To apply a gradient to a NSView, it seems best to make a class that inherits from NSView and override the draw method that sets a path and draws the gradient.
class GradientView: NSView {
var startColor = NSColor.red
var startPosition: CGFloat = 0
var endColor = NSColor.white
var endPosition: CGFloat = 1
var rotation: CGFloat = 0
override func draw(_ dirtyRect: NSRect) {
super.draw(dirtyRect)
let bgGradient = NSGradient(colorsAndLocations: (startColor, startPosition), (endColor, endPosition))
let path = NSBezierPath.init(rect: self.bounds)
bgGradient?.draw(in: path, angle: rotation)
}
}
(Let me know if there's a better/easier way...)
It works great, but now I want to send a custom collection of colors & color-stops. Note: *With this above method, I am limited to defining two (the startColor & startPosition and endColor & endPosition) only.
I'd like to use (an array ? of) tuples (containing several colors and their stop values). The API to set a color and location is NSGradient(colorsAndLocations:) and uses the tuple (NSColor, CGFloat).
I can add a parameter like var colorStopArray: [(color: NSColor, stop: CGFloat)] = [] that creates an array of (such) tuples. Also I can append values with: colorStopArray.append((startColor, startPosition)).
But how can I assign this to the API?
I've tried many methods and they all cause errors (eg. tried NSGradient(colorsAndLocations: colorStopArray.flatMap{return ($0.color, $0.stop)}) )
So how can send many custom colors+locations to the above class (or hopefully a better suggestion) to create a custom gradient (which will then be drawn on my NSView)?
With this above method, I am limited to defining two (the startColor & startPosition and endColor & endPosition) only.
No, you're not. This is a variadic; you can add as many tuples as you wish. This works fine:
let startColor = NSColor.red
let startPosition: CGFloat = 0
let middleColor = NSColor.blue
let middlePosition: CGFloat = 0.5
var endColor = NSColor.white
var endPosition: CGFloat = 1
let bgGradient = NSGradient(colorsAndLocations:
(startColor, startPosition),
(middleColor, middlePosition),
(endColor, endPosition))
If the question is "can I turn an array into a variadic?", then the answer is, unfortunately no. This feature ("splatting") is missing from the Swift language. The only way to call a variadic in Swift is as a variadic.
The initializer that lets you supply an array of colors and stop locations is init(colors:atLocations:colorSpace:). So if that's what you want to supply, call that initializer instead.
Using matt's clues, I replaced the let line (using the more appropriate API init(colors:atLocations:colorSpace:)) :
let bgGradient = NSGradient(colorsAndLocations: (startColor, startPosition), (endColor, endPosition))
with
let bgGradient = NSGradient(colors: colorStopArray.map { $0.color }, atLocations: colorStopArray.map { $0.stop }, colorSpace: NSColorSpace.genericRGB)
Now it functions as I desired.
Related
I have a spriteKit project where I have many characters across several scenes.
As a beginner I just built each one individually for each scene - which makes for a ton of extra code.
I know I could clean this up with a "Build character class" or something like that...
I am just not sure where to begin.
Here is code from two of the characters in one scene...but imagine 5-10 characters per scene?
Also is there a way a property list could be useful for storing these type of properties?
//BEAR
func buildBear() {
let bearAnimatedAtlas = SKTextureAtlas(named: "Bear")
var bearFrames: [SKTexture] = []
let numImages = bearAnimatedAtlas.textureNames.count
for i in 1...numImages {
let bearTextureName = "bear\(i)"
bearFrames.append(bearAnimatedAtlas.textureNamed(bearTextureName))
}
animatedBear = bearFrames
let firstFrameTexture = animatedBear[0]
bear = SKSpriteNode(texture: firstFrameTexture)
bear.size.height = 370
bear.size.width = 370
bear.position = CGPoint(x: 295, y: 25)
bear.zPosition = 1
bear.name = "bear"
isUserInteractionEnabled = true
addChild(bear)
}
//CAT
func buildCat() {
let catAnimatedAtlas = SKTextureAtlas(named: "Cat")
var catFrames: [SKTexture] = []
let numImages = catAnimatedAtlas.textureNames.count
for i in 1...numImages {
let catTextureName = "cat\(i)"
catFrames.append(catAnimatedAtlas.textureNamed(catTextureName))
}
animatedCat = catFrames
let firstFrameTexture = animatedCat[0]
cat = SKSpriteNode(texture: firstFrameTexture)
cat.size.height = 240
cat.size.width = 240
cat.position = CGPoint(x: 134, y: -38)
cat.zPosition = 2
cat.name = "cat"
isUserInteractionEnabled = true
addChild(cat)
}
How could I clean up something like this - I need different position/size per scene but I imagine I could just override that per scene?
I know I know how to do this! - just not where to start?
Gimme a nudge please!
One of the confusing things about the existence of so many languages is that they each have their own jargon, and their own conventions. The root of your problem, however, has nothing to do with Swift or Sprite Kit. When I read your question, I see code that could use some Abstract Data Types. In Java, you would create an Interface, in C++ you would create a "pure virtual" class. Well a rose by any other name still gets the job done. I recommend creating a Protocol, perhaps called Spritable, to define the types of objects that you intend to build into sprites. It would probably be as simple as this:
protocol Spritable {
var species: String { get }
var height: Int { get }
var width: Int { get }
}
The only other thing that differs between your two functions appears to be the starting position. Since this is not inherent in the meaning of a Spritable object, I would package that data separately. A tuple should do the job. With these revisions, your two functions can be merged into one:
func buildSprite(of creature: Spritable, at position: (x: Int, y: Int, z: Int)) {
let spriteAnimatedAtlas = SKTextureAtlas(named: creature.species)
var spriteFrames: [SKTexture] = []
let numImages = spriteAnimatedAtlas.textureNames.count
for i in 1...numImages {
let spriteTextureName = "\(creature.species.lowercased())\(i)"
spriteFrames.append(spriteAnimatedAtlas.textureNamed(spriteTextureName))
}
animatedSprite = spriteFrames
let firstFrameTexture = animatedSprite[0]
sprite = SKSpriteNode(texture: firstFrameTexture)
sprite.size.height = creature.height
sprite.size.width = creature.width
sprite.position = CGPoint(x: position.x, y: position.y)
sprite.zPosition = position.z
sprite.name = creature.species
isUserInteractionEnabled = true
addChild(sprite)
}
To build a bear, aside from a workshop, you will need to define a struct that implements Spritable:
struct Bear: Spritable {
var species: String { return "Bear" }
var height: Int
var width: Int
init(height: Int, width: Int) {
self.height = height
self.width = width
}
}
Then here would be your function call:
buildSprite(of: Bear(height: 370, width: 370), at: (295, 25, 1))
This is a pretty simple example, and could be solved in a few simpler ways than this. However, I find that the larger a project gets, the greater the benefits of organizing code around Abstract Data Types become, so it's worth taking that approach even in a simple case like this.
I ended up not using a protocol on this.
I simply built a method to construct the sprites - similar to what #TallChuck suggested.
func buildCharacter(name:String, height: CGFloat, width: CGFloat, position: CGPoint, zPosition: CGFloat) {
let animatedAtlas = SKTextureAtlas(named: name)
var animationFrames: [SKTexture] = []
let numImages = animatedAtlas.textureNames.count
for i in 1...numImages {
let textureName = "\(name)\(i)"
animationFrames.append(animatedAtlas.textureNamed(textureName))
}
animatedCharacter = animationFrames
let firstFrameTexture = animatedCharacter[0]
builtCharacter = SKSpriteNode(texture: firstFrameTexture)
builtCharacter.size.height = height
builtCharacter.size.width = width
builtCharacter.position = position
builtCharacter.zPosition = zPosition
builtCharacter.name = name
isUserInteractionEnabled = true
addChild(builtCharacter)
}
It works perfect for building and adding to the scenes - had some issues accessing the nodes names for touch detection but got it sorted. Now trying to figure out how to call actions on the nodes - its all different than the normal way. So I will prolly ask that question next. But overall reduced a ton of repeated code! Thanks for the help.
None of this has to be done in code. With Sprite Kit, you create your bear and your cat via the sprite kit editor, and then you load in the sks file by using the constructor that loads by filename.
This is a similar behavior to how game objects work in unity.
I have created a very simple scene ("SpeechScene") using Reality Composer, with a single speech callout object ("Speech Bubble") anchored to a Face anchor.
I have loaded this scene into code via the following:
let speechAnchor = try! Experience.loadSpeechScene()
arView.scene.anchors.append(speechAnchor)
let bubble = (arView.scene as? Experience.SpeechScene)?.speechBubble
It renders as expected. However, I would like to dynamically change the text of this existing entity.
I found a similar question here, but it's unclear to me how to refer to the meshResource property of a vanilla RealityKit.Entity object.
Is this possible? Thank you!
First Approach
At first you need to find out what's an hierarchy in Reality Composer's scene containing Bubble Speech object. For that I used simple print() command:
print(textAnchor.swift!.children[0].components.self) /* Bubble Plate */
print(textAnchor.swift!.children[1].components.self) /* Text Object */
Now I can extract a text entity object:
let textEntity: Entity = textAnchor.swift!.children[1].children[0].children[0]
And bubble plate entity object:
let bubbleEntity: Entity = textAnchor.swift!.children[0]
Here's a final code version that you can adapt for your needs:
import RealityKit
class GameViewController: UIViewController {
#IBOutlet var arView: ARView!
override func viewDidLoad() {
super.viewDidLoad()
let textAnchor = try! SomeText.loadTextScene()
let textEntity: Entity = textAnchor.swift!.children[1].children[0].children[0]
textAnchor.swift!.parent!.scale = [4,4,4] // Scale for both objects
var textModelComp: ModelComponent = (textEntity.components[ModelComponent])!
var material = SimpleMaterial()
material.baseColor = .color(.red)
textModelComp.materials[0] = material
textModelComp.mesh = .generateText("Obj-C",
extrusionDepth: 0.01,
font: .systemFont(ofSize: 0.08),
containerFrame: CGRect(),
alignment: .left,
lineBreakMode: .byCharWrapping)
textEntity.position = [-0.1,-0.05, 0.01]
textAnchor.swift!.children[1].children[0].children[0].components.set(textModelComp)
arView.scene.anchors.append(textAnchor)
}
}
Second Approach
And you can always use a simpler approach for this case – to create several scenes in Reality Composer, each one must contain different speech-object.
Consider, this code isn't for tracking, it's just a test for dynamically switching two objects using Tap Gesture. Then you need to adapt this code for tracking faces.
import RealityKit
class ViewController: UIViewController {
#IBOutlet var arView: ARView!
var counter = 0
var bonjourObject: FaceExperience.Bonjour? = nil
var holaObject: FaceExperience.Hola? = nil
override func viewWillAppear(_ animated: Bool) {
super.viewWillAppear(animated)
// Reality Composer Scene named "Bonjour"
// Model name – "french"
bonjourObject = try! FaceExperience.loadBonjour()
bonjourObject?.french?.scale = SIMD3(x: 2, y: 2, z: 2)
bonjourObject?.french?.position.y = 0.25
// Reality Composer Scene named "Hola"
// Model name – "spanish"
holaObject = try! FaceExperience.loadHola()
holaObject?.spanish?.scale = SIMD3(x: 2, y: 2, z: 2)
holaObject?.spanish?.position.z = 0.3
}
#IBAction func tapped(_ sender: UITapGestureRecognizer) {
if (counter % 2) == 0 {
arView.scene.anchors.removeAll()
arView.scene.anchors.append(holaObject!)
} else {
arView.scene.anchors.removeAll()
arView.scene.anchors.append(bonjourObject!)
}
counter += 1
}
}
If you want a text portion to be on the same place – just copy-paste object from one scene to another.
#maxxfrazer is correct in his assertion that currently the only way to change text dynamically is to replace the ModelComponentof the Entity assuming of course it adheres to the HasModel Protocol.
I have written a simple extension which can help with this:
//-------------------------
//MARK: - Entity Extensions
//-------------------------
extension Entity{
/// Changes The Text Of An Entity
/// - Parameters:
/// - content: String
func setText(_ content: String){ self.components[ModelComponent] = self.generatedModelComponent(text: content) }
/// Generates A Model Component With The Specified Text
/// - Parameter text: String
func generatedModelComponent(text: String) -> ModelComponent{
let modelComponent: ModelComponent = ModelComponent(
mesh: .generateText(text, extrusionDepth: TextElements().extrusionDepth, font: TextElements().font,
containerFrame: .zero, alignment: .center, lineBreakMode: .byTruncatingTail),
materials: [SimpleMaterial(color: TextElements().colour, isMetallic: true)]
)
return modelComponent
}
}
//--------------------
//MARK:- Text Elements
//--------------------
/// The Base Setup Of The MeshResource
struct TextElements{
let initialText = "Cube"
let extrusionDepth: Float = 0.01
let font: MeshResource.Font = MeshResource.Font.systemFont(ofSize: 0.05, weight: .bold)
let colour: UIColor = .white
}
In order to use it lets say you create an Entity called textEntity:
var textEntity = Entity()
You can then set the dynamically change the Text via replacing the ModelComponent and setting the MeshResource at any time by simply calling the following method:
textEntity.setText("Stack Overflow")
Of course in regard to centering or aligning the text you will need to do some simple calculations (which I have ommited here).
Hope it helps.
Find your model entity (maybe by putting a breakpoint and looking through the children initially), find the Entity that conforms to the HasModel protocol, then replace its model with a different one using generatetext:
https://developer.apple.com/documentation/realitykit/meshresource/3244422-generatetext
I've tried so many different ways of add plots package from GitHub or using core plots or combining objective-c into swift, but it appeared so many problems during the process, and during this week I didn't make a successful chart. I'm really depressed.
Is there anyone who succeed in creating a pie chart in swift? The similar questions seem don't have successful answers.
I would really appreciate your help!
Don't be depressed. You just have to add a more specific question to get more help. For example, if you start from scratch and try to integrate a plots package from Github, you have to say what package, how did you try to integrate it, what errors are you getting etc.
However, drawing a simple pie chart is pretty easy with CoreGraphics functionality. Here is a little gift from my code, this draws progress value as a simple black and white pie chart. It only has 2 sections, but you can generalize from it
#IBDesignable class ProgressPieIcon: UIView {
#IBInspectable var progress : Double = 0.0 {
didSet {
self.setNeedsDisplay()
}
}
required init(coder aDecoder: NSCoder) {
super.init(coder:aDecoder)
self.contentMode = .Redraw
}
override init(frame: CGRect) {
super.init(frame: frame)
self.backgroundColor = UIColor.clearColor()
self.contentMode = .Redraw
}
override func drawRect(rect: CGRect) {
let color = UIColor.blackColor().CGColor
let lineWidth : CGFloat = 2.0
// Calculate box with insets
let margin: CGFloat = lineWidth
let box0 = CGRectInset(self.bounds, margin, margin)
let side : CGFloat = min(box0.width, box0.height)
let box = CGRectMake((self.bounds.width-side)/2, (self.bounds.height-side)/2,side,side)
let ctx = UIGraphicsGetCurrentContext()
// Draw outline
CGContextBeginPath(ctx)
CGContextSetStrokeColorWithColor(ctx, color)
CGContextSetLineWidth(ctx, lineWidth)
CGContextAddEllipseInRect(ctx, box)
CGContextClosePath(ctx)
CGContextStrokePath(ctx)
// Draw arc
let delta : CGFloat = -CGFloat(M_PI_2)
let radius : CGFloat = min(box.width, box.height)/2.0
func prog_to_rad(p: Double) -> CGFloat {
let rad = CGFloat(p * 2 * M_PI)
return rad + delta
}
func draw_arc(s: CGFloat, e: CGFloat, color: CGColor) {
CGContextBeginPath(ctx)
CGContextMoveToPoint(ctx, box.midX, box.midY)
CGContextSetFillColorWithColor(ctx, color)
CGContextAddArc(
ctx,
box.midX,
box.midY,
radius-lineWidth/2,
s,
e,
0)
CGContextClosePath(ctx)
CGContextFillPath(ctx)
}
if progress > 0 {
let s = prog_to_rad(0)
let e = prog_to_rad(min(1.0, progress))
draw_arc(s, e, color)
}
}
We've made some changes to the Core Plot API, including replacing NSDecimal with NSDecimalNumber, for Swift compatibility. The changes are on the release-2.0 branch and not in a release package yet. See Core Plot issue #96 for more discussion of the issue.
Have you tried to look for any CorePlot tutorial with Swift ? Like this one maybe.
Otherwise, you might want to give a look to other framework.
I am trying to set a property for a particle emitter in swift, using scenekit.
I feel really stupid as I cannot figure out how to set constants that are set up as enums.
Here's the code:
//Create emitter
let emitter = SCNParticleSystem()
//set up the emitter trying to use the enum values
emitter.birthLocation = SCNParticleBirthLocation.SCNParticleBirthLocationSurface
//I have tried a few other ways including:
emitter.birthLocation = .SCNParticleBirthLocationSurface
emitter.birthLocation = SCNParticleSystem.SCNParticleBirthLocation.SCNParticleBirthLocationSurface
This isn't the only thing I have had trouble with using enums in different classes so I figure I have something fundamentally wrong.
The link to the apple framework reference in case anyone needs: https://developer.apple.com/library/prerelease/ios/documentation/SceneKit/Reference/SCNParticleSystem_Class/index.html#//apple_ref/occ/cl/SCNParticleSystem
And finally my whole code just in case it's something else entirely that I'm doing wrong
import Foundation
import SceneKit
class NextLevelScene: SCNScene{
override init(){
super.init()
let nextLevel: SCNNode = SCNNode()
let text: SCNGeometry = SCNText(string: "Next Level", extrusionDepth: 0.0)
let material = UIColor.blueColor()
text.firstMaterial!.diffuse.contents = material
nextLevel.geometry = text
let exp = SCNParticleSystem()
exp.loops = true
exp.birthRate = 5000
exp.emissionDuration = 0.2
exp.spreadingAngle = 180
exp.particleDiesOnCollision = true
exp.particleLifeSpan = 0.2
exp.particleLifeSpanVariation = 0.05
exp.particleVelocity = 20
exp.particleVelocityVariation = 3
exp.particleSize = 0.05
exp.stretchFactor = 0.02
exp.particleColor = UIColor.redColor()
exp.emitterShape = text
//Error here
exp.birthLocation = SCNParticleBirthLocation.SCNParticleBirthLocationSurface
self.addParticleSystem(exp, withTransform: SCNMatrix4MakeRotation(0, 0, 0, 0))
self.rootNode.addChildNode(nextLevel)
}
required init(coder: NSCoder) {
super.init(coder: coder)
}
}
Thanks in advance!
If you Command-click on birthLocation in the Xcode source file to get the
definition of the property, and then Command-click on SCNParticleBirthLocation
then you'll see that it is declared in Swift as
enum SCNParticleBirthLocation : Int {
case Surface //particles are emitted on the surface of the emitter shape.
case Volume //particles are emitted inside the volume of the emitter shape.
case Vertex //particles are emitted on the vertices of the emitter shape.
}
so you simply need
emitter.birthLocation = SCNParticleBirthLocation.Surface
which can be shortened (due to automatic type inference) to
emitter.birthLocation = .Surface
How do i return the radius of a sphere geometry(SCNSphere) used to set up a SCNNode.
I want to use the radius in a method where i move some child nodes in relation to a parent node. The code below fails since radius appears to be unknown to the resulting node, should i not pass the node to the method?
Also my array index fails saying that Int is not a Range.
I am trying to build something from this
import UIKit
import SceneKit
class PrimitivesScene: SCNScene {
override init() {
super.init()
self.addSpheres();
}
func addSpheres() {
let sphereGeometry = SCNSphere(radius: 1.0)
sphereGeometry.firstMaterial?.diffuse.contents = UIColor.redColor()
let sphereNode = SCNNode(geometry: sphereGeometry)
self.rootNode.addChildNode(sphereNode)
let secondSphereGeometry = SCNSphere(radius: 0.5)
secondSphereGeometry.firstMaterial?.diffuse.contents = UIColor.greenColor()
let secondSphereNode = SCNNode(geometry: secondSphereGeometry)
secondSphereNode.position = SCNVector3(x: 0, y: 1.25, z: 0.0)
self.rootNode.addChildNode(secondSphereNode)
self.attachChildrenWithAngle(sphereNode, children:[secondSphereNode, sphereNode], angle:20)
}
func attachChildrenWithAngle(parent: SCNNode, children:[SCNNode], angle:Int) {
let parentRadius = parent.geometry.radius //This fails cause geometry does not know radius.
for var index = 0; index < 3; ++index{
children[index].position=SCNVector3(x:Float(index),y:parentRadius+children[index].radius/2, z:0);// fails saying int is not convertible to range.
}
}
required init(coder aDecoder: NSCoder) {
fatalError("init(coder:) has not been implemented")
}
}
The issue with radius is that parent.geometry returns a SCNGeometry and not an SCNSphere. If you need to get the radius, you'll need to cast parent.geometry to SCNSphere first. To be safe, it's probably best to use some optional binding and chaining to do that:
if let parentRadius = (parent.geometry as? SCNSphere)?.radius {
// use parentRadius here
}
You'll also need to do that when accessing the radius on your children nodes. If you put all that together and clean things up a little, you get something like this:
func attachChildrenWithAngle(parent: SCNNode, children:[SCNNode], angle:Int) {
if let parentRadius = (parent.geometry as? SCNSphere)?.radius {
for var index = 0; index < 3; ++index{
let child = children[index]
if let childRadius = (child.geometry as? SCNSphere)?.radius {
let radius = parentRadius + childRadius / 2.0
child.position = SCNVector3(x:CGFloat(index), y:radius, z:0.0);
}
}
}
}
Note though that you're calling attachChildrenWithAngle with an array of 2 children:
self.attachChildrenWithAngle(sphereNode, children:[secondSphereNode, sphereNode], angle:20)
If you do that, you're going to get a runtime crash in that for loop when accessing the 3rd element. You'll either need to pass an array with 3 children every time you call that function, or change the logic in that for loop.
At this point I'm only working with spheres, hence the radius is quite easy to manipulate. But still, I think you should try to look at the geometry of the SCNNode. I use this function besides the code snippet you posted (the two works together just fine for me).
func updateRadiusOfNode(_ node: SCNNode, to radius: CGFloat) {
if let sphere = (node.geometry as? SCNSphere) {
if (sphere.radius != radius) {
sphere.radius = radius
}
}
}
Also the scaling factor (at the radius) is also important. I use a baseRadius then i'm multiplying that with
maxDistanceObserved / 2
Always updating the max distance if a hitTest has futher results.