| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153 |
- package handler
- import (
- "errors"
- "fmt"
- "io/fs"
- "os"
- "path/filepath"
- "strings"
- "sync"
- )
- type PathInfo struct {
- Dirs []string // 目录路径(包含所有层级)
- Files []string // 文件路径
- }
- var pathPool = sync.Pool{
- New: func() interface{} {
- return &PathInfo{
- Dirs: make([]string, 0, 1024), // 预分配容量
- Files: make([]string, 0, 4096),
- }
- },
- }
- func GoHeavy(info *PathInfo) (PathInfo, error) {
- m := make(map[string]bool)
- for i := range info.Files {
- path := info.Files[i]
- lastIndex := strings.LastIndex(path, "/")
- if lastIndex == -1 {
- lastIndex = strings.LastIndex(path, "\\")
- }
- if lastIndex == -1 {
- continue
- }
- //if _, ok := m[path[:lastIndex]]; !ok {
- // fmt.Println(path[:lastIndex])
- //}
- m[path[:lastIndex]] = true
- }
- var dirs = make([]string, 0, len(m))
- fmt.Println(len(info.Dirs))
- for i := range info.Dirs {
- path := info.Dirs[i]
- if !m[path] {
- fmt.Println(info.Dirs[i])
- dirs = append(dirs, info.Dirs[i])
- }
- }
- info.Dirs = dirs
- return *info, nil
- }
- func GetPathInfo(root string) (*PathInfo, error) {
- //获取文件名
- index := strings.LastIndex(root, "/")
- if index == -1 {
- index = strings.LastIndex(root, "\\")
- }
- //RootFileName := ""
- //if index == -1 {
- // RootFileName = root
- //} else {
- // RootFileName = root[:index+1]
- //}
- // 复用对象减少内存分配[9](@ref)
- info := pathPool.Get().(*PathInfo)
- // 路径有效性校验[5,6](@ref)
- root = filepath.Clean(root)
- stat, err := os.Stat(root)
- if err != nil {
- if errors.Is(err, fs.ErrNotExist) {
- return nil, &os.PathError{Op: "stat", Path: root, Err: err} // 增强错误信息[8](@ref)
- }
- return nil, err
- }
- // 文件直接返回[4](@ref)
- if !stat.IsDir() {
- info.Files = append(info.Files, root)
- return info, nil
- }
- // 使用WalkDir优化遍历性能[4,9](@ref)
- err = filepath.WalkDir(root, func(path string, d fs.DirEntry, err error) error {
- if err != nil {
- // 处理遍历错误但继续执行[7](@ref)
- if errors.Is(err, fs.ErrPermission) {
- return nil // 跳过权限错误
- }
- return err
- }
- // 排除根目录自身
- if path == root {
- return nil
- }
- // 动态扩容检测[9](@ref)
- if len(info.Dirs)+len(info.Files) > 1e8 {
- return errors.New("超出内存安全阈值(1,000,000条路径)")
- }
- //fmt.Println(path[len(RootFileName):])
- if d.IsDir() {
- info.Dirs = append(info.Dirs, path)
- } else {
- info.Files = append(info.Files, path)
- }
- return nil
- })
- if err != nil {
- return nil, &os.PathError{Op: "walk", Path: root, Err: err} // 错误包装[6](@ref)
- }
- return info, nil
- }
- // 流式处理版本(适用于超大数据集)
- func StreamPathInfo(root string) (<-chan string, <-chan error) {
- paths := make(chan string, 1000)
- errs := make(chan error, 1)
- go func() {
- defer close(paths)
- defer close(errs)
- err := filepath.WalkDir(root, func(path string, d fs.DirEntry, err error) error {
- if err != nil {
- return err
- }
- if path != root {
- select {
- case paths <- path:
- default:
- return errors.New("消费者处理过慢")
- }
- }
- return nil
- })
- if err != nil {
- errs <- &os.PathError{Op: "stream", Path: root, Err: err}
- }
- }()
- return paths, errs
- }
|