def dynamic_chunk_size(bandwidth):  
    if bandwidth < 10: return 5*1024*1024  
    elif 10 <= bandwidth < 100: return 20*1024*1024  
    else: return 200*1024*1024  

// 前端切割文件
const sliceFile = async (file) => {
  const chunkSize = file.size > 1024 * 1024 * 1024 ? 200 * 1024 * 1024 : 5 * 1024 * 1024;
  let offset = 0;

  while (offset < file.size) {
    const chunk = file.slice(offset, offset + chunkSize);
    chunks.push({
      index: Math.floor(offset / chunkSize),
      hash: await calculateMD5(chunk)
    });
    offset += chunkSize;
  }
};

CREATE TABLE upload_task (
    task_id VARCHAR(64) PRIMARY KEY COMMENT '任务唯一标识(UUIDv4)',
    file_hash CHAR(40) NOT NULL COMMENT '文件SHA1指纹(全大写HEX字符串)',
    total_chunks INT NOT NULL COMMENT '总分片数量(1-65535)',
    completed_chunks LONGTEXT NOT NULL COMMENT '已上传分片序号数组(JSON数组格式，如[1,2,3])',
    status ENUM('PENDING','UPLOADING','MERGING','COMPLETED') NOT NULL DEFAULT 'PENDING' COMMENT '任务状态:待处理/上传中/合并中/已完成',
    created_at DATETIME(3) NOT NULL DEFAULT CURRENT_TIMESTAMP(3) COMMENT '任务创建时间(带毫秒)',
    updated_at DATETIME(3) NOT NULL DEFAULT CURRENT_TIMESTAMP(3) ON UPDATE CURRENT_TIMESTAMP(3) COMMENT '最后更新时间',
    expire_time DATETIME COMMENT '任务过期时间(默认创建后7天自动清理)',
    retry_count TINYINT UNSIGNED NOT NULL DEFAULT 0 COMMENT '合并失败重试次数(0-255)',
    storage_path VARCHAR(512) COMMENT '最终存储路径(合并完成后更新)',
    INDEX idx_file_hash (file_hash) USING BTREE COMMENT '文件哈希索引',
    INDEX idx_status (status) USING BTREE COMMENT '状态索引',
    INDEX idx_expire (expire_time) USING BTREE COMMENT '过期时间索引'
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_bin
COMMENT='文件分片任务元数据表'

-- 查询待合并任务  有分片上传完成时执行
SELECT * FROM upload_task 
WHERE status = 'UPLOADING' and task_id = #{taskId}
AND JSON_LENGTH(completed_chunks) = total_chunks
AND created_at >= DATE_SUB(NOW(), INTERVAL 2 HOUR);

@PostMapping("/upload-chunk")
public ResponseEntity<?> uploadChunk(
    @RequestParam String fileHash,
    @RequestParam int chunkNumber,
    @RequestParam MultipartFile chunk) {

    // 验证分片是否已存在（幂等设计）
    if (storageService.isChunkExists(fileHash, chunkNumber)) {
        return ResponseEntity.ok().build();
    }

    // 存储分片到MinIO临时目录
    String path = String.format("/temp/%s/%d", fileHash, chunkNumber);
    storageService.upload(path, chunk.getInputStream());

    // 更新任务进度
    taskService.updateChunkStatus(fileHash, chunkNumber);

    return ResponseEntity.ok().build();
}

public void mergeFile(String fileHash) throws IOException {
    List<String> chunkPaths = getSortedChunkPaths(fileHash);

   ForkJoinPool customPool = new ForkJoinPool(Runtime.getRuntime().availableProcessors());
    customPool.submit(() -> 
    IntStream.range(0, chunkPaths.size()).parallel().forEach(i -> {
        String chunkPath = chunkPaths.get(i);
        long position = offsets.get(i);
        try (FileChannel source = FileChannel.open(Paths.get(chunkPath), StandardOpenOption.READ);
             FileChannel dest = FileChannel.open(finalPath, StandardOpenOption.WRITE)) {
            source.transferTo(0, source.size(), dest.position(position));
        } catch (IOException e) {
            throw new RuntimeException(e);
        }
    })
).get();

    // 生成最终文件的MD5校验码
    String mergedHash = DigestUtils.md5Hex(Files.newInputStream(finalPath));
    if (!mergedHash.equals(fileHash)) {
        throw new RuntimeException("File integrity check failed!");
    }
}

// 未优化的浏览器端哈希计算（可能阻塞主线程）
async function calculateHash(chunk) {
  return crypto.subtle.digest('SHA-256', chunk); 
}