public override void GetContent(Invert.Core.GraphDesigner.IDocumentationBuilder _)
{
base.GetContent(_);
_.Paragraph("Commands define a property of type ISignal<TCommandClass> on an element. These act as a delegate to a pre-initialized handler." +
"The handlers are setup in the controllers initialize method to point to a method, this makes a controller a factory for a view-model." +
"");
var ele = new ScaffoldGraph()
.BeginNode<ElementNode>("Player")
.AddItem<CommandsChildItem>("Hit")
.EndNode();
_.Title2("Executing a command from a viewmodel.");
_.CodeSnippet("MyViewModel.CommandName.OnNext(new {CommandName}Command() { });");
_.Break();
_.Title2("Generated Model Commands");
_.TemplateExample<ViewModelTemplate, ElementNode>(ele as ElementNode, true, "CommandItems", "Bind");
_.Break();
_.Title2("Generated Controller Command Handlers");
_.TemplateExample<ControllerTemplate, ElementNode>(ele as ElementNode, true, "CommandItems", "CommandMethod");
_.Break();
_.Break();
_.Title3("Also See");
_.LinkToPage<Controllers>();
_.LinkToPage<CreationAndInitialization>();
}
public override void GetContent(Invert.Core.GraphDesigner.IDocumentationBuilder _)
{
base.GetContent(_);
_.Paragraph("uFrame utilizes a new concept developed to fit the uFrame and RX paradigm of programming. " +
"We like to call this \"Re-active\" state machines. ");
_.Paragraph("The main concept of state machines in uFrame MVVM is that when defining them in your " +
"diagrams you don't need to focus on anything other than the states and transitions that " +
"make up the machine. Making the machine come to life is a matter of wiring them to " +
"computed properties, command bindings, and view bindings. But initially, you can focus " +
"purely on the high-level rather than implementation, once commands and transitions " +
"are all wired together transitions and states can be easily tweaked with minimal to " +
"zero changes in your code.");
_.Paragraph("Reactive State Machines employ the concept of data subscriptions causing transitions, rather than polling data every frame, this can give a definite increase in performance.");
_.Break();
_.ImageByUrl("http://i.imgur.com/CPymbPH.png");
_.Break();
_.Title2("Manually triggering transitions");
_.Paragraph("To set a transition in code, you'll need to access the state machine property. This should be the property that is connected to your state machine.");
_.CodeSnippet("{ViewModel}.{StateMachinePropertyName}Property.Transition(\"NAME OF TRANSITION HERE\");");
_.Break();
_.Title2("Manually setting the state");
_.CodeSnippet("{ViewModel}.{StateMachinePropertyName}Property.SetState<STATE_CLASS_NAME>();");
_.Title3("or");
_.CodeSnippet("{ViewModel}.{StateMachinePropertyName}Property.SetState(\"STATE_NAME\");");
_.Break();
// _.AlsoSeePages(typeof(UsingStateMachines));
}
public override void GetContent(Invert.Core.GraphDesigner.IDocumentationBuilder _)
{
base.GetContent(_);
_.Paragraph("Usually a property binding is a one-way stream of information, where a View is simply receiving information about a property's changing values. In a game environment however, there are times where it makes sense to allow Views to actually determine these values as well, for things like a GameObject's position, rotation, etc. This is where two-way bindings are needed.");
_.Paragraph("Scene Properties are two-way bindings, and allows for a View to calculate and set a property on its ViewModel. This is done in an observable way, and when adding a scene property to a view, 3 specific methods are made available. For example, adding a Position scene property on a PlayerView will result in these underlying base methods:");
var graph = new ScaffoldGraph();
PropertiesChildItem positionProperty;
var player = graph.BeginNode<ElementNode>("Player")
.AddItem("Position", out positionProperty, "Vector3")
.EndNode();
var view = (ViewNode) graph
.BeginNode<ViewNode>("PlayerView")
.EndNode()
;
graph.AddConnection(player, view.ElementInputSlot);
graph.AddConnection(positionProperty,view.ScenePropertiesInputSlot);
_.Break();
_.Title2("Under the hood. Generated Scene Properties");
_.TemplateExample<ViewTemplate, ViewNode>(view, true,new [] {"ResetProperty", "CalculateProperty", "GetPropertyObservable", "Bind"});
_.Break();
_.Title3("ResetPosition");
_.Paragraph("ResetPosition() is mostly used to initialize the binding, is called in the View's base Bind() method, and typically doesn't need to be overridden and altered.");
_.Break();
_.Title3("CalculatePosition");
_.Paragraph("CalculatePosition() is the main method you would override on your generated PlayerView, where you would return a Vector3 to give the player's position.");
_.ShowGist("2e5c9b6d07b76bcaedb1","CalculatePosition.cs");
_.Break();
_.Title3("GetPositionObservable");
_.Paragraph("GetPositionObservable() should only be overridden in cases where you have a more convenient or performant method of observing the scene property change, because as you see, this calculation is happening every Update by default. In this case, we know that ViewBase is already monitoring a TransformChangedObservable (and specifically a PositionChangedObservable as well), so on our PlayerView we would override the GetPositionObservable like this:");
_.ShowGist("3d0e2f8bd65a044e82d6","GetPositionObservable.cs");
_.Break();
}
public override void GetContent(Invert.Core.GraphDesigner.IDocumentationBuilder _)
{
base.GetContent(_);
ElementNode elementNode;
ScaffoldComputed(out elementNode);
_.ImageByUrl("http://i.imgur.com/qeDjrVR.png");
_.Paragraph("A Computed Property is an extremely powerful feature added to uFrame's ViewModel layer. These properties can calculate their value based on other properties, and are recalculated whenever one of those change. So for example, if you have a boolean IsDead computed property on a PlayerViewModel, dependent on a Player.Health property, uFrame will generate these three things on that PlayerViewModel for you to override:");
_.Break();
_.Title2("Under the hood");
_.TemplateExample<ViewModelTemplate, ElementNode>(elementNode, true, "ComputedDependents", "ResetComputed", "Compute");
_.Title3("ResetIsDead");
_.Paragraph("Usually fine without overriding, used to set up the computed observable");
_.Title3("ComputeIsDead");
_.Paragraph("A \"getter\" used to compute the property, based on dependents");
_.Title3("GetIsDeadDepedents");
_.Paragraph("This where you would override and provide additional dependents, if needed");
_.Break();
_.Note("It's important to correctly set all the dependents, because the computed only knows when it needs to be recalculated by observing these dependents for changes. This is done in the diagram by dragging a link from a ViewModel's property to the computed, or in code by overriding the Get{ComputedName}Dependents() function.");
}
public override void GetContent(Invert.Core.GraphDesigner.IDocumentationBuilder _)
{
base.GetContent(_);
_.Title2("What Is It?");
_.Paragraph("Views are perhaps the easiest layer to understand in the context of Unity, due to the fact that they exist specifically to interact with the UnityEngine namespace and environment. Views are the \"presentation\" layer, where the data of ViewModels are represented through the implementation of bindings. The idea is that for the most part, ViewModel data exists somewhere already, and Views merely \"bind\" to that data in order to represent changes in a way that Unity and players can understand.");
_.Title2("Where does it exist in Unity?");
_.Paragraph("Views generated by uFrame inherit from Monobehaviour, and therefore are much like normal Unity components. Building on top of Monobehaviour, uFrame ties into Unity methods like Update, Start, OnEnable, and OnDestroy in order to implement necessary MVCVM functionality within the Unity environment. Every View that uFrame generates is meant to exist as a component on a particular GameObject. For instance, a PlayerView should probably exist on some kind of Player GameObject, and a PlayerHUDView should probably exist on some kind of GUI GameObject to bind to and express a player's stats and other properties.");
_.Title2("What Else?");
_.Title3("The Presentation Layer");
_.Paragraph("Views are the so-called \"presentation\" layer, where a programmer will implement the logic of how the abstracted ViewModel data is represented in a particular environment. So if you have a PlayerViewModel, you may decide to represent that in any number of ways, including:");
_.Paragraph(" - a PlayerView that represents the player as an animated character moving in space");
_.Paragraph(" - a PlayerHUDView that represents the player's health, stamina, energy and other stats in your GUI");
_.Paragraph(" - and maybe a PlayerMapView that represents your player's position relative to some kind of GUI Map object");
_.Paragraph("All of these would typically want to bind to the same player ViewModel \"instance\", such that they are said to share the same ViewModel, representing the data in different ways. The most important distinction is that each of these views should concern themselves with ONLY their own representation, meaning that the PlayerView in the above example should not be updating GUI elements, but rather leave that to the PlayerHUDView or possibly PlayerMapView. Views should be as independent as possible, handling just themselves and their own interactions. Since any number of Views can bind to the same ViewModel instance, it is up to you to determine how many Views are needed and what their individual responsibilities will be in representing that data inside Unity.");
ImportantMethods(_);
}
