import Photos
import Vision

class ClassifyPhoto {
    
    struct PhotoSizeInfo {
        var totalSize: Int64 = 0
        var count: Int = 0
    }

    struct ClassifiedPhotos {
        var screenshots: [PHAsset] = []
        var locations: [String: [PHAsset]] = [:] // 按地点分组
        var people: [String: [PHAsset]] = [:]     // 按人物分组
        var similarPhotos: [[PHAsset]] = [] // 存储相似照片组
        var blurryPhotos: [PHAsset] = [] // 添加模糊照片数组
        
        // 添加容量信息
        var screenshotsSize: PhotoSizeInfo = PhotoSizeInfo()
        var locationsSize: PhotoSizeInfo = PhotoSizeInfo()
        var peopleSize: PhotoSizeInfo = PhotoSizeInfo()
        var similarPhotosSize: PhotoSizeInfo = PhotoSizeInfo()
        var blurryPhotosSize: PhotoSizeInfo = PhotoSizeInfo() // 添加模糊照片容量信息
    }
    
    // 添加位置缓存
    private var locationCache: [String: String] = [:]

    func classifyPhotos(
            assets: PHFetchResult<PHAsset>,
            progressHandler: @escaping (String, Float) -> Void,
            completion: @escaping (ClassifiedPhotos) -> Void
        ) {
            
        // 在后台队列处理
        DispatchQueue.global(qos: .userInitiated).async {
            var result = ClassifiedPhotos()
            let group = DispatchGroup()
            
            // 开始处理
            DispatchQueue.main.async {
                progressHandler("正在加载照片...", 0.0)
            }
            
            // 先处理模糊照片检测（占进度的 30%）
            group.enter()
            progressHandler("正在检测模糊照片...", 0.0)
            self.detectBlurryPhotos(from: assets) { blurryPhotos in
                result.blurryPhotos = blurryPhotos
                progressHandler("模糊照片检测完成", 0.3)
                group.leave()
            }
            
            // 1. 检测截图 (占总进度的 20%)
            group.enter()
            self.fetchScreenshots(from: assets) { screenshots in
                result.screenshots = screenshots
                DispatchQueue.main.async {
                    progressHandler("正在检测截图...", 0.3)
                }
                group.leave()
            }
            
            // 2. 检测相似照片 (占总进度的 80%)
            group.enter()
            self.detectSimilarPhotos(
                assets: assets,
                progressHandler: { stage, progress in
                    // 将相似照片检测的进度映射到 20%-100% 的范围
                    let mappedProgress = 0.3 + (progress * 0.6)
                    DispatchQueue.main.async {
                        progressHandler(stage, mappedProgress)
                    }
                }
            ) { similarPhotos in
                result.similarPhotos = similarPhotos
                group.leave()
            }
            
            // 3. 按地点分类 (占总进度的 20%)
//            group.enter()
//            self.classifyByLocation(assets: assets) { locationGroups in
//                result.locations = locationGroups
//                DispatchQueue.main.async {
//                    progressHandler("正在按地点分类...", 0.8)
//                }
//                group.leave()
//            }
            
//             4. 按人物分类 (占总进度的 20%)
            group.enter()
            self.classifyByPeople(assets: assets) { peopleGroups in
                result.people = peopleGroups
                DispatchQueue.main.async {
                    progressHandler("正在按人物分类...", 1.0)
                }
                group.leave()
            }
            
            //                // 添加模糊照片检测
            //                group.enter()
            //                self.detectBlurryPhotos(from: assets) { blurryPhotos in
            //                    result.blurryPhotos = blurryPhotos
            //                    DispatchQueue.main.async {
            //                        progressHandler("正在检测模糊照片...", 1.0)
            //                    }
            //                    group.leave()
            //                }
            
            // 在所有分类完成后计算大小
            group.notify(queue: .main) {
                let sizeGroup = DispatchGroup()
                
                // 计算模糊照片大小
                sizeGroup.enter()
                self.calculateAssetsSize(result.blurryPhotos) { sizeInfo in
                    result.blurryPhotosSize = sizeInfo
                    sizeGroup.leave()
                }
                
                
                // 计算相似照片大小
                sizeGroup.enter()
                let similarAssets = Array(result.similarPhotos.flatMap { $0 })
                self.calculateAssetsSize(similarAssets) { sizeInfo in
                    result.similarPhotosSize = sizeInfo
                    sizeGroup.leave()
                }
                
                // 计算截图大小
                sizeGroup.enter()
                self.calculateAssetsSize(result.screenshots) { sizeInfo in
                    result.screenshotsSize = sizeInfo
                    sizeGroup.leave()
                }
                
//                // 计算地点照片大小
//                sizeGroup.enter()
//                let locationAssets = Array(result.locations.values.flatMap { $0 })
//                self.calculateAssetsSize(locationAssets) { sizeInfo in
//                    result.locationsSize = sizeInfo
//                    sizeGroup.leave()
//                }
                
                // 计算人物照片大小
                sizeGroup.enter()
                let peopleAssets = Array(result.people.values.flatMap { $0 })
                self.calculateAssetsSize(peopleAssets) { sizeInfo in
                    result.peopleSize = sizeInfo
                    sizeGroup.leave()
                }
                
                // 所有大小计算完成后回调
                sizeGroup.notify(queue: .main) {
                    progressHandler("分类完成", 1.0)
                    completion(result)
                }
            }
        }
    }
    
    // 添加内存清理辅助方法
    private func cleanupMemory() {
        // 清理图像缓存
        URLCache.shared.removeAllCachedResponses()
        
        // 强制进行一次垃圾回收
        autoreleasepool {
            let _ = [String](repeating: "temp", count: 1)
        }
        
        #if os(iOS)
        // 发送低内存警告
        UIApplication.shared.perform(Selector(("_performMemoryWarning")))
        #endif
    }

    func detectSimilarPhotos(
            assets: PHFetchResult<PHAsset>,
            progressHandler: @escaping (String, Float) -> Void,
            completion: @escaping ([[PHAsset]]) -> Void
        ) {
            var similarGroups: [[PHAsset]] = []
            let group = DispatchGroup()
                    
            if #available(iOS 13.0, *) {
                var imageFeatures: [(asset: PHAsset, feature: VNFeaturePrintObservation)] = []
                
                // 创建处理队列
                let processingQueue = DispatchQueue(label: "com.app.similarPhotos", qos: .userInitiated)
                let semaphore = DispatchSemaphore(value: 4) // 增加并发数以提高效率
                
                // 1. 提取所有图片的特征
                let totalAssets = assets.count
                var processedAssets = 0
                
                progressHandler("正在加载照片...", 0.0)
                
                for i in 0..<assets.count {
                    let asset = assets[i]
                    group.enter()
                    semaphore.wait()
                    
                    let options = PHImageRequestOptions()
                    options.deliveryMode = .fastFormat // 使用快速模式
                    options.isSynchronous = false
                    options.resizeMode = .fast
                    
                    DispatchQueue.global(qos: .background).async {
                        PHImageManager.default().requestImage(
                            for: asset,
                            targetSize: CGSize(width: 128, height: 128), // 降低分辨率
                            contentMode: .aspectFit,
                            options: options
                        ) { image, _ in
                            defer {
                                semaphore.signal()
                            }
                            
                            guard let image = image,
                                  let cgImage = image.cgImage else {
                                group.leave()
                                return
                            }
                            
                            processingQueue.async {
                                do {
                                    let requestHandler = VNImageRequestHandler(cgImage: cgImage, options: [:])
                                    let request = VNGenerateImageFeaturePrintRequest()
                                    try requestHandler.perform([request])
                                    
                                    if let result = request.results?.first as? VNFeaturePrintObservation {
                                        imageFeatures.append((asset, result))
                                        
                                        // 更新特征提取进度
                                        processedAssets += 1
                                        let progress = Float(processedAssets) / Float(totalAssets)
                                        progressHandler("正在提取特征...", progress * 0.6)
                                    }
                                } catch {
                                    print("特征提取失败: \(error)")
                                }
                                group.leave()
                            }
                        }
                    }
                }
                
                group.notify(queue: processingQueue) {
                    progressHandler("正在比较相似度...", 0.6)
                    
                    // 近似度
                    let similarityThreshold: Float = 0.7
                    var similarGroups: [[PHAsset]] = []
                    
                    // 使用并行处理来加速比较
                    let processingGroup = DispatchGroup()
                    let processingQueue = DispatchQueue(label: "com.yourapp.similarity.processing", attributes: .concurrent)
                    let resultsQueue = DispatchQueue(label: "com.yourapp.similarity.results")
                    let semaphore = DispatchSemaphore(value: 4) // 减少并发数量
                    
                    // 创建一个线程安全的数据结构来存储结果
                    var processedIndices = Atomic<Set<Int>>(Set<Int>())
                    var groupResults = Atomic<[Int: [PHAsset]]>([:])
                    
                    // 分批处理，每批处理一部分数据
                    let batchSize = min(50, imageFeatures.count)
                    // 修复 Float 转换错误
                    let batchCount = Float(imageFeatures.count) / Float(batchSize)
                    let batches = batchCount.isFinite ? Int(ceil(batchCount)) : 1
                    
                    for batchIndex in 0..<batches {
                        let startIndex = batchIndex * batchSize
                        let endIndex = min(startIndex + batchSize, imageFeatures.count)
                        
                        for i in startIndex..<endIndex {
                            // 检查是否已处理
                            if processedIndices.value.contains(i) { continue }
                            
                            semaphore.wait()
                            processingGroup.enter()
                            
                            processingQueue.async {
                                // 再次检查，因为可能在等待期间被其他线程处理
                                if processedIndices.value.contains(i) {
                                    semaphore.signal()
                                    processingGroup.leave()
                                    return
                                }
                                
                                var similarAssets: [PHAsset] = [imageFeatures[i].asset]
                                processedIndices.mutate { $0.insert(i) }
                                
                                for j in (i + 1)..<imageFeatures.count {
                                    // 检查是否已处理
                                    if processedIndices.value.contains(j) { continue }
                                    
                                    do {
                                        var distance: Float = 0
                                        try imageFeatures[i].feature.computeDistance(&distance, to: imageFeatures[j].feature)
                                        
                                        // 检查距离值是否有效
                                        if distance.isNaN || distance.isInfinite {
                                            print("警告: 检测到无效的距离值")
                                            continue
                                        }
                                        
                                        // 确保距离在有效范围内
                                        distance = max(0, min(1, distance))
                                        
                                        let similarity = 1 - distance
                                        if similarity >= similarityThreshold {
                                            similarAssets.append(imageFeatures[j].asset)
                                            processedIndices.mutate { $0.insert(j) }
                                        }
                                    } catch {
                                        print("相似度计算失败: \(error)")
                                    }
                                }
                                
                                // 只保存有多个相似图像的组
                                if similarAssets.count > 1 {
                                    resultsQueue.async {
                                        groupResults.mutate { $0[i] = similarAssets }
                                    }
                                }
                                
                                // 更新进度 - 添加安全检查
                                if imageFeatures.count > 0 {
                                    let processedCount = Float(processedIndices.value.count)
                                    let totalCount = Float(imageFeatures.count)
                                    
                                    // 确保进度值有效
                                    var progress: Float = 0
                                    if processedCount.isFinite && totalCount.isFinite && totalCount > 0 {
                                        progress = processedCount / totalCount
                                        // 限制进度范围
                                        progress = max(0, min(1, progress))
                                    }
                                    
                                    DispatchQueue.main.async {
                                        progressHandler("正在比较相似度...", 0.6 + progress * 0.4)
                                    }
                                }
                                
                                semaphore.signal()
                                processingGroup.leave()
                            }
                        }
                    }
                    
                    processingGroup.wait()
                    
                    // 整理结果
                    similarGroups = Array(groupResults.value.values)
                    
                    // 按照照片数量降序排序
                    similarGroups.sort { $0.count > $1.count }
                    
                    DispatchQueue.main.async {
                        completion(similarGroups)
                    }
                }
            }
        }

    func classifyByLocation(assets: PHFetchResult<PHAsset>,
                                  completion: @escaping ([String: [PHAsset]]) -> Void) {
        var locationGroups: [String: [PHAsset]] = [:]
        let group = DispatchGroup()
        let geocodeQueue = DispatchQueue(label: "com.app.geocoding")
        let semaphore = DispatchSemaphore(value: 10) // 限制并发请求数

        assets.enumerateObjects { asset, _, _ in
            if let location = asset.location {
                group.enter()
                semaphore.wait()
                
                geocodeQueue.async {
                    let geocoder = CLGeocoder()
                    geocoder.reverseGeocodeLocation(location) { placemarks, error in
                        defer {
                            semaphore.signal()
                            group.leave()
                        }
                        
                        if let placemark = placemarks?.first {
                            let locationName = self.formatLocationName(placemark)
                            DispatchQueue.main.async {
                                if locationGroups[locationName] == nil {
                                    locationGroups[locationName] = []
                                }
                                locationGroups[locationName]?.append(asset)
                            }
                        }
                    }
                }
            }
        }

        // 等待所有地理编码完成后回调
        group.notify(queue: .main) {
            completion(locationGroups)
        }
    }

    // 格式化地点名称（只返回城市名）
    func formatLocationName(_ placemark: CLPlacemark) -> String {
        if let city = placemark.locality {
            return city
        } else if let area = placemark.administrativeArea {
            return area
        }
        return "其他"
    }
    
    func classifyByPeople(assets: PHFetchResult<PHAsset>,
                         completion: @escaping ([String: [PHAsset]]) -> Void) {
        // 创建结果字典
        var peopleGroups: [String: [PHAsset]] = [:]
        peopleGroups["包含人脸的照片"] = []
        
        // 使用主队列确保安全完成
        let mainCompletion: ([String: [PHAsset]]) -> Void = { result in
            DispatchQueue.main.async {
                completion(result)
            }
        }
        
        // 限制处理的照片数量，防止内存过载
        let totalCount = min(500, assets.count)
        if totalCount == 0 {
            mainCompletion(peopleGroups)
            return
        }
        
        // 创建专用队列
        let processingQueue = DispatchQueue(label: "com.app.peopleDetection", qos: .userInitiated, attributes: .concurrent)
        let resultQueue = DispatchQueue(label: "com.app.peopleResult", qos: .userInitiated)
        
        // 使用NSLock替代原子操作，更安全
        let resultLock = NSLock()
        
        // 创建进度追踪
        let processedCount = Atomic<Int>(0)
        
        // 分批处理，每批处理一部分数据
        let batchSize = 20
        let batches = Int(ceil(Float(totalCount) / Float(batchSize)))
        
        // 创建一个组来等待所有操作完成
        let group = DispatchGroup()
        
        // 创建一个Vision请求处理器
        let faceDetectionRequest = VNDetectFaceRectanglesRequest()
        
        // 防止过早释放
        var strongSelf: AnyObject? = self
        
        for batchIndex in 0..<batches {
            let startIndex = batchIndex * batchSize
            let endIndex = min(startIndex + batchSize, totalCount)
            
            // 每批处理前进入组
            group.enter()
            
            // 使用延迟减轻主线程压力
            DispatchQueue.global(qos: .userInitiated).asyncAfter(deadline: .now() + Double(batchIndex) * 0.3) { [weak self] in
                guard let self = self else {
                    group.leave()
                    return
                }
                
                // 创建批次内的处理组
                let batchGroup = DispatchGroup()
                // 限制并发数
                let batchSemaphore = DispatchSemaphore(value: 2)
                
                for i in startIndex..<endIndex {
                    batchGroup.enter()
                    batchSemaphore.wait()
                    
                    processingQueue.async {
                        // 使用自动释放池减少内存占用
                        autoreleasepool {
                            let asset = assets[i]
                            
                            // 降低处理图片的分辨率
                            let options = PHImageRequestOptions()
                            options.deliveryMode = .fastFormat
                            options.isSynchronous = false
                            options.resizeMode = .fast
                            options.isNetworkAccessAllowed = false
                            
                            PHImageManager.default().requestImage(
                                for: asset,
                                targetSize: CGSize(width: 120, height: 120),
                                contentMode: .aspectFit,
                                options: options
                            ) { image, info in
                                defer {
                                    batchSemaphore.signal()
                                    batchGroup.leave()
                                }
                                
                                // 检查是否是降级的图像
                                if let degraded = info?[PHImageResultIsDegradedKey] as? Bool, degraded {
                                    return
                                }
                                
                                guard let image = image, let cgImage = image.cgImage else {
                                    return
                                }
                                
                                // 使用简化的人脸检测
                                let handler = VNImageRequestHandler(cgImage: cgImage, options: [:])
                                
                                do {
                                    try handler.perform([faceDetectionRequest])
                                    if let results = faceDetectionRequest.results, !results.isEmpty {
                                        // 检测到人脸，添加到数组
                                        resultLock.lock()
                                        peopleGroups["包含人脸的照片"]?.append(asset)
                                        resultLock.unlock()
                                    }
                                } catch {
                                    print("人脸检测失败: \(error)")
                                }
                                
                                // 更新进度
                                processedCount.mutate { $0 += 1 }
                            }
                        }
                    }
                }
                
                // 等待批次内所有处理完成
                batchGroup.wait()
                
                // 每批处理完后清理内存
                self.cleanupMemory()
                
                // 批次完成
                group.leave()
            }
        }
        
        // 设置超时保护
        let timeoutWorkItem = DispatchWorkItem {
            print("人脸检测超时，返回当前结果")
            mainCompletion(peopleGroups)
            strongSelf = nil
        }
        
        // 30秒后超时
        DispatchQueue.global().asyncAfter(deadline: .now() + 30, execute: timeoutWorkItem)
        
        // 等待所有检测完成后更新结果
        group.notify(queue: .main) {
            // 取消超时
            timeoutWorkItem.cancel()
            
            // 最终清理内存
            self.cleanupMemory()
            
            // 返回结果
            mainCompletion(peopleGroups)
            
            // 释放引用
            strongSelf = nil
        }
    }

    // 优化的人脸检测方法
    private func optimizedFaceDetection(in image: UIImage, request: VNDetectFaceRectanglesRequest, completion: @escaping (Bool) -> Void) {
        guard let cgImage = image.cgImage else {
            completion(false)
            return
        }
        
        // 在后台线程执行检测
        DispatchQueue.global(qos: .userInitiated).async {
            autoreleasepool {
                let handler = VNImageRequestHandler(cgImage: cgImage, options: [:])
                
                do {
                    try handler.perform([request])
                    let hasFace = request.results?.isEmpty == false
                    completion(hasFace)
                } catch {
                    print("人脸检测失败: \(error)")
                    completion(false)
                }
            }
        }
    }
    
//    func classifyByPeople(assets: PHFetchResult<PHAsset>,
//                         completion: @escaping ([String: [PHAsset]]) -> Void) {
//        var peopleGroups: [String: [PHAsset]] = [:]
//        let group = DispatchGroup()
//        
//        // 创建专用队列和信号量控制并发
//        let processingQueue = DispatchQueue(label: "com.app.peopleDetection", qos: .userInitiated, attributes: .concurrent)
//        let resultQueue = DispatchQueue(label: "com.app.peopleResult", qos: .userInitiated)
//        let semaphore = DispatchSemaphore(value: 4) // 限制并发数
//        
//        // 创建进度追踪
//        var processedCount = 0
//        let totalCount = assets.count
//        
//        // 分批处理，每批处理一部分数据
//        let batchSize = 50
//        let batches = Int(ceil(Float(assets.count) / Float(batchSize)))
//        
//        for batchIndex in 0..<batches {
//            let startIndex = batchIndex * batchSize
//            let endIndex = min(startIndex + batchSize, assets.count)
//            
//            // 使用自动释放池减少内存占用
//            autoreleasepool {
//                for i in startIndex..<endIndex {
//                    let asset = assets[i]
//                    group.enter()
//                    semaphore.wait()
//                    
//                    // 降低处理图片的分辨率
//                    let options = PHImageRequestOptions()
//                    options.deliveryMode = .fastFormat
//                    options.isSynchronous = false
//                    options.resizeMode = .fast
//                    
//                    processingQueue.async {
//                        // 使用自动释放池减少内存占用
//                        autoreleasepool {
//                            let _ = PHImageManager.default().requestImage(
//                                for: asset,
//                                targetSize: CGSize(width: 128, height: 128), // 降低分辨率
//                                contentMode: .aspectFit,
//                                options: options
//                            ) { image, _ in
//                                defer {
//                                    semaphore.signal()
//                                }
//                                
//                                guard let image = image else {
//                                    group.leave()
//                                    return
//                                }
//                                
//                                // 使用 Vision 框架检测人脸
//                                guard let ciImage = CIImage(image: image) else {
//                                    group.leave()
//                                    return
//                                }
//                                
//                                let request = VNDetectFaceRectanglesRequest()
//                                let handler = VNImageRequestHandler(ciImage: ciImage, options: [:])
//                                
//                                do {
//                                    try handler.perform([request])
//                                    if let results = request.results, !results.isEmpty {
//                                        // 检测到人脸，添加到数组
//                                        resultQueue.async {
//                                            if peopleGroups["包含人脸的照片"] == nil {
//                                                peopleGroups["包含人脸的照片"] = []
//                                            }
//                                            peopleGroups["包含人脸的照片"]?.append(asset)
//                                        }
//                                    }
//                                } catch {
//                                    print("人脸检测失败: \(error)")
//                                }
//                                
//                                // 更新进度
//                                resultQueue.async {
//                                    processedCount += 1
//                                    let progress = Float(processedCount) / Float(totalCount)
//                                    if processedCount % 100 == 0 || processedCount == totalCount {
//                                        DispatchQueue.main.async {
//                                            print("人脸检测进度: \(Int(progress * 100))%")
//                                        }
//                                    }
//                                }
//                                
//                                group.leave()
//                            }
//                        }
//                    }
//                }
//            }
//            
//            // 每批处理完后清理内存
//            cleanupMemory()
//        }
//        
//        // 等待所有检测完成后更新结果
//        group.notify(queue: .main) {
//            completion(peopleGroups)
//        }
//    }

//    // 添加内存清理方法（如果还没有）
//    private func cleanupMemory() {
//        // 强制清理内存
//        autoreleasepool {
//            // 触发内存警告，促使系统回收内存
//            UIApplication.shared.performSelector(onMainThread: #selector(UIApplication.beginIgnoringInteractionEvents), with: nil, waitUntilDone: true)
//            UIApplication.shared.performSelector(onMainThread: #selector(UIApplication.endIgnoringInteractionEvents), with: nil, waitUntilDone: true)
//        }
//    }

    // 按人物分类
//    func classifyByPeople(assets: PHFetchResult<PHAsset>,
//                                completion: @escaping ([String: [PHAsset]]) -> Void) {
//        var peopleGroups: [String: [PHAsset]] = [:]
//       let group = DispatchGroup()
//
//       DispatchQueue.global(qos: .background).async {
//           // 创建一个数组来存储检测到人脸的照片
//           var facesArray: [PHAsset] = []
//
//           // 遍历所有照片
//           assets.enumerateObjects { asset, _, _ in
//               group.enter()
//
//               // 获取照片的缩略图进行人脸检测
//               let options = PHImageRequestOptions()
//               options.isSynchronous = false
//               options.deliveryMode = .fastFormat
//
//               PHImageManager.default().requestImage(
//                   for: asset,
//                   targetSize: CGSize(width: 128, height: 128), // 使用较小的尺寸提高性能
//                   contentMode: .aspectFit,
//                   options: options
//               ) { image, _ in
//                   guard let image = image else {
//                       group.leave()
//                       return
//                   }
//
//                   // 使用 Vision 框架检测人脸
//                   guard let ciImage = CIImage(image: image) else {
//                       group.leave()
//                       return
//                   }
//
//                   let request = VNDetectFaceRectanglesRequest()
//                   let handler = VNImageRequestHandler(ciImage: ciImage)
//
//                   do {
//                       try handler.perform([request])
//                       if let results = request.results, !results.isEmpty {
//                           // 检测到人脸，添加到数组
//                           DispatchQueue.main.async {
//                               facesArray.append(asset)
//                           }
//                       }
//                   } catch {
//                       print("人脸检测失败: \(error)")
//                   }
//
//                   group.leave()
//               }
//           }
//
//           // 等待所有检测完成后更新结果
//           group.notify(queue: .main) {
//               if !facesArray.isEmpty {
//                   peopleGroups["包含人脸的照片"] = facesArray
//               }
//               completion(peopleGroups)
//           }
//       }
//    }

    // 识别截图
    func fetchScreenshots(from assets: PHFetchResult<PHAsset>,
                                completion: @escaping ([PHAsset]) -> Void) {
        var screenshots: [PHAsset] = []

        // 获取系统的截图智能相册
        let screenshotAlbums = PHAssetCollection.fetchAssetCollections(
            with: .smartAlbum,
            subtype: .smartAlbumScreenshots,
            options: nil
        )

        // 从截图相册中获取所有截图
        screenshotAlbums.enumerateObjects { collection, _, _ in
            let fetchOptions = PHFetchOptions()
            let screenshotAssets = PHAsset.fetchAssets(in: collection, options: fetchOptions)

            screenshotAssets.enumerateObjects { asset, _, _ in
                screenshots.append(asset)
            }
        }

        completion(screenshots)
    }
    
    // 修改辅助方法以接受 PHFetchResult<PHAsset>
//    private func detectScreenshots(assets: PHFetchResult<PHAsset>, completion: @escaping ([PHAsset]) -> Void) {
//        let processingQueue = DispatchQueue(label: "com.yourapp.screenshots.processing", attributes: .concurrent)
//        let resultQueue = DispatchQueue(label: "com.yourapp.screenshots.results")
//        let group = DispatchGroup()
//        let semaphore = DispatchSemaphore(value: 4) // 限制并发数
//        
//        let screenshots = Atomic<[PHAsset]>([])
//        
//        // 分批处理
//        let totalCount = assets.count
//        let batchSize = 50
//        let batches = Int(ceil(Float(totalCount) / Float(batchSize)))
//        
//        for batchIndex in 0..<batches {
//            let startIndex = batchIndex * batchSize
//            let endIndex = min(startIndex + batchSize, totalCount)
//            
//            processingQueue.async {
//                autoreleasepool {
//                    for i in startIndex..<endIndex {
//                        semaphore.wait()
//                        group.enter()
//                        
//                        let asset = assets.object(at: i)
//                        
//                        // 检测是否为截图的逻辑
//                        // ...
//                        
//                        // 模拟检测逻辑
//                        let isScreenshot = asset.pixelWidth == asset.pixelHeight * 16 / 9 ||
//                                          asset.pixelHeight == asset.pixelWidth * 16 / 9
//                        
//                        if isScreenshot {
//                            resultQueue.async {
//                                screenshots.mutate { $0.append(asset) }
//                            }
//                        }
//                        
//                        semaphore.signal()
//                        group.leave()
//                    }
//                }
//            }
//        }
//        
//        group.notify(queue: .main) {
//            completion(screenshots.value)
//        }
//    }
    
    // ... existing code ...

    func detectBlurryPhotos(from assets: PHFetchResult<PHAsset>, completion: @escaping ([PHAsset]) -> Void) {
        var blurryPhotos: [PHAsset] = []
        let group = DispatchGroup()
        let processingQueue = DispatchQueue(label: "com.app.blurryDetection", attributes: .concurrent)
        let resultQueue = DispatchQueue(label: "com.app.blurryResult")
        let semaphore = DispatchSemaphore(value: 8) // 增加并发数
        
        // 创建进度追踪
        var processedCount = 0
        let totalCount = assets.count
        
        // 分批处理，每批处理一部分数据
        let batchSize = 50
        let batches = Int(ceil(Float(assets.count) / Float(batchSize)))
        
        for batchIndex in 0..<batches {
            let startIndex = batchIndex * batchSize
            let endIndex = min(startIndex + batchSize, assets.count)
            
            autoreleasepool {
                for i in startIndex..<endIndex {
                    let asset = assets[i]
                    group.enter()
                    semaphore.wait()
                    
                    let options = PHImageRequestOptions()
                    options.deliveryMode = .fastFormat // 使用快速模式
                    options.isSynchronous = false
                    options.resizeMode = .fast
                    
                    // 进一步降低处理图片的分辨率
                    PHImageManager.default().requestImage(
                        for: asset,
                        targetSize: CGSize(width: 64, height: 64), // 降低分辨率到64x64
                        contentMode: .aspectFit,
                        options: options
                    ) { image, _ in
                        defer {
                            semaphore.signal()
                        }
                        
                        guard let image = image else {
                            group.leave()
                            return
                        }
                        
                        processingQueue.async {
                            // 使用更高效的模糊检测
                            let isBlurry = self.fastBlurCheck(image)
                            
                            if isBlurry {
                                resultQueue.async {
                                    blurryPhotos.append(asset)
                                }
                            }
                            
                            // 更新进度
                            resultQueue.async {
                                processedCount += 1
                                let progress = Float(processedCount) / Float(totalCount)
                                if processedCount % 100 == 0 || processedCount == totalCount {
                                    DispatchQueue.main.async {
                                        print("模糊检测进度: \(Int(progress * 100))%")
                                    }
                                }
                            }
                            
                            group.leave()
                        }
                    }
                }
            }
            
            // 每批处理完后清理内存
            cleanupMemory()
        }
        
        group.notify(queue: .main) {
            completion(blurryPhotos)
        }
    }

    // 更高效的模糊检测方法
    private func fastBlurCheck(_ image: UIImage) -> Bool {
        guard let cgImage = image.cgImage else { return false }
        
        // 使用更小的采样区域
        let width = cgImage.width
        let height = cgImage.height
        let pixelStride = 4 // 改名为pixelStride，避免与函数名冲突
        
        // 提前检查图像尺寸是否合法
        guard width > (2 * pixelStride), height > (2 * pixelStride) else {
            return false
        }
        
        // 使用vImage进行快速处理
        var buffer = [UInt8](repeating: 0, count: width * height)
        
        let colorSpace = CGColorSpaceCreateDeviceGray()
        guard let context = CGContext(
            data: &buffer,
            width: width,
            height: height,
            bitsPerComponent: 8,
            bytesPerRow: width,
            space: colorSpace,
            bitmapInfo: CGImageAlphaInfo.none.rawValue
        ) else {
            return false
        }
        
        context.draw(cgImage, in: CGRect(x: 0, y: 0, width: width, height: height))
        
        // 使用拉普拉斯算子的简化版本
        var score: Double = 0
        var sampledPixels = 0
        
        // 只采样图像的一部分区域
        let sampleRows = 10
        let sampleCols = 10
        
        // 计算步长
        let rowStep = max(1, height / sampleRows)
        let colStep = max(1, width / sampleCols)
        
        // 使用Swift的stride函数
        for y in Swift.stride(from: pixelStride, to: height - pixelStride, by: rowStep) {
            for x in Swift.stride(from: pixelStride, to: width - pixelStride, by: colStep) {
                let current = Int(buffer[y * width + x])
                let left = Int(buffer[y * width + (x - pixelStride)])
                let right = Int(buffer[y * width + (x + pixelStride)])
                let top = Int(buffer[(y - pixelStride) * width + x])
                let bottom = Int(buffer[(y + pixelStride) * width + x])
                
                // 简化的边缘检测
                let dx = abs(left - right)
                let dy = abs(top - bottom)
                score += Double(max(dx, dy))
                sampledPixels += 1
            }
        }
        
        // 避免除以零
        guard sampledPixels > 0 else { return false }
        
        // 归一化分数
        let normalizedScore = score / Double(sampledPixels)
        
        // 调整阈值
        let threshold = 15.0
        return normalizedScore < threshold
    }

    // ... existing code ...
    
//    func detectBlurryPhotos(from assets: PHFetchResult<PHAsset>, completion: @escaping ([PHAsset]) -> Void) {
//        var blurryPhotos: [PHAsset] = []
//        let group = DispatchGroup()
//        let processingQueue = DispatchQueue(label: "com.app.blurryDetection", attributes: .concurrent)
//        let resultQueue = DispatchQueue(label: "com.app.blurryResult")
//        let semaphore = DispatchSemaphore(value: 5) // 增加并发数
//        
//        // 创建进度追踪
//        var processedCount = 0
//        let totalCount = assets.count
//        
//        for i in 0..<assets.count {
//            let asset = assets[i]
//            group.enter()
//            semaphore.wait()
//            
//            let options = PHImageRequestOptions()
//            options.deliveryMode = .fastFormat // 使用快速模式
//            options.isSynchronous = false
//            options.resizeMode = .fast
//            
//            // 降低处理图片的分辨率
//            PHImageManager.default().requestImage(
//                for: asset,
//                targetSize: CGSize(width: 128, height: 128), // 降低分辨率
//                contentMode: .aspectFit,
//                options: options
//            ) { image, _ in
//                defer {
//                    semaphore.signal()
//                }
//                
//                guard let image = image,
//                      let cgImage = image.cgImage else {
//                    group.leave()
//                    return
//                }
//                
//                processingQueue.async {
//                    // 快速模糊检测
//                    let isBlurry = self.quickBlurCheck(cgImage)
//                    
//                    if isBlurry {
//                        resultQueue.async {
//                            blurryPhotos.append(asset)
//                        }
//                    }
//                    
//                    // 更新进度
//                    resultQueue.async {
//                        processedCount += 1
//                        let progress = Float(processedCount) / Float(totalCount)
//                        DispatchQueue.main.async {
//                            print("模糊检测进度: \(Int(progress * 100))%")
//                        }
//                    }
//                    
//                    group.leave()
//                }
//            }
//        }
//        
//        group.notify(queue: .main) {
//            completion(blurryPhotos)
//        }
//    }
//
//    // 快速模糊检测方法
//    private func quickBlurCheck(_ image: CGImage) -> Bool {
//        
//        let width = image.width
//        let height = image.height
//        let stride = 2 // 跳过一些像素以加快速度
//        
//        // 提前检查图像尺寸是否合法
//        guard width > (2 * stride), height > (2 * stride) else {
//            return false // 小尺寸图像直接判定为模糊或清晰
//        }
//        
//        var buffer = [UInt8](repeating: 0, count: width * height)
//        
//        let colorSpace = CGColorSpaceCreateDeviceGray()
//        guard let context = CGContext(
//            data: &buffer,
//            width: width,
//            height: height,
//            bitsPerComponent: 8,
//            bytesPerRow: width,
//            space: colorSpace,
//            bitmapInfo: CGImageAlphaInfo.none.rawValue
//        ) else {
//            return false
//        }
//        
//        context.draw(image, in: CGRect(x: 0, y: 0, width: width, height: height))
//        
//        // 使用简化的拉普拉斯算子
//        var score: Double = 0
//        
//        for y in stride..<(height-stride) where y % stride == 0 {
//            for x in stride..<(width-stride) where x % stride == 0 {
//                let current = Int(buffer[y * width + x])
//                let left = Int(buffer[y * width + (x - stride)])
//                let right = Int(buffer[y * width + (x + stride)])
//                let top = Int(buffer[(y - stride) * width + x])
//                let bottom = Int(buffer[(y + stride) * width + x])
//                
//                // 简化的边缘检测
//                let dx = abs(left - right)
//                let dy = abs(top - bottom)
//                score += Double(max(dx, dy))
//            }
//        }
//        
//        // 归一化分数
//        let normalizedScore = score / Double((width * height) / (stride * stride))
//        
//        // 调整阈值（可能需要根据实际效果调整）
//        let threshold = 20.0
//        return normalizedScore < threshold
//    }
}

extension ClassifyPhoto {
    
    // 获取资源大小的辅助方法
    func getAssetSize(_ asset: PHAsset, completion: @escaping (Int64) -> Void) {
        DispatchQueue.global(qos: .background).async {
            let resources = PHAssetResource.assetResources(for: asset)
            if let resource = resources.first {
                var size: Int64 = 0
                if let unsignedInt64 = resource.value(forKey: "fileSize") as? CLong {
                    size = Int64(unsignedInt64)
                }
                DispatchQueue.main.async {
                    completion(size)
                }
            } else {
                DispatchQueue.main.async {
                    completion(0)
                }
            }
        }
    }

    // 计算资产组的总大小
    func calculateAssetsSize(_ assets: [PHAsset], completion: @escaping (PhotoSizeInfo) -> Void) {
        print("正在计算图片组容量大小")
        let group = DispatchGroup()
        var totalSize: Int64 = 0
        
        for asset in assets {
            group.enter()
            getAssetSize(asset) { size in
                totalSize += size
                group.leave()
            }
        }
        
        group.notify(queue: .main) {
            completion(PhotoSizeInfo(totalSize: totalSize, count: assets.count))
        }
    }
}

extension ClassifyPhoto {
    
    // 添加一个处理 P3 色彩空间图像的辅助方法
    private func processImageWithSafeColorSpace(_ image: UIImage) -> UIImage? {
        autoreleasepool {
            guard let cgImage = image.cgImage else { return image }
            
            // 检查色彩空间
            if let colorSpace = cgImage.colorSpace,
               (colorSpace.name as String?) == CGColorSpace.displayP3 as String {
                
                // 转换为 sRGB 色彩空间
                let sRGBColorSpace = CGColorSpaceCreateDeviceRGB()
                if let context = CGContext(
                    data: nil,
                    width: cgImage.width,
                    height: cgImage.height,
                    bitsPerComponent: 8,
                    bytesPerRow: 0,
                    space: sRGBColorSpace,
                    bitmapInfo: CGImageAlphaInfo.premultipliedLast.rawValue
                ) {
                    context.draw(cgImage, in: CGRect(x: 0, y: 0, width: cgImage.width, height: cgImage.height))
                    if let convertedImage = context.makeImage() {
                        return UIImage(cgImage: convertedImage, scale: image.scale, orientation: image.imageOrientation)
                    }
                }
            }
            
            return image
        }
    }

    // 修改图像请求方法，添加色彩空间处理
    private func requestImageWithSafeProcessing(
        for asset: PHAsset,
        targetSize: CGSize,
        contentMode: PHImageContentMode,
        options: PHImageRequestOptions?,
        completion: @escaping (UIImage?) -> Void
    ) {
        PHImageManager.default().requestImage(
            for: asset,
            targetSize: targetSize,
            contentMode: contentMode,
            options: options
        ) { image, info in
            guard let image = image else {
                completion(nil)
                return
            }
            
            // 处理可能的 P3 色彩空间图像
            DispatchQueue.global(qos: .userInitiated).async {
                let processedImage = self.processImageWithSafeColorSpace(image)
                DispatchQueue.main.async {
                    completion(processedImage)
                }
            }
        }
    }
}


class Atomic<T> {
    private var value_: T
    private let lock = NSLock()
    
    init(_ value: T) {
        self.value_ = value
    }
    
    var value: T {
        lock.lock()
        defer { lock.unlock() }
        return value_
    }
    
    func mutate(_ mutation: (inout T) -> Void) {
        lock.lock()
        defer { lock.unlock() }
        mutation(&value_)
    }
}