A demonstrator of Factory pattern in Augmented Reality context. Goal is create a set of predefined virtual objects and manipulate them without knowledge of RealityKit. I intented show the fusion of consolidated practices & an innovate framework. Tradition & disruption.
User can choose the creation of two basic shapes of virtual objects: a cube or a sphere. These objects have a basic setting to be manipulated with gestures (drag, rotation, ...) or not. Hence there are four possible objects / products that can be created.
With the factory pattern approach these predefined virtual objects are created on consistent way and anyone may create them without any knowledge of Realitykit with intuitive calls. Created virtual objects aka "Products" with the factory pattern are conformed to VirtualObject protocol
This is the call of the creation of a sphere with gestures with the factory pattern
And these are the required lines of code in RealityKit to create the sphere with gestures
It's really easier, faster and intuitive the first way.
Two features are added to the objects/products created by the factory pattern to grant a way to manipulate objects with "classic" calls: change the height and change the color. It also allows add some checks, for example: it checks height never goes behind the "ground"
This is the call to change the color with protocol related to the factyory pattern (a random color is the new one)
This is how RealityKit change color under the hood
This is the call to change the height with protocol related to the factyory pattern
This is how RealityKit change height under the hood
Thought saiving in lines of code is limited and RealityKit lines seems intuive while you are reading them...to code you previously need dig in the ModelEntity documentation to figure out how change this stuff. I chose these ones to get an easy to understand demostrator. And the height check allows the project incorporate this is consistency check (we could consider a behaviour bug if the object starts going down our ground)
While system is detecting the plane user receives feedback from ARCoaching. Once it is detected, inputs to system comes from user detecting taps and other gestures (if allowed by virtual object on the screen and SwitUI controls; obviously the movement of the device is other input.
When app is lanched while the base plane is detected, ARCoacing is used to give feedback to the user
Once the plane is detected, control goes to user.
User selects the type of virtual object in a Picker
The virtual object is created with a tap in the screen. Where? on the point of tap
Once the virtual object is shown, some buttons to control it are displayed. We can change height or color, also delete the object. Gestures like drag are recognized if virtual object allows them. It is used the MVVM pattern.