setLinkHeadShapeSize(2);
diagram.setBounds(new Rect(0, 0, 2000, 2000));

// draw a pie when the user creates a node
diagram.addEventListener(Events.nodeCreated, onNodeCreated);

});

The Behavior enumeration lists various modes of behavior for the diagram. We choose Behavior.LinkShapes, which creates nodes, when the mouse is dragged
over an empty area and connects the nodes if the mouse is dragged from an existing node. We set tbe bounds of the diagram to a big Rect which guarantees that the user can draw nodes anywhere on the visible area. When the user draws towards the edge of the browser, the diagram control expands automatically.

Finally, we add an event handler for the nodeCreated event.

//nodeCreated event handler
function onNodeCreated(sender, args)
{
	var node = args.getNode();
	
	var nBounds = node.getBounds ();
	var topLeftCoord = diagram.docToClient(nBounds.topLeft());
	var bottomRightCoord = diagram.docToClient(nBounds.bottomRight());
	.......................................................
}

At first we get the node that was created. Then we need to get its actual size, for which we use the docToClient method that converts between diagram and browser measure units. We get the two coordinates of the node’s bounding rectangle.

We create then a Canvas, which takes the size of the node’s rectangle:

............................................
.....................................................
var pieChartCanvas = document.createElement('canvas');
pieChartCanvas.width = bottomRightCoord.x - topLeftCoord.x;
pieChartCanvas.height = bottomRightCoord.y - topLeftCoord.y;

We then add this temp canvas to the body of the web page and call a method “createPie” where we draw the pie chart. Once the chart is drawn we get the URL of the image and set is to the node with the setImageLocation method. We remove the Canvas from the tree with the elements because we want to use for the next node.

//create a temporary Canvas for the pie chart
//to draw itself upon
document.body.appendChild(pieChartCanvas);
createPie(pieChartCanvas);
var pieImageLocation = pieChartCanvas.toDataURL();
node.setImageLocation(pieImageLocation);
document.body.removeChild(pieChartCanvas);

III. The Chart

We create a pie chart in the Canvas provided to the createPie method:

//draw a pie chart on the provided canvas
function createPie(pieChartCanvas)
{
	var pieChart = new Controls.PieChart(pieChartCanvas);
	pieChart.startAngle = 45;			
	pieChart.showLegend = false;
	pieChart.title = "Sales";
	pieChart.titleMargin = new Charting.Margins(0, 10, 0, 0);
	pieChart.chartPadding = 3;
...........................

We set some appearance properties to make the chart look the way we want – change the start angle of the pie, add a title and title margin, hide the legend. Then we create a PieSeries which holds the data and the labels of the pie. We assign the PieSeries to the series property of the pie chart:

// create sample data
var values = new Collections.List([20.00, 30.00, 15.00, 40.00]);
pieChart.series = new Charting.PieSeries(
	values,
	new Collections.List(["20", "30", "15", "40"]),
null);

We style the chart with an instance of the PerElementSeriesStyle class, which colors all elements of a Series with the consequtive Brush and stroke from its brushes and strokes collections. Then we adjust the dataLabelsFontSize to match the size of the canvas. We make the labels be drawn with a white brush using the dataLabelsBrush property. Finally we call draw to render the pie.

var brushes = new Collections.List(
		[
			new Drawing.Brush("#0455BF"),
			new Drawing.Brush("#033E8C"),
			new Drawing.Brush("#F24405"),
			new Drawing.Brush("#F21905")
		]);

	var seriesBrushes = new Collections.List();
	seriesBrushes.add(brushes);

	var strokes = new Collections.List(
		[
			new Drawing.Brush("#c0c0c0")
		]);

	var seriesStrokes = new Collections.List();
	seriesStrokes.add(strokes);

	pieChart.plot.seriesStyle = new Charting.PerElementSeriesStyle(seriesBrushes, seriesStrokes);
	pieChart.theme.highlightStroke = new Drawing.Brush("#000063");
	pieChart.theme.dataLabelsFontSize = pieChartCanvas.height/20;
	pieChart.theme.dataLabelsBrush = new Drawing.Brush("#FFFFFF");

	pieChart.draw();

And with that the sample is ready. You can download the source code together with all MindFusion JavaScript libraries used from:

https://mindfusion.eu/samples/javascript/diagram/PieNodes.zip

About Diagramming for JavaScript: This native JavaScript library provides developers with the ability to create and customize any type of diagram, decision tree, flowchart, class hierarchy, graph, genealogy tree and more. The control offers rich event set, numerous customization options, animations, graph operations, styling and themes. You have more than 100 predefined nodes, table nodes and more than 15 automatic layout algorithms. Learn more about Diagramming for JavaScript at https://mindfusion.eu/javascript-diagram.html.

About Charting for JavaScript: MindFusion library for interactive charts and gauges. It supports all common chart types including 3D bar charts. Charts can have a grid, a legend, unlimited number of axes and series. Scroll, zoom and pan are supported out of the box. You can easily create your own chart series by implementing the Series interface.
The gauges library is part of Charting for JavaScript. It supports oval and linear gauge with several types of labels and ticks. Various samples show you how the implement the gauges to create and customize all popular gauge types: car dashboard, clock, thermometer, compass etc. Learn more about Charting and Gauges for JavaScript at https://mindfusion.eu/javascript-chart.html.

Creating custom CompositeNode components

In this post we’ll examine how CompositeNode components work in MindFusion.Diagramming for Windows Forms, and in the process create a custom radio button component. You can find the completed sample project here: RadioComponent.zip

CompositeNode was created as alternative of the ControlNode class, which lets you present any Windows Forms control as a diagram node. ControlNode has many advantages, such as letting you design the hosted user controls using Visual Studio designer, reusing them in other parts of the user interface, and including complex framework or third-party controls as their children. From the fact that each user control creates a hierarchy of Win32 windows come some disadvantages too:

  • ControlNodes cannot mix with other diagram elements in the Z order but are always drawn on top
  • performance deteriorates if showing hundreds of nodes
  • mouse events might not reach the diagram if hosted controls capture mouse input
  • print and export might not be able to reproduce the appearance of hosted controls without additional work (handling PaintControl event)

On the other hand, CompositeNode does all its drawing in DiagramView control’s canvas and is not affected by the issues listed above. CompositeNode lets you build node’s UI by composing hierarchy of components derived from ComponentBase class. Pre-defined components include layout panels, read-only or editable text fields, images, borders, buttons, check-boxes and sliders. If the UI component you need isn’t provided out of the box, you could still implement it as a custom class that derives from ComponentBase or more specific type and overriding the GetDesiredSize, ArrangeComponents and Draw methods. Lets see how that works using a RadioButtonComponent as an example.

Derive RadioButtonComponent from CheckBoxComponent so we reuse its IsChecked and Content properties:

class RadioButtonComponent : CheckBoxComponent
{
}

CompositeNode relies on a dynamic layout system that lets components determine their size by overriding GetDesiredSize method, and arranging children in allocated size by means of ArrangeComponents method. For radio button we’ll call its base class to measure content size and add enough space for drawing the radio graphics element (a circle) horizontally, while fitting it in measured height:

float RadioSize(SizeF size)
{
	return Math.Min(size.Width, size.Height);
}

public override SizeF GetDesiredSize(SizeF availableSize, IGraphics graphics)
{
	var s = base.GetDesiredSize(availableSize, graphics);
	s.Width += RadioSize(s);
	return s;
}

ArrangeComponents calls the base class to arrange its content on the right side of available space:

public override void ArrangeComponents(RectangleF availableSpace, IGraphics graphics)
{
	var radioSize = RadioSize(availableSpace.Size);
	availableSpace.X += radioSize;
	availableSpace.Width -= radioSize;
	base.ArrangeComponents(availableSpace, graphics);
}

Now override Draw and render standard radio button graphics on the left side of the component, and content on the right side:

public override void Draw(IGraphics graphics, RenderOptions options)
{
	var radioSize = RadioSize(Bounds.Size);
	var r = radioSize / 2 - 1;
	var cr = r - 1;

	graphics.FillEllipse(Brushes.White, Bounds.X + 1, Bounds.Y + 1, 2 * r, 2 * r);
	using (var pen = new System.Drawing.Pen(Color.Black, 0.1f))
		graphics.DrawEllipse(pen, Bounds.X + 1, Bounds.Y + 1, 2 * r, 2 * r);
	if (IsChecked)
		graphics.FillEllipse(Brushes.Black, Bounds.X + 2, Bounds.Y + 2, 2 * cr, 2 * cr);

	GraphicsState s = graphics.Save();
	graphics.TranslateTransform(radioSize - 1 + Bounds.X, Bounds.Y);
	Content.Draw(graphics, options);
	graphics.Restore(s);
}

We’ll want only one radio from a group to be selected. For our purposes we can count all radio buttons placed inside same stack panel as part of same group. Override the OnClick method to unselect all buttons in parent panel and select the clicked one:

protected override void OnClicked(EventArgs e)
{
	var parentStack = Parent as StackPanel;
	if (parentStack != null)
	{
		foreach (var child in parentStack.Components)
		{
			var radio = child as RadioButtonComponent;
			if (radio != null)
				radio.IsChecked = false;
		}
	}
	this.IsChecked = true;
}

That’s it, the radio button component is ready with just a screenful of code 🙂 Let’s check how it works by creating an OptionNode class that shows a group of radio buttons and exposes a property to access or change selected one:

class OptionNode : CompositeNode
{
}

You could create the stack panel and radio buttons from code if you need more dynamic configuration, e.g. one with variable number of radio buttons. For this example we’ll just load a fixed template consisting of four buttons from XML:

const string Template = @"
	<simplepanel>

        <shape name="" shape""="" shape="" roundrect""="">

		<border padding="" 2""="">

			<stackpanel name="" radiogroup""="" orientation="" vertical""="" spacing="" 1""="" horizontalalignment="" center""="">
				<radiobuttoncomponent padding="" 2""="">
					<radiobuttoncomponent.content>
						<text text="" option="" 1""="" font="" verdana,="" 3world,="" style="Bold&quot;&quot;">
					</text></radiobuttoncomponent.content>
				</radiobuttoncomponent>
				<radiobuttoncomponent padding="" 2""="">
					<radiobuttoncomponent.content>
						<text text="" option="" 2""="" font="" verdana,="" 3world,="" style="Bold&quot;&quot;">
					</text></radiobuttoncomponent.content>
				</radiobuttoncomponent>
				<radiobuttoncomponent padding="" 2""="">
					<radiobuttoncomponent.content>
						<text text="" option="" 3""="" font="" verdana,="" 3world,="" style="Bold&quot;&quot;">
					</text></radiobuttoncomponent.content>
				</radiobuttoncomponent>
				<radiobuttoncomponent padding="" 2""="">
					<radiobuttoncomponent.content>
						<text text="" option="" 4""="" font="" verdana,="" 3world,="" style="Bold&quot;&quot;">
					</text></radiobuttoncomponent.content>
				</radiobuttoncomponent>
			</stackpanel>

		</border>

    </shape></simplepanel>";

The template can be loaded using the XmlLoader class. We’ll also store a reference to the stack panel so we can access its child radio buttons:

public OptionNode()
{
	Load();
}

public OptionNode(Diagram d)
	: base(d)
{
	Load();
}

private void Load()
{
	Components.Add(XmlLoader.Load(Template, this, null));

	radioGroup = FindComponent("RadioGroup") as StackPanel;
}

StackPanel radioGroup;

Now implement a SelectedOption property that lets us select a radio button by its index. Define it as nullable integer so we can represent missing select too:

public int? SelectedOption
{
	get
	{
		for (int i = 0; i < radioGroup.Components.Count; i++)
		{
			var radioButton = (RadioButtonComponent)radioGroup.Components[i];
			if (radioButton.IsChecked)
				return i;
		}
		return null;
	}
	set
	{
		for (int i = 0; i < radioGroup.Components.Count; i++)
		{
			var radioButton = (RadioButtonComponent)radioGroup.Components[i];
			radioButton.IsChecked = value == i;
		}
	}
}

Let’s try it – create a few nodes and run the application, you’ll see the screen shown below:

var node1 = new OptionNode();
node1.Bounds = new RectangleF(20, 20, 30, 40);
node1.SelectedOption = 0;
diagram.Nodes.Add(node1);

var node2 = new OptionNode();
node2.Bounds = new RectangleF(90, 20, 30, 40);
node2.SelectedOption = 1;
diagram.Nodes.Add(node2);

var node3 = new OptionNode();
node3.Bounds = new RectangleF(20, 80, 30, 40);
node3.SelectedOption = null;
diagram.Nodes.Add(node3);

var node4 = new OptionNode();
node4.Bounds = new RectangleF(90, 80, 30, 40);
node4.SelectedOption = 3;
diagram.Nodes.Add(node4);

for (int i = 0; i < diagram.Nodes.Count - 1; i++)
	diagram.Factory.CreateDiagramLink(
		diagram.Nodes[i], diagram.Nodes[i + 1]);

Radio buttons in MindFusion diagram nodes

To be fair, this kind of nodes is simple enough to implement using standard TableNode class where radio button graphics are either custom drawn or set as Image inside table cells in first column, and text displayed in second column. However the radio buttons can be mixed with other components in CompositeNodes to implement more complex user interfaces than ones possible with tables.

For more information on MindFusion flow diagramming libraries for various desktop, web and mobile platforms, see MindFusion.Diagramming Pack page.

Enjoy!