react-native/docs/Animations.md

id	title	layout	category	permalink	next
animations	Animations	docs	Guides	docs/animations.html	accessibility

Fluid, meaningful animations are essential to the mobile user experience. Like everything in React Native, Animation APIs for React Native are currently under development, but have started to coalesce around two complementary systems: LayoutAnimation for animated global layout transactions, and Animated for more granular and interactive control of specific values.

Animated

The Animated library is designed to make it very easy to concisely express a wide variety of interesting animation and interaction patterns in a very performant way. Animated focuses on declarative relationships between inputs and outputs, with configurable transforms in between, and simple start/stop methods to control time-based animation execution. For example, a complete component with a simple spring bounce on mount looks like this:

class Playground extends React.Component {
  constructor(props: any) {
    super(props);
    this.state = {
      bounceValue: new Animated.Value(0),
    };
  }
  render(): ReactElement {
    return (
      <Animated.Image                         // Base: Image, Text, View
        source={{uri: 'http://i.imgur.com/XMKOH81.jpg'}}
        style={{
          flex: 1,
          transform: [                        // `transform` is an ordered array
            {scale: this.state.bounceValue},  // Map `bounceValue` to `scale`
          ]
        }}
      />
    );
  }
  componentDidMount() {
    this.state.bounceValue.setValue(1.5);     // Start large
    Animated.spring(                          // Base: spring, decay, timing
      this.state.bounceValue,                 // Animate `bounceValue`
      {
        toValue: 0.8,                         // Animate to smaller size
        friction: 1,                          // Bouncier spring
      }
    ).start();                                // Start the animation
  }
}

bounceValue is initialized as part of state in the constructor, and mapped to the scale transform on the image. Behind the scenes, the numeric value is extracted and used to set scale. When the component mounts, the scale is set to 1.5 and then a spring animation is started on bounceValue which will update all of its dependent mappings on each frame as the spring animates (in this case, just the scale). This is done in an optimized way that is faster than calling setState and re-rendering. Because the entire configuration is declarative, we will be able to implement further optimizations that serialize the configuration and runs the animation on a high-priority thread.

Core API

Most everything you need hangs directly off the Animated module. This includes two value types, Value for single values and ValueXY for vectors, three animation types, spring, decay, and timing, and three component types, View, Text, and Image. You can make any other component animated with Animated.createAnimatedComponent.

The three animation types can be used to create almost any animation curve you want because each can be customized:

• spring: Simple single-spring physics model that matches Origami.
• friction: Controls "bounciness"/overshoot. Default 7.
• tension: Controls speed. Default 40.
• decay: Starts with an initial velocity and gradually slows to a stop.
• velocity: Initial velocity. Required.
• deceleration: Rate of decay. Default 0.997.
• timing: Maps time range to easing value.
• duration: Length of animation (milliseconds). Default 500.
• easing: Easing function to define curve. See Easing module for several predefined functions. iOS default is Easing.inOut(Easing.ease).
• delay: Start the animation after delay (milliseconds). Default 0.

Animations are started by calling start. start takes a completion callback that will be called when the animation is done. If the animation is done because it finished running normally, the completion callback will be invoked with {finished: true}, but if the animation is done because stop was called on it before it could finish (e.g. because it was interrupted by a gesture or another animation), then it will receive {finished: false}.

Composing Animations

Animations can also be composed with parallel, sequence, stagger, and delay, each of which simply take an array of animations to execute and automatically calls start/stop as appropriate. For example:

Animated.sequence([            // spring to start and twirl after decay finishes
  Animated.decay(position, {   // coast to a stop
    velocity: {x: gestureState.vx, y: gestureState.vy}, // velocity from gesture release
    deceleration: 0.997,
  }),
  Animated.parallel([          // after decay, in parallel:
    Animated.spring(position, {
      toValue: {x: 0, y: 0}    // return to start
    }),
    Animated.timing(twirl, {   // and twirl
      toValue: 360,
    }),
  ]),
]).start();                    // start the sequence group

By default, if one animation is stopped or interrupted, then all other animations in the group are also stopped. Parallel has a stopTogether option that can be set to false to disable this.

Interpolation

Another powerful part of the Animated API is the interpolate function. It allows input ranges to map to different output ranges. For example, a simple mapping to convert a 0-1 range to a 0-100 range would be

value.interpolate({
  inputRange: [0, 1],
  outputRange: [0, 100],
});

interpolate supports multiple range segments as well, which is handy for defining dead zones and other handy tricks. For example, to get an negation relationship at -300 that goes to 0 at -100, then back up to 1 at 0, and then back down to zero at 100 followed by a dead-zone that remains at 0 for everything beyond that, you could do:

value.interpolate({
  inputRange: [-300, -100, 0, 100, 101],
  outputRange: [300,    0, 1,   0,   0],
});

Which would map like so:

Input	Output
-400	450
-300	300
-200	150
-100	0
-50	0.5
0	1
50	0.5
100	0
101	0
200	0

interpolation also supports arbitrary easing functions, many of which are already implemented in the Easing class including quadradic, exponential, and bezier curves as well as functions like step and bounce. interpolation also has configurable behavior for extrapolation, the default being 'extend', but 'clamp' is also very useful to prevent the output value from exceeding outputRange.

Tracking Dynamic Values

Animated values can also track other values. Just set the toValue of an animation to another animated value instead of a plain number, for example with spring physics for an interaction like "Chat Heads", or via timing with duration: 0 for rigid/instant tracking. They can also be composed with interpolations:

Animated.spring(follower, {toValue: leader}).start();
Animated.timing(opacity, {
  toValue: pan.x.interpolate({
    inputRange: [0, 300],
    outputRange: [1, 0],
  }),
}).start();

ValueXY is a handy way to deal with 2D interactions, such as panning/dragging. It is a simple wrapper that basically just contains two Animated.Value instances and some helper functions that call through to them, making ValueXY a drop-in replacement for Value in many cases. For example, in the code snippet above, leader and follower could both be of type ValueXY and the x and y values will both track as you would expect.

Input Events

Animated.event is the input side of the Animated API, allowing gestures and other events to map directly to animated values. This is done with a structured map syntax so that values can be extracted from complex event objects. The first level is an array to allow mapping across multiple args, and that array contains nested objects. In the example, you can see that scrollX maps to event.nativeEvent.contentOffset.x (event is normally the first arg to the handler), and pan.x and pan.y map to gestureState.dx and gestureState.dy, respectively (gestureState is the second arg passed to the PanResponder handler).

onScroll={Animated.event(
  [{nativeEvent: {contentOffset: {x: scrollX}}}]   // scrollX = e.nativeEvent.contentOffset.x
)}
onPanResponderMove={Animated.event([
  null,                                          // ignore the native event
  {dx: pan.x, dy: pan.y}                         // extract dx and dy from gestureState
]);

Responding to the Current Animation Value

You may notice that there is no obvious way to read the current value while animating - this is because the value may only be known in the native runtime due to optimizations. If you need to run JavaScript in response to the current value, there are two approaches:

• spring.stopAnimation(callback) will stop the animation and invoke callback with the final value - this is useful when making gesture transitions.
• spring.addListener(callback) will invoke callback asynchronously while the animation is running, providing a recent value. This is useful for triggering state changes, for example snapping a bobble to a new option as the user drags it closer, because these larger state changes are less sensitive to a few frames of lag compared to continuous gestures like panning which need to run at 60fps.

Future Work

As previously mentioned, we're planning on optimizing Animated under the hood to make it even more performant. We would also like to experiment with more declarative and higher level gestures and triggers, such as horizontal vs. vertical panning.

The above API gives a powerful tool for expressing all sorts of animations in a concise, robust, and performant way. Check out more example code in UIExplorer/AnimationExample. Of course there may still be times where Animated doesn't support what you need, and the following sections cover other animation systems.

LayoutAnimation

LayoutAnimation allows you to globally configure create and update animations that will be used for all views in the next render/layout cycle. This is useful for doing flexbox layout updates without bothering to measure or calculate specific properties in order to animate them directly, and is especially useful when layout changes may affect ancestors, for example a "see more" expansion that also increases the size of the parent and pushes down the row below which would otherwise require explicit coordination between the components in order to animate them all in sync.

Note that although LayoutAnimation is very powerful and can be quite useful, it provides much less control than Animated and other animation libraries, so you may need to use another approach if you can't get LayoutAnimation to do what you want.

var App = React.createClass({
  componentWillMount() {
    // Animate creation
    LayoutAnimation.spring();
  },

  getInitialState() {
    return { w: 100, h: 100 }
  },

  _onPress() {
    // Animate the update
    LayoutAnimation.spring();
    this.setState({w: this.state.w + 15, h: this.state.h + 15})
  },

  render: function() {
    return (
      <View style={styles.container}>
        <View style={[styles.box, {width: this.state.w, height: this.state.h}]} />
        <TouchableOpacity onPress={this._onPress}>
          <View style={styles.button}>
            <Text style={styles.buttonText}>Press me!</Text>
          </View>
        </TouchableOpacity>
      </View>
    );
  }
});

Run this example

This example uses a preset value, you can customize the animations as you need, see LayoutAnimation.js for more information.

requestAnimationFrame

requestAnimationFrame is a polyfill from the browser that you might be familiar with. It accepts a function as its only argument and calls that function before the next repaint. It is an essential building block for animations that underlies all of the JavaScript-based animation APIs. In general, you shouldn't need to call this yourself - the animation API's will manage frame updates for you.

react-tween-state (Not recommended - use Animated instead)

react-tween-state is a minimal library that does exactly what its name suggests: it tweens a value in a component's state, starting at a from value and ending at a to value. This means that it generates the values in between those two values, and it sets the state on every requestAnimationFrame with the intermediary value.

Tweening definition from Wikipedia

"... tweening is the process of generating intermediate frames between two images to give the appearance that the first image evolves smoothly into the second image. [Tweens] are the drawings between the key frames which help to create the illusion of motion."

The most obvious way to animate from one value to another is linearly: you subtract the end value from the start value and divide the result by the number of frames over which the animation occurs, and then add that value to the current value on each frame until the end value is reached. Linear easing often looks awkward and unnatural, so react-tween-state provides a selection of popular easing functions that can be applied to make your animations more pleasing.

This library does not ship with React Native - in order to use it on your project, you will need to install it with npm i react-tween-state --save from your project directory.

var tweenState = require('react-tween-state');

var App = React.createClass({
  mixins: [tweenState.Mixin],

  getInitialState() {
    return { opacity: 1 }
  },

  _animateOpacity() {
    this.tweenState('opacity', {
      easing: tweenState.easingTypes.easeOutQuint,
      duration: 1000,
      endValue: this.state.opacity === 0.2 ? 1 : 0.2,
    });
  },

  render() {
    return (
      <View style={{flex: 1, justifyContent: 'center', alignItems: 'center'}}>
        <TouchableWithoutFeedback onPress={this._animateOpacity}>
          <View ref={component => this._box = component}
                style={{width: 200, height: 200, backgroundColor: 'red',
                        opacity: this.getTweeningValue('opacity')}} />
        </TouchableWithoutFeedback>
      </View>
    )
  },
});

Run this example

Here we animated the opacity, but as you might guess, we can animate any numeric value. Read more about react-tween-state in its README.

Rebound (Not recommended - use Animated instead)

Rebound.js is a JavaScript port of Rebound for Android. It is similar in concept to react-tween-state: you have an initial value and set an end value, then Rebound generates intermediate values that you can use for your animation. Rebound is modeled after spring physics; we don't provide a duration when animating with springs, it is calculated for us depending on the spring tension, friction, current value and end value. Rebound is used internally by React Native on Navigator and WarningBox.

Notice that Rebound animations can be interrupted - if you release in the middle of a press, it will animate back from the current state to the original value.

var rebound = require('rebound');

var App = React.createClass({
  // First we initialize the spring and add a listener, which calls
  // setState whenever it updates
  componentWillMount() {
    // Initialize the spring that will drive animations
    this.springSystem = new rebound.SpringSystem();
    this._scrollSpring = this.springSystem.createSpring();
    var springConfig = this._scrollSpring.getSpringConfig();
    springConfig.tension = 230;
    springConfig.friction = 10;

    this._scrollSpring.addListener({
      onSpringUpdate: () => {
        this.setState({scale: this._scrollSpring.getCurrentValue()});
      },
    });

    // Initialize the spring value at 1
    this._scrollSpring.setCurrentValue(1);
  },

  _onPressIn() {
    this._scrollSpring.setEndValue(0.5);
  },

  _onPressOut() {
    this._scrollSpring.setEndValue(1);
  },

  render: function() {
    var imageStyle = {
      width: 250,
      height: 200,
      transform: [{scaleX: this.state.scale}, {scaleY: this.state.scale}],
    };

    var imageUri = "https://facebook.github.io/react-native/img/ReboundExample.png";

    return (
      <View style={styles.container}>
        <TouchableWithoutFeedback onPressIn={this._onPressIn}
                                  onPressOut={this._onPressOut}>
          <Image source={{uri: imageUri}} style={imageStyle} />
        </TouchableWithoutFeedback>
      </View>
    );
  }
});

Run this example

You can also clamp the spring values so that they don't overshoot and oscillate around the end value. In the above example, we would add this._scrollSpring.setOvershootClampingEnabled(true) to change this. See the below gif for an example of where in your interface you might use this.

Screenshot from react-native-scrollable-tab-view. You can run a simlar example here.

A sidenote about setNativeProps

As mentioned in the Direction Manipulation section, setNativeProps allows us to modify properties of native-backed components (components that are actually backed by native views, unlike composite components) directly, without having to setState and re-render the component hierarchy.

We could use this in the Rebound example to update the scale - this might be helpful if the component that we are updating is deeply nested and hasn't been optimized with shouldComponentUpdate.

// Outside of our React component
var precomputeStyle = require('precomputeStyle');

// Back inside of the App component, replace the scrollSpring listener
// in componentWillMount with this:
this._scrollSpring.addListener({
  onSpringUpdate: () => {
    if (!this._photo) { return }
    var v = this._scrollSpring.getCurrentValue();
    var newProps = precomputeStyle({transform: [{scaleX: v}, {scaleY: v}]});
    this._photo.setNativeProps(newProps);
  },
});

// Lastly, we update the render function to no longer pass in the
// transform via style (avoid clashes when re-rendering) and to set the
// photo ref
render: function() {
  return (
    <View style={styles.container}>
      <TouchableWithoutFeedback onPressIn={this._onPressIn} onPressOut={this._onPressOut}>
        <Image ref={component => this._photo = component}
               source={{uri: "https://facebook.github.io/react-native/img/ReboundExample.png"}}
               style={{width: 250, height: 200}} />
      </TouchableWithoutFeedback>
    </View>
  );
}

Run this example

It would not make sense to use setNativeProps with react-tween-state because the updated tween values are set on the state automatically by the library - Rebound on the other hand gives us an updated value for each frame with the onSpringUpdate function.

If you find your animations with dropping frames (performing below 60 frames per second), look into using setNativeProps or shouldComponentUpdate to optimize them. You may also want to defer any computationally intensive work until after animations are complete, using the InteractionManager. You can monitor the frame rate by using the In-App Developer Menu "FPS Monitor" tool.

Navigator Scene Transitions

As mentioned in the Navigator Comparison, Navigator is implemented in JavaScript and NavigatorIOS is a wrapper around native functionality provided by UINavigationController, so these scene transitions apply only to Navigator. In order to re-create the various animations provided by UINavigationController and also make them customizable, React Native exposes a NavigatorSceneConfigs API.

var SCREEN_WIDTH = require('Dimensions').get('window').width;
var BaseConfig = Navigator.SceneConfigs.FloatFromRight;

var CustomLeftToRightGesture = Object.assign({}, BaseConfig.gestures.pop, {
  // Make it snap back really quickly after canceling pop
  snapVelocity: 8,

  // Make it so we can drag anywhere on the screen
  edgeHitWidth: SCREEN_WIDTH,
});

var CustomSceneConfig = Object.assign({}, BaseConfig, {
  // A very tighly wound spring will make this transition fast
  springTension: 100,
  springFriction: 1,

  // Use our custom gesture defined above
  gestures: {
    pop: CustomLeftToRightGesture,
  }
});

Run this example

For further information about customizing scene transitions, read the source.

AnimationExperimental (Deprecated)

As the name would suggest, this was only ever an experimental API and it is not recommended to use this on your apps. It has some rough edges and is not under active development. It is built on top of CoreAnimation explicit animations.

If you choose to use it anyways, here is what you need to know:

• You will need to include RCTAnimationExperimental.xcodeproj and add libRCTAnimationExperimental.a to Build Phases.
• Suited only for static "fire and forget" animations - not continuous gestures.
• Hit detection will not work as expected because animations occur on the presentation layer.

var AnimationExperimental = require('AnimationExperimental');

var App = React.createClass({
  componentDidMount() {
    AnimationExperimental.startAnimation(
      {
        node: this._box,
        duration: 1000,
        easing: 'easeInOutBack',
        property: 'scaleXY',
        toValue: { x: 1, y: 1 },
      },
    );
  },

  render() {
    return (
      <View style={{flex: 1, justifyContent: 'center', alignItems: 'center'}}>
        <View ref={component => this._box = component}
              style={{width: 200, height: 200, backgroundColor: 'red'}} />
      </View>
    )
  },
});

Now to demonstrate a known issue, and one of the reasons why it is recommended not to use AnimationExperimental currently, let's try to animate opacity from 1 to 0.5:

AnimationExperimental.startAnimation(
  {
    node: this._box,
    duration: 1000,
    easing: 'easeInOutBack',
    property: 'opacity',
    fromValue: 1,
    toValue: 0.5,
  },
);

Pop (Unsupported, not recommended)

Facebook Pop "supports spring and decay dynamic animations, making it useful for building realistic, physics-based interactions."

This is not officially supported or recommended because the direction is to move towards JavaScript-driven animations, but if you must use it, you can find the code to integrate with React Native here. Please do not open questions specific to Pop on the React Native issues, StackOverflow is a better place to answer those questions as it is not considered to be part of the core.