Jetpack Compose版来啦！高仿微信朋友圈大图缩放、切换、预览功能

最近在学习Jetpack Compose，想着能否用Jetpack Compose实现微信一些重要界面以及功能。好消息是已经实现了微信聊天界面相关功能以及交互，最近又搞了搞朋友圈的整体交互，网上看了看，关于compose动画相关知识比较少，所以打算通过最近学习的compose手势动画相关知识实现该功能。

本文主要讲述如何通过compose手势动画实现微信大图缩放、切换、预览功能。

先上动图

在实现上述功能时首先我们需要了解一下 Compose 为我们提供的一些手势动画。

使用`PointerInput Modifier`

对于所有手势操作的处理都需要封装到这个 Modifier 中，我们知道 Modifier 时用来修饰 UI 组件的，所以将手势操作的处理封装在 Modifier 符合开发者设计直觉，这同时也做到了手势处理逻辑与 UI 视图的解耦，从而提高复用性。

Modifier 为我们提供了很多手势事件，比如：Transformer Modifier、Draggable Modifier、Rotation Modifier以及滚动事件、点击事件等等都能看到PointerInput Modifier的身影。因为这类上层的手势处理 Modifier 都是基于基础Modifier.pointInput()来实现的，所以自定义手势必然要在这个 Modifier 中进行。

//Transformer Modifier
fun Modifier.transformable(
    state: TransformableState,
    lockRotationOnZoomPan: Boolean = false,
    enabled: Boolean = true
) = composed(
    factory = {
        ...
        if (enabled) Modifier.pointerInput(Unit, block) else Modifier
    },
)

//Draggable Modifier
internal fun Modifier.draggable(
    stateFactory: @Composable () -> PointerAwareDraggableState,
    canDrag: (PointerInputChange) -> Boolean,
    orientation: Orientation,
    enabled: Boolean = true,
    interactionSource: MutableInteractionSource? = null,
    startDragImmediately: () -> Boolean,
    onDragStarted: suspend CoroutineScope.(startedPosition: Offset) -> Unit = {},
    onDragStopped: suspend CoroutineScope.(velocity: Float) -> Unit = {},
    reverseDirection: Boolean = false
): Modifier = composed(
) {
    ...
    Modifier.pointerInput(orientation, enabled, reverseDirection) {
       ...
    }
}

通过 PointerInput Modifier 实现我们可以看出，我们所定义的自定义手势处理流程均发生在 PointerInputScope 中，suspend 关键字也告知我们自定义手势处理流程是发生在协程中。这其实是无可厚非的，在探索重组工作原理的过程中我们也经常能够看到协程的身影。

fun Modifier.pointerInput(
    key1: Any?,
    block: suspend PointerInputScope.() -> Unit
): Modifier = composed(
    inspectorInfo = debugInspectorInfo {
        name = "pointerInput"
        properties["key1"] = key1
        properties["block"] = block
    }
) {
    val density = LocalDensity.current
    val viewConfiguration = LocalViewConfiguration.current
    remember(density) { SuspendingPointerInputFilter(viewConfiguration, density) }.apply {
        val filter = this
        LaunchedEffect(this, key1) {
            filter.coroutineScope = this
            block()
        }
    }
}

接下来我们重点看看 PointerInputScope作用域，本文将着重解释一下部分我们用到的API，有想了解更全的可以移步大神文章：# 使用Jetpack Compose完成自定义手势处理。

点击类型基础 API

API介绍

API名称	作用
detectTapGestures	监听点击手势

我们知道，Clickable Modifier是compose给我们提供的单击事件，与 Clickable Modifier 不同的是，detectTapGestures 可以监听更多的点击事件。作为手机监听的基础 API，必然不会存在 Clickable Modifier 所拓展的涟漪效果。

detectTapGestures包括四个函数回调，分别为：

onDoubleTap (可选)：双击时回调
onLongPress (可选)：长按时回调
onPress (可选)：按下时回调
onTap (可选)：轻触时回调

suspend fun PointerInputScope.detectTapGestures(
    onDoubleTap: ((Offset) -> Unit)? = null,
    onLongPress: ((Offset) -> Unit)? = null,
    onPress: suspend PressGestureScope.(Offset) -> Unit = NoPressGesture,
    onTap: ((Offset) -> Unit)? = null
) = coroutineScope {
    ...
}

💡 Tips

onPress 普通按下事件

onDoubleTap 前必定会先回调 2 次 Press

onLongPress 前必定会先回调 1 次 Press（时间长）

onTap 前必定会先回调 1 次 Press（时间短）

例子如下：

@Composable
fun TapGestureSample() {
    var boxSize = 100.dp
    Box(
        Modifier.fillMaxSize()
    ) {
        Text(text = "detectTapGestures\t监听点击手势", fontSize = 30.sp)

        Text(
            text = "",
            fontSize = 16.sp,
            modifier = Modifier.align(Alignment.BottomCenter)
        )

        Box(
            Modifier
                .size(boxSize)
                .align(Alignment.Center)
                .background(Color.Green)
                .pointerInput(Unit) {
                    detectTapGestures(
                        onDoubleTap = { offset: Offset ->
                            //双击时回调
                            println("detectTapGestures obDoubleTap[双击时回调] offset:$offset")
                        },
                        onLongPress = { offset: Offset ->
                            //长按时回调
                            println("detectTapGestures onLongPress[长按时回调] offset:$offset")
                        },
                        onPress = { offset: Offset ->
                            //按下时回调
                            println("detectTapGestures onPress[按下时回调] offset:$offset")
                        },
                        onTap = { offset: Offset ->
                            //轻触时回调
                            println("detectTapGestures onTap[轻触时回调] offset:$offset")
                        }
                    )
                }
        )
    }
}

将上述例子运行一下就明白了，此处就不录gif了。

手势检测

transformable 修饰符

接下来我们通过rememberTransformableState检测用于平移、缩放和旋转的多点触控手势，我们可以使用transformable修饰符。此修饰符本身不会转换元素，只会检测手势。

rememberTransformableState内部是通过协程作用域来事实检测触控手势改变的。

例子如下：

@Composable
fun TransformableSample() {
    // set up all transformation states
    var scale by remember { mutableStateOf(1f) }
    var rotation by remember { mutableStateOf(0f) }
    var offset by remember { mutableStateOf(Offset.Zero) }
    val state = rememberTransformableState { zoomChange, offsetChange, rotationChange ->
        scale *= zoomChange
        rotation += rotationChange
        offset += offsetChange
    }
    Box(
        Modifier
            // apply other transformations like rotation and zoom
            // on the pizza slice emoji
            .graphicsLayer(
                scaleX = scale,
                scaleY = scale,
                rotationZ = rotation,
                translationX = offset.x,
                translationY = offset.y
            )
            // add transformable to listen to multitouch transformation events
            // after offset
            .transformable(state = state)
            .background(Color.Blue)
            .fillMaxSize()
    )
}

需要部分如下依赖

def accompanist_version = "0.24.3-alpha"
//compose的viewpager库
implementation "com.google.accompanist:accompanist-pager:$accompanist_version"  

//指示器
implementation "com.google.accompanist:accompanist-pager-indicators:$accompanist_version"

//CoilImage是google推荐我们去使用的加载网络图片的开源库
implementation "com.google.accompanist:accompanist-coil:0.13.0"

功能实现

我们回到我们的项目中，如上图所示，我们拆分一下该功能的实现。

1: 实现图片横向水平滚动HorizontalPager
2: 底部的水平切换的指示器:HorizontalPagerIndicator
3: 双击放大和缩小
4: 双指缩放
5: 图片如有放大，切换时放大图还原至原始大小

HorizontalPager

HorizontalPager 是其中一种布局，他将所有子项摆放在一条水平行上，允许用户在子项之间水平滑动。

ccdf5c9e05ce4baa9e89de02bc43670a_tplv-k3u1fbpfcp-zoom-in-crop-mark_1304_0_0_0.awebp

/**
 * 界面状态变更
 */
val pageState = rememberPagerState(initialPage = currentIndex)

HorizontalPager(
    count = 图片数量,
    state = pageState, //图片状态
    contentPadding = PaddingValues(horizontal = 0.dp), //图片间的间距
    modifier = Modifier.fillMaxSize()
) { page ->
    println("ImageBrowserItem current page: $page")
    ImageBrowserItem(images[page], page, this)
}

如果你想跳转到某一个特定页面，你可以在 CoroutineScope 中选择使用 rememberPagerState(initialPage = currentIndex)或 pagerState.scrollToPage(index) 、 pagerState.animateScrollToPage(index) 选一即可。

HorizontalPagerIndicator

HorizontalPagerIndicator用来标识 HorizontalPager 或 VerticalPager 的水平布局指示器，表示当前活动页面和使用 Shape 绘制的总页面。需要通过pageState绑定。

HorizontalPagerIndicator(
    pagerState = pageState, //需要通过pageState绑定
    activeColor = Color.White,
    inactiveColor = WeComposeTheme.colors.onBackground,
    modifier = Modifier
        .align(Alignment.BottomCenter)
        .padding(60.dp)
)

双击放大和缩小

对于我们要实现的双击事件来说，当双击时获取到已经缩放的scale，，则将当前图片缩放至原始图的两倍，也就是双击放大两倍，再次双击还原到原图大小，并且偏移量Offset恢复到中心点位置。如下部分代码：

...
Modifier.pointerInput(Unit) {
    detectTapGestures(
        onDoubleTap = {
            println("ImageBrowserItem detectTapGestures onDoubleTap offset: $it")
            scale = if (scale <= 1f) {
                2f
            } else {
                1f
            }
            offset = Offset.Zero
        },
        onTap = {

        }
    )

双指缩放

对于我们要实现的双指缩放来说，我们只需要处理缩放大小即可。当我们监听rememberTransformableState变换时，scale放大的5倍时就停止继续放大。

如下部分代码：

/**
 * 监听手势状态变换
 */
var state =
    rememberTransformableState(onTransformation = { zoomChange, panChange, rotationChange ->
        scale = (zoomChange * scale).coerceAtLeast(1f)
        scale = if (scale > 5f) {
            5f
        } else {
            scale
        }
        println("ImageBrowserItem detectTapGestures rememberTransformableState scale: $scale")
    })
    ...
Modifier
    .transformable(state = state)
    .graphicsLayer{  //布局缩放、旋转、移动变换
        scaleX = scale
        scaleY = scale
        translationX = offset.x
        translationY = offset.y
    }

切换恢复图片大小

在 pager 组件的 content scope 中允许开发者很轻松地拿到 currentPage 与 currentPageOffset 引用。可以使用这些值来计算效果。我们提供了 calculateCurrentOffsetForPage() 扩展函数去计算某一个特定页面的偏移量。

例子如下：

@OptIn(ExperimentalPagerApi::class)
@Composable
private fun Sample() {
    Scaffold(
        topBar = {
            TopAppBar(
                title = { Text(stringResource(R.string.horiz_pager_with_transition_title)) },
                backgroundColor = MaterialTheme.colors.surface,
            )
        },
        modifier = Modifier.fillMaxSize()
    ) { padding ->
        HorizontalPagerWithOffsetTransition(Modifier.padding(padding))
    }
}

@OptIn(ExperimentalPagerApi::class, ExperimentalCoilApi::class)
@Composable
fun HorizontalPagerWithOffsetTransition(modifier: Modifier = Modifier) {
    HorizontalPager(
        count = 10,
        // Add 32.dp horizontal padding to 'center' the pages
        contentPadding = PaddingValues(horizontal = 32.dp),
        modifier = modifier.fillMaxSize()
    ) { page ->
        Card(
            Modifier
                .graphicsLayer {
                    // Calculate the absolute offset for the current page from the
                    // scroll position. We use the absolute value which allows us to mirror
                    // any effects for both directions
                    val pageOffset = calculateCurrentOffsetForPage(page).absoluteValue

                    // We animate the scaleX + scaleY, between 85% and 100%
                    lerp(
                        start = 0.85f,
                        stop = 1f,
                        fraction = 1f - pageOffset.coerceIn(0f, 1f)
                    ).also { scale ->
                        scaleX = scale
                        scaleY = scale
                    }

                    // We animate the alpha, between 50% and 100%
                    alpha = lerp(
                        start = 0.5f,
                        stop = 1f,
                        fraction = 1f - pageOffset.coerceIn(0f, 1f)
                    )
                }
                .fillMaxWidth()
                .aspectRatio(1f)
        ) {
            Box {
                Image(
                    painter = rememberImagePainter(
                        data = rememberRandomSampleImageUrl(width = 600),
                    ),
                    contentDescription = null,
                    modifier = Modifier.fillMaxSize(),
                )

                ProfilePicture(
                    Modifier
                        .align(Alignment.BottomCenter)
                        .padding(16.dp)
                        // We add an offset lambda, to apply a light parallax effect
                        .offset {
                            // Calculate the offset for the current page from the
                            // scroll position
                            val pageOffset =
                                this@HorizontalPager.calculateCurrentOffsetForPage(page)
                            // Then use it as a multiplier to apply an offset
                            IntOffset(
                                x = (36.dp * pageOffset).roundToPx(),
                                y = 0
                            )
                        }
                )
            }
        }
    }
}

@OptIn(ExperimentalCoilApi::class)
@Composable
private fun ProfilePicture(modifier: Modifier = Modifier) {
    Card(
        modifier = modifier,
        shape = CircleShape,
        border = BorderStroke(4.dp, MaterialTheme.colors.surface)
    ) {
        Image(
            painter = rememberImagePainter(rememberRandomSampleImageUrl()),
            contentDescription = null,
            modifier = Modifier.size(72.dp),
        )
    }
}

我们可以在pager切换时通过calculateCurrentOffsetForPage(page).absoluteValue拿到当前pager的偏移量，当pageOffet == 1.0f 时证明pager已切换至下一页，此时我们恢复scale = 1f到原始大小即可，部分代码如下：

 Modifier
    .transformable(state = state)
    .graphicsLayer{  //布局缩放、旋转、移动变换
        scaleX = scale
        scaleY = scale
        translationX = offset.x
        translationY = offset.y

        val pageOffset = pagerScope.calculateCurrentOffsetForPage(page = page).absoluteValue
        if (pageOffset == 1.0f) {
            scale = 1.0f
        }
        println("ImageBrowserItem pagerScope calculateCurrentOffsetForPage pageOffset: $pageOffset")
    }

到这里我们就整个实现了大图缩放、切换、预览功能，完整代码如下:

import androidx.compose.foundation.Image
import androidx.compose.foundation.gestures.*
import androidx.compose.foundation.layout.*
import androidx.compose.material.Surface
import androidx.compose.material.swipeable
import androidx.compose.runtime.*
import androidx.compose.ui.Alignment
import androidx.compose.ui.Modifier
import androidx.compose.ui.draw.rotate
import androidx.compose.ui.geometry.Offset
import androidx.compose.ui.graphics.Color
import androidx.compose.ui.graphics.graphicsLayer
import androidx.compose.ui.input.pointer.pointerInput
import androidx.compose.ui.layout.ContentScale
import androidx.compose.ui.tooling.preview.Preview
import androidx.compose.ui.unit.dp
import androidx.lifecycle.viewmodel.compose.viewModel
import com.eegets.wechatcompose.ui.find.model.ImageBrowserModel
import com.eegets.wechatcompose.ui.theme.WeComposeTheme
import com.google.accompanist.coil.rememberCoilPainter
import com.google.accompanist.pager.*
import kotlinx.coroutines.InternalCoroutinesApi
import kotlinx.coroutines.flow.collect
import kotlin.math.absoluteValue

/**
 * 大图预览
 */
@OptIn(ExperimentalPagerApi::class, InternalCoroutinesApi::class)
@Composable
fun ImageBrowserScreen(images: List<Image>, selectImage: Image) {

    var currentIndex = 0

    images.forEachIndexed { index, image ->
        if (image.url == selectImage.url) {
            currentIndex = index
            return@forEachIndexed
        }
    }
    /**
     * 界面状态变更
     */
    val pageState = rememberPagerState(initialPage = currentIndex)

    Box {
        HorizontalPager(
            count = images.size,
            state = pageState,
            contentPadding = PaddingValues(horizontal = 0.dp),
            modifier = Modifier.fillMaxSize()
        ) { page ->
            println("ImageBrowserItem current page: $page")
            ImageBrowserItem(images[page], page, this)
        }

        HorizontalPagerIndicator(
            pagerState = pageState,
            activeColor = Color.White,
            inactiveColor = WeComposeTheme.colors.onBackground,
            modifier = Modifier
                .align(Alignment.BottomCenter)
                .padding(60.dp)
        )

        LaunchedEffect(pageState) {
            snapshotFlow { pageState }.collect { pageState ->
                println("ImageBrowserItem LaunchedEffect pageState currentPageOffset: $pageState.currentPageOffset")
            }
        }
    }
}

@OptIn(ExperimentalPagerApi::class)
@Composable
fun ImageBrowserItem(image: Image, page: Int = 0,  pagerScope: PagerScope) {
    /**
     * 缩放比例
     */
    var scale by remember { mutableStateOf(1f) }

    /**
     * 偏移量
     */
    var offset  by remember { mutableStateOf(Offset.Zero) }

    /**
     * 监听手势状态变换
     */
    var state =
        rememberTransformableState(onTransformation = { zoomChange, panChange, rotationChange ->
            scale = (zoomChange * scale).coerceAtLeast(1f)
            scale = if (scale > 5f) {
                5f
            } else {
                scale
            }
            println("ImageBrowserItem detectTapGestures rememberTransformableState scale: $scale")
        })

    Surface(
        modifier = Modifier
            .fillMaxSize(),
        color = Color.Black,
    ) {
        Image(
            painter = rememberCoilPainter(
                request = image.url
            ),
            contentDescription = "",
            contentScale = ContentScale.Fit,
            modifier = Modifier
                .transformable(state = state)
                .graphicsLayer{  //布局缩放、旋转、移动变换
                    scaleX = scale
                    scaleY = scale
                    translationX = offset.x
                    translationY = offset.y

                    val pageOffset = pagerScope.calculateCurrentOffsetForPage(page = page).absoluteValue
                    if (pageOffset == 1.0f) {
                        scale = 1.0f
                    }
                    println("ImageBrowserItem pagerScope calculateCurrentOffsetForPage pageOffset: $pageOffset")
                }
                .pointerInput(Unit) {
                    detectTapGestures(
                        onDoubleTap = {
                            println("ImageBrowserItem detectTapGestures onDoubleTap offset: $it")
                            scale = if (scale <= 1f) {
                                2f
                            } else {
                                1f
                            }
                            offset = Offset.Zero
                        },
                        onTap = {

                        }
                    )
                }
        )
    }
}

@Preview
@Composable
fun ImageBrowserScreenPreview() {
    ImageBrowserScreen(
        images = mutableListOf(),
        selectImage = Image(url = "https://wx4.sinaimg.cn/orj360/001YqBPrly1h0ha9bxk27j63do52iu0y02.jpg")
    )
}

调用

val images = mutableListOf(
    Image("https://wx4.sinaimg.cn/orj360/001YqBPrly1h0ha93zif8j63gg56ob2c02.jpg"),
    Image("https://wx3.sinaimg.cn/orj360/001YqBPrly1h0ha99r0vej652i3dox6r02.jpg"),
    Image("https://wx2.sinaimg.cn/orj360/001YqBPrly1h0ha96rjbij63do52inpf02.jpg"),
    Image("https://wx1.sinaimg.cn/orj360/001YqBPrly1h0ha9hek4cj652i3do1l202.jpg"),
    Image("https://wx4.sinaimg.cn/orj360/001YqBPrly1h0ha9bxk27j63do52iu0y02.jpg"),
    Image("https://wx3.sinaimg.cn/orj360/001YqBPrly1h0ha9e50phj652u3dwhdv02.jpg"),
    Image("https://wx4.sinaimg.cn/orj360/001YqBPrly1h0ha91e0jxj63gg56o7wk02.jpg"),
    Image("https://wx4.sinaimg.cn/orj360/001YqBPrly1h0ha9jfmouj635x4rgb2b02.jpg"),
    Image("https://wx1.sinaimg.cn/orj360/001YqBPrly1h0ha9lzkk3j656o3gge8402.jpg")
)

val selectImage = Image("https://wx4.sinaimg.cn/orj360/001YqBPrly1h0ha93zif8j63gg56ob2c02.jpg")

ImageBrowserScreen(images = images, selectImage = selectImage)

该功能只是仿Jetpack Compose微信的小部分功能，所以暂无源码，所后期会将仿微信全部代码上传至Github。

参考资料

# 使用Jetpack Compose完成自定义手势处理

# 多点触控：平移、缩放、旋转

# Jetpack Navigation Compose Animation

Jetpack Compose版来啦！高仿微信朋友圈大图缩放、切换、预览功能

使用PointerInput Modifier

点击类型基础 API

API介绍

例子如下：

手势检测

transformable 修饰符

例子如下：

需要部分如下依赖

功能实现

HorizontalPager

HorizontalPagerIndicator

双击放大和缩小

双指缩放

切换恢复图片大小

例子如下：

参考资料

使用`PointerInput Modifier`