scene — 2D Games and Animations

The scene module provides an easy way to create hardware-accelerated 2D graphics and animations, specifically games.

Overview

Introduction

For every scene-based project, you first have to create a subclass of Scene that is responsible for drawing your content, responding to touch events etc. To actually get a scene on screen, you instantiate your subclass, and pass it to the run() function. The Scene class provides various methods that you can override to customize its behavior, for example Scene.setup(), which is called just before your scene becomes visible on screen:

from scene import *

class MyScene (Scene):
    def setup(self):
        self.background_color = 'green'

run(MyScene())

In the minimal example above, the Scene.setup() method simply sets the background color of the scene to green.

To add actual content to your scene, you typically create Node objects – there are a couple of different subclasses of Node corresponding to different content, e.g. a SpriteNode renders an image, a LabelNode text, etc. A Scene is also a subclass of Node, and nodes can contain other nodes (forming a tree or “scene graph”). This is often useful to move or rotate a group of objects as a single entity. All nodes have position, rotation, scale, and alpha (opacity) attributes that determine how the node and its children are drawn. The default position is (0, 0) which corresponds to the lower-left corner of the screen.

In the following example, the setup() method is extended to add a space ship at the center of the screen:

from scene import *

class MyScene (Scene):
    def setup(self):
        self.background_color = 'midnightblue'
        self.ship = SpriteNode('spc:PlayerShip1Orange')
        self.ship.position = self.size / 2
        self.add_child(self.ship)

run(MyScene())

A couple of things to note here:

  • The string that was passed to the SpriteNode initializer is the name of a built-in image. You can access built-in images in Pythonista using the [+] button at the top of the editor.
  • For convenience, the objects representing sizes and positions in the scene module support a couple of standard operators – in this case, the scene’s size (which is set automatically before setup() is called) is simply divided by two, resulting in the center of the screen. Sizes and points can be used interchangeably, so assigning a size to the position attribute works. You could also use a simple tuple of two numbers.
  • By default, the position attribute of a SpriteNode corresponds to its center. In some cases it may be more convenient to set the position of one of the sprite’s corners instead – you can use the anchor_point attribute to change this behavior.

Touches and Simple Animations

Let’s extend this example once more to make it respond to touches. To do this, we simply override the Scene‘s touch_began() method:

from scene import *

class MyScene (Scene):
    def setup(self):
        self.background_color = 'midnightblue'
        self.ship = SpriteNode('spc:PlayerShip1Orange')
        self.ship.position = self.size / 2
        self.add_child(self.ship)

    def touch_began(self, touch):
        x, y = touch.location
        move_action = Action.move_to(x, y, 0.7, TIMING_SINODIAL)
        self.ship.run_action(move_action)

run(MyScene())

This last example introduces the Action class. Actions allow you to easily animate a Node‘s attributes, in this case its position. The timing mode (TIMING_SINODIAL) specifies the type of interpolation between the start and end values. If you leave out this argument, you get the default (linear) interpolation which results in a more abrupt movement. There are a lot of other Action factory methods for rotating or scaling nodes, changing their opacity (alpha), or even removing them from their parent node. You can also combine multiple Actions using Action.group() and Action.sequence(), or create an action that repeats another action using Action.repeat().

Similar to touch_began(), there are methods for detecting when a touch moves or ends – touch_moved() and touch_ended().

Advanced Animation and Motion Control

While Actions provide an easy-to-use high-level API for animating a scene’s content, you can also animate changes on a frame-by-frame basis by overriding the Scene‘s update() method. The following example demonstrates this by moving the space ship in response to the device’s orientation (using the gravity() function). By default, the update() method is called 60 times per second, so the ship’s position is set directly, instead of relying on Action-based animations:

from scene import *
import sound

class MyScene (Scene):
    def setup(self):
        self.background_color = 'midnightblue'
        self.ship = SpriteNode('spc:PlayerShip1Orange')
        self.ship.position = self.size / 2
        self.add_child(self.ship)

    def update(self):
        x, y, z = gravity()
        pos = self.ship.position
        pos += (x * 15, y * 15)
        # Don't allow the ship to move beyond the screen bounds:
        pos.x = max(0, min(self.size.w, pos.x))
        pos.y = max(0, min(self.size.h, pos.y))
        self.ship.position = pos

    def touch_began(self, touch):
        laser = SpriteNode('spc:LaserBlue9', position=self.ship.position, z_position=-1, parent=self)
        laser.run_action(Action.sequence(Action.move_by(0, 1000), Action.remove()))
        sound.play_effect('arcade:Laser_1')

run(MyScene(), PORTRAIT)

The gravity() function used here provides a very easy-to-use method of determining the device’s orientation. For more fine-grained control, you may want to use the motion module instead. In motion-controlled games, auto-rotation can be annoying, so this scene is restricted to portrait orientation using an additional argument to the run() function.

As a bonus, the space ship now shoots a laser when you touch the screen. The z_position attribute that is set when initializing the laser sprite determines the drawing order – it is set to -1 here so that the laser doesn’t appear on top of the ship. After the laser is added to the scene, it runs a sequence of two actions: The first one moves it by 1000 points upwards, and then it’s removed from the scene using the special remove action (which is not really an animation, but provided as an Action so that it can be part of an animation sequence).

This was just a very basic overview – there’s a lot more you can do with the scene module. The included Examples folder contains a couple of complete games that you can use to learn more advanced techniques.

Geometry

The coordinate system of a scene has its origin (0, 0) in the bottom-left corner.

As briefly mentioned in the overview, the scene module provides a couple of convenient classes for representing 2D geometry, primarily Rect, Point and Size. The latter two are basically identical, but used in different contexts.

These geometry objects are used for various Node attributes, e.g. Node.position, SpriteNode.size, Node.frame etc. When you assign a new value to these attributes, you don’t have to create a Point or Size object explicitly though – any sequence of two numbers (e.g. a tuple or list) will do just as well.

All of these classes behave like sequences, so you can e.g. access a Point‘s x coordinate using point[0] (though it’s typically more convenient to use point.x).

For Vector2, Point and Size, a couple of math operators are supported for convenience:

  • Multiplying a vector with a scalar (i.e. a number) multiplies both components with that number. The same goes for division. For example, Size(100, 200) * 0.5 results in Size(50, 100).
  • Multiplying a vector with another vector does component-wise multiplication, e.g. Point(100, 100) * (2, 3) results in Point(200, 300).
  • Adding a vector to another vector behaves the same way, e.g. Point(100, 100) + (20, 50) results in Point(120, 150).

For hit-testing purposes, you can check if a Point lies within a Rect using the in operator. To check if one rectangle intersects with another, you can use the Rect.intersects() method (or Rect.intersection() if you need to know exactly where the rectangles intersect).

Colors

Node attributes that represent a color, e.g. SpriteNode.color or Scene.background_color can be set in a variety of ways:

  • As an HTML hex color string, e.g. '#ff0000' for red.
  • As a CSS color name, e.g. 'green', 'blue', 'magenta'...
  • As a tuple of 3 or 4 numbers, corresponding to the red, green, blue, and alpha components in the range 0.0 - 1.0. The alpha component defaults to 1.0, i.e. fully opaque.
  • As a single number between 0.0 and 1.0, representing a grayscale value (0.0 = black, 1.0 = white).

Regardless of how you set a color attribute, reading it back will always result in a 4-tuple (r, g, b, a).

Classic Render Loop

In addition to the node-based approach described above, the scene module also supports a classic rendering loop that was the primary interface in previous versions of Pythonista. In most cases, a node-based scene will result in significantly better performance, but the classic rendering loop may be easier to understand if you already know similar programming environments, e.g. Processing.

To use this mode, you also have to create a subclass of Scene, but instead of adding nodes to it, you simply override the draw() method. This method is called for every frame (i.e. 60 times per second), and you can use module-level drawing functions to “paint” your content in every frame. Images and shapes that you draw to the screen in this way are not preserved from one frame to the next – in every call of the draw() method, you basically start with a blank screen.

Similar to classic OpenGL, you set up global state variables (like fill color and coordinate transformations) before calling the actual drawing functions. The following example draws a couple of rotated rectangles:

from scene import *

class MyScene (Scene):
    def draw(self):
        background('gray')
        colors = ['red', 'green', 'blue', 'yellow']
        # Move to the center of the screen:
        translate(self.size.w/2, self.size.h/2)
        for color in colors:
            # Set the current fill color (this affects the drawing commands that follow)
            fill(color)
            # Rotate the transformation matrix (this is cumulative):
            rotate(-20)
            # Note: The coordinates are relative to the center of the screen because of the previous `translate` command:
            rect(-50, -50, 100, 100)

run(MyScene())

If you’ve used a previous version of Pythonista, you may also be familiar with the Layer and Animation classes that were conceptually similar to Node and Action. These classes are still available for backwards compatibility, but it is strongly recommended to use the new Node/Action APIs instead because they provide significantly better performance.

Note

The functions you can use for drawing in this mode are documented in the scene_drawing module, though you don’t have to import scene_drawing explicitly if you want to use them.

Integration with the ui Module

The easiest way to run a scene is to use the run() function, which presents the scene in full-screen. Sometimes this is not what you want, especially if the scene you’re building is not actually a game. For more flexibility, you can also create a SceneView explicitly, set its scene attribute to the scene you want to present, and then add it to a view hierarchy that you created using the ui module.

You can also use this approach to add traditional UI elements (e.g. text fields) to your game. For this purpose, you don’t have to create a SceneView – you can simply use the view that was created automatically, using your Scene‘s view attribute (this will only work if a scene is already being presented of course).

The ui module can also be useful for rendering vector images (shapes, text etc.) that you want to use as textures. A Texture can be initialized from a ui.Image for this purpose, typically one that you create using an ui.ImageContext. The ShapeNode and LabelNode classes use this approach to render their content.

Node Classes

Scene

class scene.Scene

A scene is the root node in a tree of nodes (Node objects). These nodes provide content that the scene animates and renders for display. To display a scene, you typically call the module-level run() function to present it in full-screen. You can also create a SceneView explicitly, if you want to display a scene within other UI content.

A scene calculates the contents of a new frame by processing the following actions in order:

  1. The scene calls its update() method.
  2. The scene executes actions on its children.
  3. The scene calls its did_evaluate_actions() method.
  4. The scene renders all of its nodes and updates the view to display the new contents.

You typically create at least one subclass of Scene to handle touch events (by overriding touch_began(), touch_moved(), touch_ended()) and any other interaction with your game’s content.

Note that Scene is a subclass of EffectNode, so you can use a custom Shader for full-scene post-processing effects. Unlike a vanilla EffectNode however, a scene’s effects_enabled attribute is set to False by default.

Scene.setup()

This gets called once, just before the scene is presented on screen. You can use this to set up your scene. The size and bounds attributes are already valid at this point, so you can use them to determine the layout of your content.

Scene.touch_began(touch)

This method is called when a touch begins on the scene’s view. You typically don’t call this directly, but implement it as part of your Scene subclass.

You can inspect the touch object’s location attribute to get the position of the touch in the scene’s coordinate system. To distinguish between simultaneous touches, you can use the touch’s touch_id attribute.

Scene.touch_moved(node, touch)

This method is called when a touch moves in the scene’s view. You typically don’t call this directly, but implement it as part of your Scene subclass.

You can inspect the touch object’s location attribute to get the position of the touch in the scene’s coordinate system. To distinguish between simultaneous touches, you can use the touch’s touch_id attribute.

Scene.touch_ended(node, touch)

This method is called when a touch ends in the scene’s view. You typically don’t call this directly, but implement it as part of your Scene subclass.

You can inspect the touch object’s location attribute to get the position of the touch in the scene’s coordinate system. To distinguish between simultaneous touches, you can use the touch’s touch_id attribute.

Scene.did_change_size()

This method gets called whenever the scene’s size changes, usually when the screen rotates. You typically override this to reposition your content. The size attribute of the scene is already set to the new size when this is called.

Scene.did_evaluate_actions()

This method gets called after a scene has finished processing the actions of its child nodes.

Scene.update()

Performs any scene-specific updates that need to occur before scene actions are evaluated.

Do not call this method directly; it is called exactly once per frame, so long as the scene is presented in a view and is not paused. By default, this method does nothing. Your scene subclass should override this method and perform any necessary updates to the scene.

Scene.pause()

Gets called automatically when the home button is pressed while a scene is running. You can override this to save persistent state for example. The default implementation does nothing.

Scene.resume()

Gets called automatically when a scene is resumed (after being sent to the background with the home button). The default implementation does nothing.

Scene.stop()

Gets called automatically when a scene is stopped (by tapping the “x” button). You can override this to save persistent state. The default implementation does nothing.

Scene.present_modal_scene(other_scene)

Present another scene on top of this one. This can be useful for overlay menus etc. While the scene is being presented, it receives all touch events.

Scene.dismiss_modal_scene()

Close a scene that is being presented modally using Scene.present_modal_scene().

Scene Attributes

scene.bounds

A Rect that has the origin (0, 0) and the size of the drawable area.

Scene.dt

The time (in seconds) that has passed since the last invocation of update(). You can use this to calculate the progress of custom animations.

Scene.size

The size of the entire drawable area.

Scene.t

The time (in seconds) that has passed since the scene was started. You can use this to calculate the progress of custom animations.

Scene.touches

A dictionary of all touches that are currently active. The keys correspond to the touch_id attribute of the Touch objects.

Scene.background_color

The background color of the scene.

The default value is dark gray.

Scene.size

The dimensions of the scene in points (read-only). This corresponds to the size of the scene’s view.

Scene.view

The View that is currently presenting the scene. May be None if the scene is not currently being presented. (read-only)

Scene.presented_scene

The scene that is currently being presented using Scene.present_modal_scene() (if any, None otherwise).

Scene.presenting_scene

The presenting scene, if this scene is being presented by another scene (using present_modal_scene()).

Node

class scene.Node([position=(0, 0), z_position=0.0, scale=1.0, x_scale=1.0, y_scale=1.0, alpha=1.0, speed=1.0, parent=None])

The Node class is the fundamental building block of a scene. The basic Node class doesn’t draw anything by itself – its primary role is to provide baseline behavior that its subclasses use. Nodes can also contain other nodes to modify them as a group. For drawing actual content, you will typically use one of Node‘s subclasses:

  • SpriteNode – A node that draws a textured sprite
  • LabelNode – A specialized SpriteNode that renders a text string.
  • ShapeNode – A specialized SpriteNode that renders a shape based on a bezier path (e.g. a circle or rounded rectangle).
  • EffectNode – A node that applies effects to its children using a custom shader.

Nodes are organized hierarchically into node trees, similar to how views and subviews work. Most commonly, a node tree is defined with a Scene node as the root node and other content nodes as descendants. The scene node runs an animation loop that processes actions on the nodes, and then renders the contents of the node tree for display.

Every node in a node tree provides a coordinate system to its children. After a child is added to the node tree, it is positioned inside its parent’s coordinate system by setting its position attribute. A node’s coordinate system can be scaled and rotated by changing its x_scale, y_scale, and rotation attributes. When a node’s coordinate system is scaled or rotated, this transformation is applied both to the node’s own content and to that of its descendants.

The Node class does not perform any drawing of its own. However, many subclasses render visual content and so the Node class understands some visual concepts:

The frame attribute provides the bounding rectangle for a node’s visual content, modified by the scale and rotation attributes. The frame is non-empty if the node’s class draws content. Each node subclass determines the size of this content differently. In some subclasses, the size of the node’s content is declared explicitly, such as in the SpriteNode class. In other subclasses, the content size is calculated implicitly by the class using other object properties. For example, a LabelNode object determines its content size using the label’s text and font characteristics.

A node’s bbox is the largest rectangle that includes the frame of the node and the frames of all its descendants.

Other attributes, such as the alpha attribute, affect how the node and its descendants are drawn.

Any node in the tree may run actions (Action objects), which are used to animate the properties of a node, e.g. to move it to a new position smoothly.

Node.add_child(node)

Adds a node to the end of the receiver’s list of child nodes.

Node.remove_from_parent()

Removes the node from its parent node.

Node.remove_action(key)

Ends and removes a specific action, identified by its key, from the node. The key is an arbitrary string that was passed to the Node.run_action() method.

Node.remove_all_actions()

Ends and removes all actions from the node.

When an action is removed from the node, any remaining animation the action would perform is skipped; however, previous changes are not reverted.

Node.render_to_texture([crop_rect])

Creates a Texture object by rendering a snapshot of this node and its children.

Node.point_to_scene(point)

Convert a point from this node’s coordinate system to the coordinate system of its containing scene. If the node is not part of a scene, a ValueError is raised.

Node.point_from_scene(point)

Convert a point from the coordinate system of this node’s scene to the local coordinate system of the node. If the node is not part of a scene, a ValueError is raised.

Node.run_action(action[, key])

Adds an action to the list of actions executed by the node.

If an action using the same key (an arbitrary string) is already running, it is removed before the new action is added.

Node Attributes

Node.bbox

Calculates a rectangle in the parent’s coordinate system that contains the content of the node and all of its descendants.

Node.alpha

The transparency value applied to the node’s contents.

The Node class does not perform drawing, but many of its subclasses do. When a node or any of its descendants are drawn, the alpha component of each pixel is multiplied by the node’s alpha property and then clamped to the range 0.0-1.0. This modified alpha value is used to blend the pixel into the framebuffer. Subclasses that render content define properties that determine the blending operations used in conjunction with the alpha value to blend pixels into the parent’s framebuffer.

Node.frame

A rectangle in the parent’s coordinate system that contains the node’s content, ignoring the node’s children. (read-only)

Node.children

A list of this node’s child nodes. Note that modifying this list has no effect – use Node.add_child() and Node.remove_from_parent() instead.

Node.parent

The node’s parent node. (read-only)

Node.paused

A Boolean value that determines whether actions on the node and its descendants are processed.

Node.position

The position (x, y) of the node in its parent’s coordinate system.

Node.scene

The scene node that contains the node. (read-only)

If the node is not embedded in a scene, the value is None.

Node.speed

A speed modifier applied to all actions executed by a node and its descendants.

Node.x_scale

A scaling factor that multiplies the width of a node and its children.

The x_scale attribute scales the width of the node and all of its descendants. The scale value affects how a node’s frame is calculated, its hit test area, how it is drawn, and other similar characteristics. The default value is 1.0.

Node.y_scale

A scaling factor that multiplies the height of a node and its children.

The y_scale attribute scales the height of the node and all of its descendants. The scale value affects how a node’s frame is calculated, its hit test area, how it is drawn, and other similar characteristics. The default value is 1.0.

Node.z_position

The z position of a node determines the order in which it’s drawn, relative to its siblings. The default value is 0.0; nodes with larger values are rendered in front of nodes with smaller values.

Node.rotation

The node’s rotation about the z axis (in radians).

The default value is 0.0, which indicates no rotation. A positive value indicates a counterclockwise rotation. When the coordinate system is rotated, it affects the node and its descendants.

SpriteNode

class scene.SpriteNode([texture, position=(0, 0), z_position=0.0, scale=1.0, x_scale=1.0, y_scale=1.0, alpha=1.0, speed=1.0, parent=None, size=None, color='white', blend_mode=0])

A SpriteNode is a node that draws a textured image, a colored square, or a textured image blended with a color. You can also provide a custom shader to create your own rendering effects.

When initializing a SpriteNode, you can provide the texture as either a Texture object, a ui.Image, or the name of a built-in image (a string).

SpriteNode Attributes

SpriteNode.anchor_point

Defines the point in the sprite that corresponds to the node’s position.

You specify the value for this property in the unit coordinate space. The default value is (0.5, 0.5), which means that the sprite is centered on its position.

SpriteNode.blend_mode

The blend mode used to draw the sprite into the parent’s framebuffer.

The possible values for this property are listed in Blend Modes. The default value is BLEND_NORMAL.

SpriteNode.color

The sprite’s color.

If the texture attribute is set, the image is tinted with the given color. If the texture attribute is None, the color is used to draw a colored rectangle.

SpriteNode.shader

A property that determines whether the sprite is rendered using a custom shader.

The default value is None, which means that the normal behavior for sprite rendering is performed. If a Shader is attached to this attribute, the custom shader is used to render the sprite.

SpriteNode.size

The dimensions of the sprite, in points (width, height). When setting the texture attribute, the size is automatically set to the dimensions of the texture.

SpriteNode.texture

The Texture used to draw the sprite.

If the value is None, the sprite is drawn as a colored rectangle using its color attribute. Otherwise, the texture is used to draw the sprite.

EffectNode

class scene.EffectNode([position=(0, 0), z_position=0.0, scale=1.0, x_scale=1.0, y_scale=1.0, alpha=1.0, speed=1.0, parent=None])

An EffectNode object can be used to apply post-processing effects.

Each time a new frame is rendered using the effect node, the effect node follows these steps:

  • The effect node draws its children into a private framebuffer.
  • It blends the contents of its private framebuffer into its parent’s framebuffer, using one of the standard blend modes, similar to a SpriteNode. You can also use a custom Shader to render the result for more advanced post-processing effects.

EffectNode Attributes

EffectNode.crop_rect

A rectangle in the effect node’s coordinate system that determines how much of its children the effect node renders.

By default, an effect node automatically determines the size and position of its crop rect based on the accumulated frames of its children. In some cases, this may be inefficient, especially if the size of the children changes frequently, or if some of the children are off-screen.

EffectNode.blend_mode

The blend mode used to draw the filtered image into the parent’s framebuffer.

The possible values for this property are listed in Blend Modes. The default value is BLEND_NORMAL.

EffectNode.effects_enabled

When set to False, the effect node renders like a regular Node (shader, blend mode, and crop rect are ignored). The default is True.

EffectNode.shader

A custom shader that is called when the effect node is blended into the parent’s framebuffer.

The default value is None, meaning that default blending behavior executes. If a shader is specified, it is called when the rasterized image is blended into the parent’s framebuffer.

LabelNode

class scene.LabelNode(text, font=('Helvetica', 20), *args, **kwargs)

A LabelNode is a specialized subclass of SpriteNode that automatically generates a texture by rendering a string into an image, using the given font. When you set the LabelNode.text or LabelNode.font attributes, the texture is updated automatically.

By default, the text is centered on the node’s position. You can change this by adjusting the Node.anchor_point attribute.

LabelNode Attributes

LabelNode.text

The string value of the label.

LabelNode.font

The font of the label, as a 2-tuple of font name and font size (in points).

ShapeNode

class scene.ShapeNode(path=None, fill_color='white', stroke_color='clear', shadow=None, *args, **kwargs)

A ShapeNode is a specialized subclass of SpriteNode that renders a ui.Path (a vector shape).

The shape has customizable fill and stroke colors, and can optionally be rendered with a drop shadow.

ShapeNode Attributes

ShapeNode.path

A ui.Path object that represents the shape to be rendered. The ui.Path.line_width attribute determines the width of the shape’s outline (stroke).

ShapeNode.fill_color

The color that is used to fill the shape.

ShapeNode.stroke_color

The color that is used for the shape’s outline.

ShapeNode.shadow

A 4-tuple of (color, x_offset, y_offset, radius) that specifies an optional drop shadow. If set to None, no drop shadow is rendered.

Other Classes

SceneView

class scene.SceneView

SceneView is a subclass of ui.View that draws a Scene‘s content and implements a rendering loop.

Typically, you will not create a SceneView explicitly – calling the run() function does this automatically. If you need to modify the view’s attributes after running your scene this way, you can access its view using the Scene.view attribute.

For debugging purposes, a SceneView will show standard output and exceptions in a “status line” at the bottom, so you can use print statements during development, even if your scene covers the whole screen.

SceneView Attributes

SceneView.scene

The scene that is currently presented in the view. Without setting this attribute, the view will be empty.

SceneView.paused

Set this to True to pause the rendering loop.

SceneView.frame_interval

By default, the rendering loop updates 60 times per second, which corresponds to a frame interval of 1. Setting this to a higher value reduces the update frequency, e.g. a value of 2 corresponds to 30 frames per second. If you don’t create a SceneView explicitly, you can also pass this value as an argument to the run() function.

SceneView.anti_alias

Set this to True to enable 4x multisampling. This has a (sometimes significant) performance cost, and is disabled by default.

SceneView.shows_fps

Set to True to enable a debugging overlay that shows the current framerate.

Shader

class scene.Shader(shader_src)

A Shader object represents a custom OpenGL fragment shader that can be used to modify the rendering behavior of SpriteNode and EffectNode objects (via their respective shader attributes).

Shader programming is a complex topic, and a complete introduction is beyond the scope of this document, but the basic concept is really quite simple. A fragment shader is essentially a function/program that is executed directly on the GPU, and that is responsible for producing a color value for every pixel. Shaders are written in GLSL (GL Shading Language), which is very similar to C.

The default shader of a SpriteNode basically just uses its texture coordinate input (which is set automatically) to look up the corresponding color in its associated Texture. A simple custom shader might either adjust the resulting color value (for example, converting the colors to grayscale), or modify the texture coordinates to produce morphological effects.

Shaders have two kinds of inputs: Varyings and uniforms. Essentially, a uniform has the same value for every pixel/fragment, while a varying is interpolated. An example for a varying would be the texture coordinates, which are obviously different for every pixel. An example for a uniform would be a sprite’s size or the current timestamp.

A couple of varyings and uniforms are set automatically when a SpriteNode or EffectNode with a custom shader is rendered:

  • uniform float u_time – the current timestamp of the scene’s animation loop
  • uniform vec2 u_sprite_size – The size of the sprite (in points)
  • uniform float u_scale – The scale factor of the screen (typically 2.0 for retina screens) – this can be used to convert u_sprite_size to actual screen pixels.
  • uniform sampler2D u_texture – The texture of the sprite (for an EffectNode, this texture contains the rendering of its children)
  • uniform vec4 u_color – The premultiplied color of the sprite (corresponding to the SpriteNode.color attribute)
  • varying vec2 v_tex_coord – The current texture (UV) coordinates

Even though these uniforms and varyings are set automatically, you have to declare them in your shader if you want to use them.

You can also declare and set custom uniforms using the Shader.set_uniform() method. Custom uniforms can be textures (sampler2D), floats, and 2-/3-/4-component vectors (vec2/vec3/vec4).

The following example produces an interesting ‘ripple’ effect on the Pythonista icon. It also demonstrates how you can modify shader behavior by setting uniforms – in this case, the center of the ripple effect is changed in response to touch events.:

from scene import *

ripple_shader = '''
precision highp float;
varying vec2 v_tex_coord;
// These uniforms are set automatically:
uniform sampler2D u_texture;
uniform float u_time;
uniform vec2 u_sprite_size;
// This uniform is set in response to touch events:
uniform vec2 u_offset;

void main(void) {
    vec2 p = -1.0 + 2.0 * v_tex_coord + (u_offset / u_sprite_size * 2.0);
    float len = length(p);
    vec2 uv = v_tex_coord + (p/len) * 1.5 * cos(len*50.0 - u_time*10.0) * 0.03;
    gl_FragColor = texture2D(u_texture,uv);
}
'''

class MyScene (Scene):
    def setup(self):
            self.sprite = SpriteNode('test:Pythonista', parent=self)
            self.sprite.shader = Shader(ripple_shader)
            self.did_change_size()

    def did_change_size(self):
            # Center the image:
            self.sprite.position = self.size/2

    def touch_began(self, touch):
            self.set_ripple_center(touch)

    def touch_moved(self, touch):
            self.set_ripple_center(touch)

    def set_ripple_center(self, touch):
            # Center the ripple effect on the touch location by setting the `u_offset` shader uniform:
            dx, dy = self.sprite.position - touch.location
            self.sprite.shader.set_uniform('u_offset', (dx, dy))

run(MyScene())
Shader.get_uniform(name)

Return the current value of the uniform with the given name. Note that the uniform must be of type float, vec2, vec3 or vec4 (i.e. sampler/texture uniforms are not supported by this method). For invalid uniform names, None is returned.

Shader.set_uniform(name, value)

Set the uniform with the given name to a new value. The value may be a single number for float or int uniforms, a sequence of numbers for uniforms of type vec2, vec3 or vec4, or a Texture object for sampler2D uniforms. Other types of uniforms are not supported.

Action

class scene.Action

An Action object is an animation that usually changes a Node‘s attributes over time. It can be added to a node using its Node.run_action() method.

Different types of actions (for different Node attributes) are created using different class methods, e.g. Action.move_to(), Action.rotate_by() etc.

Once an action has been added to a Node, it cannot be changed anymore, but you can remove it (and stop the animation) using Node.remove_action() or Node.remove_all_actions().

Some actions modify the behavior of other actions:

An Action.sequence() has multiple child actions. Each action in the sequence begins after the previous action ends.

An Action.group() has multiple child actions. All actions stored in the group begin executing at the same time. The entire group finishes after all actions have finished. This is particularly useful in combination with Action.sequence().

An Action.repeat() action stores a single child action. When the child action completes, it is restarted.

Groups, sequences, and repeating actions can be nested. The ability to combine actions together allows you to add very sophisticated behaviors to a node.

The default duration of animated actions is 0.5 seconds.

classmethod Action.call(func[, duration])

Creates a custom action that calls a function (or other callable object).

If no duration parameter is passed, the function is called exactly once and must not take any parameters.

If the duration parameter is used, the node that the action is executed on, and the current progress (between 0.0 and 1.0) are passed to the function, and it must have the following signature:

def custom_action(node, progress):
    pass # do something with the node here...
classmethod Action.fade_by(alpha[, duration, timing_mode])

Creates an action that adjusts the alpha value of a node by a relative value.

When the action executes, the Node.alpha attribute animates to its new value.

classmethod Action.fade_to(alpha[, duration, timing_mode])

Creates an action that adjusts the alpha value of a node to a new value.

When the action executes, the Node.alpha attribute animates to its new value.

classmethod Action.group(actions...)

Creates an action that runs a collection of actions in parallel.

When the action executes, the actions that comprise the group all start immediately and run in parallel. The duration of the group action is the longest duration among the collection of actions. If an action in the group has a duration less than the group’s duration, the action completes, then idles until the group completes the remaining actions. This matters most when creating a repeating action that repeats a group.

You can pass the group’s actions as individual arguments or as a single sequence argument (e.g. a list).

classmethod Action.move_by(dx, dy[, duration, timing_mode])

Creates an action that moves a node relative to its current position.

classmethod Action.move_to(x, y[, duration, timing_mode])

Creates an action that moves a node to a new position.

classmethod Action.repeat(action, repeat_count)

Creates an action that repeats another action a specified number of times. If repeat_count is <=0, the action repeats forever (or until it is removed explicitly).

classmethod Action.rotate_by(radians[, duration, timing_mode])

Creates an action that rotates the node by a relative value.

classmethod Action.rotate_to(radians[, duration, timing_mode])

Creates an action that rotates the node counterclockwise to an absolute angle.

classmethod Action.scale_by(scale[, duration, timing_mode])

Creates an action that changes the x and y scale values of a node by a relative value.

classmethod Action.scale_to(scale[, duration, timing_mode])

Creates an action that changes the x and y scale values of a node.

classmethod Action.scale_x_to(scale[, duration, timing_mode])

Creates an action that changes the x scale value of a node.

classmethod Action.scale_y_to(scale[, duration, timing_mode])

Creates an action that changes the y scale value of a node.

classmethod Action.set_uniform(name, value[, duration, timing_mode])

Creates an action that animates the given shader uniform to a new value. Note that this action only has an effect on SpriteNode and EffectNode objects. The value must be either a single number (for float uniforms) or a sequence of 2-4 numbers (for vec2, vec3 and vec4 uniforms).

classmethod Action.sequence(actions...)

Creates an action that runs a collection of actions sequentially.

When the action executes, the first action in the sequence starts and runs to completion. Subsequent actions in the sequence run in a similar fashion until all of the actions in the sequence have executed. The duration of the sequence action is the sum of the durations of the actions in the sequence.

You can pass the sequence’s actions as individual arguments or as a single sequence argument (e.g. a list).

classmethod Action.wait(wait_duration)

Creates an action that idles for a specified period of time.

When the action executes, the action waits for the specified amount of time, then ends. This is typically used as part of a sequence of actions to insert a delay between two other actions.

Action Attributes

Action.duration

The duration of the action, in seconds. For some types of actions (e.g. groups, sequences), the duration is ignored. Note that changing the duration (or any other attributes) of an action after it has been added to a Node has no effect.

Action.timing_mode

The timing mode used to execute an action.

The possible values for this property are listed under Timing Modes. The default value is TIMING_LINEAR.

For some types of actions (e.g. groups, sequences), the timing mode is ignored. Note that changing the timing mode (or any other attributes) of an action after it has been added to a Node has no effect.

Texture

class scene.Texture(image)

Texture objects are used by SpriteNode objects to render their content. A texture is an image that has been loaded into the GPU’s memory. Textures can be initialized using either the name of a built-in image (a string), or a ui.Image object.

Texture.filtering_mode

The filtering mode that is used for scaling the texture. Can be one of the constants listed under Texture Filtering Modes (FILTERING_LINEAR by default).

Texture.size

The size of the texture in pixels. Note that this does not take the screen’s scale factor into account.

Touch

class scene.Touch(x, y, prev_x, prev_y, touch_id)
Instances of this class are passed to the Scene.touch_began(), Scene.touch_moved() and Scene.touch_ended() methods. Scene objects also have a touches attribute (a dictionary

that maps touch_id to Touch objects).

Touch.location

The current location of the touch as a Point object.

Touch.prev_location

The previous location of the touch (before it moved) as a Point object.

Geometry Types

Vector2

class scene.Vector2(x, y)

The Vector2 class is the base class for Point and Size.

Vectors (and subsequently points and sizes) support the addition (+), subtraction (-), multiplication (*) and division (/) operators. Addition and subtraction are supported for two vectors. Multiplication and division can also be applied to a vector and a scalar (number).

Passing a vector to the built-in abs() function calculates the length of a vector. This can be particularly useful in combination with subtraction, making it very easy to calculate the distance between two points, e.g. abs(p2 - p1).

It is also possible to determine whether a point lies within a Rect, using the in operator.

In most ways, Vector2 behaves like a sequence, similar to a 2-tuple. For example, you can alternatively access its x component using subscript notation (v[0]). Vectors also support iteration, argument unpacking, etc.

Whenever a Vector2, Point or Size object is used as an attribute in the scene module, you can alternatively provide any sequence of 4 numbers (e.g. a list or tuple).

Point

class scene.Point(x, y)

Subclass of Vector2, used to represent positions. This class doesn’t provide any additional methods or attributes.

Size

class scene.Size(w, h)

Subclass of Vector2, used to represent sizes. The only difference compared to Vector2 is that the x and y components can alternatively be accessed using Size.w (width), and Size.h (height).

Rect

class scene.Rect(x, y, w, h)

The Rect class is used for bounding boxes and other rectangle values, e.g. the Node.frame attribute. A rectangle is represented as (x, y, w[idth], h[eight]), with (x, y) being its lower-left corner.

In most ways, Rect behaves like a sequence, similar to a 4-tuple. For example, you can alternatively access its x component using subscript notation (r[0]). Rectangles also support iteration, argument unpacking, etc.

Whenever a Rect object is used as an attribute in the scene module, you can alternatively provide any sequence of 4 numbers (e.g. a list or tuple).

Rect.x

The x component of the rectangle’s lower-left corner.

Rect.y

The y component of the rectangle’s lower-left corner.

Rect.w
Rect.width

The width of the rectangle.

Rect.h
Rect.height

The height of the rectangle.

Rect.origin

Equivalent to Point(rect.x, rect.y)

Rect.size

Equivalent to Size(rect.w, rect.h)

Rect.min_x

Equivalent to min(rect.x, rect.x + rect.w) (the x component of the left edge)

Rect.max_x

Equivalent to max(rect.x, rect.x + rect.w) (the x component of the right edge)

Rect.min_y

Equivalent to min(rect.y, rect.y + rect.h) (the y component of the bottom edge)

Rect.max_y

Equivalent to max(rect.y, rect.y + rect.h) (the y component of the top edge)

Rect.center([p])

When called without arguments, return the center of the rectangle. When a Point is passed as an argument, the rectangle’s x and y values are adjusted, so that the new center of the rectangle is p.

Rect.contains_point(p)

Return True if the given point lies within the bounds of the rectangle, False otherwise.

Rect.contains_rect(other_rect)

Return True if the given rectangle lies entirely within the bounds of this rectangle, False otherwise.

Rect.intersects(other_rect)

Return True if this rectangle intersects with the other rectangle, False otherwise.

Rect.intersection(other_rect)

Return a Rect that corresponds to the intersection of this rectangle with the other one.

Rect.union(other_rect)

Return the smallest Rect that encloses both rectangles.

Rect.translate(x, y)

Equivalent to Rect(r.x + x, r.y + y, r.w, r.h)

Rect.inset(top, left[, bottom, right])

Return a rectangle that is adjusted by the given edge insets. bottom/right are optional and default to the same value as top/left.

Functions

scene.gravity()

Return the current gravity vector (x, y, z) – each component value will be between 0.0 and 1.0. This can be used to determine the current orientation of the device, typically for motion-controlled games.

scene.get_screen_size()

Return the screen size (in points) as a Size object. Note that the value is orientation-dependent – you may want to use min(get_screen_size()) to get the smallest dimension in any orientation.

scene.get_screen_scale()

Return the scale factor of the current device’s screen. For retina screens this will usually be 2.0 or 3.0, and 1.0 for non-retina screens.

scene.get_image_path(name)

Return the absolute path of the file that corresponds to a built-in image name.

scene.run(scene, orientation=DEFAULT_ORIENTATION, frame_interval=1, anti_alias=False, show_fps=False, multi_touch=True)

Runs the given Scene object. By default, the scene runs in the current device orientation, set the parameter to PORTRAIT or LANDSCAPE to force a specific orientation (and to disallow auto-rotation).

By default, the scene’s update() method is called 60 times per second. Set the frame_interval parameter to 2 for 30fps, 3 for 20, etc.

Constants

Orientations

The following constants are used for the orientation parameter of the run() function.

scene.DEFAULT_ORIENTATION

Start the scene in the current device orientation, allow auto-rotation.

scene.PORTRAIT

Force portrait orientation.

scene.LANDSCAPE

Force landscape orientation.

Blend Modes

scene.BLEND_NORMAL

The source and destination colors are blended by multiplying the source alpha value.

scene.BLEND_ADD

The source and destination colors are added together.

scene.BLEND_MULTIPLY

The source color is multiplied by the destination color.

Texture Filtering Modes

The following constants can be used for the Texture.filtering_mode attribute.

scene.FILTERING_LINEAR

Linear interpolation, the default filtering mode

scene.FILTERING_NEAREST

Nearest-neighbor interpolation – this is the recommended mode for pixel art because it results in sharp edges when scaling up low-resolution artwork.

Timing Modes

The following constants can be used for the timing_mode attribute of Action objects to determine their interpolation function.

scene.TIMING_LINEAR

Simple linear interpolation (the default).

scene.TIMING_EASE_IN

Ease-in interpolation

scene.TIMING_EASE_IN_2

Alternative ease-in interpolation (quadratic)

scene.TIMING_EASE_OUT

Ease-out interpolation

scene.TIMING_EASE_OUT_2

Alternative ease-out interpolation (quadratic)

scene.TIMING_EASE_IN_OUT

Ease-in-ease-out interpolation.

scene.TIMING_SINODIAL

Similar to TIMING_EASE_IN_OUT, but with slightly less pronounced easing phases.

scene.TIMING_ELASTIC_IN

“Rubberband” effect at the start of the animation.

scene.TIMING_ELASTIC_OUT

“Rubberband” effect at the end of the animation.

scene.TIMING_ELASTIC_IN_OUT

“Rubberband” effect at both ends of the animation.

scene.TIMING_BOUNCE_IN

“Bouncing” effect at the start of the animation.

scene.TIMING_BOUNCE_OUT

“Bouncing” effect at the end of the animation.

scene.TIMING_BOUNCE_IN_OUT

“Bouncing” effect at both ends of the animation.

scene.TIMING_EASE_BACK_IN

Overshooting animation, eased at the start.

scene.TIMING_EASE_BACK_OUT

Overshooting animation, eased at the end.

scene.TIMING_EASE_BACK_IN_OUT

Overshooting animation, eased at the start and end.