1
0

unarchiveFile.go 5.5 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252
  1. package archive_helper
  2. import (
  3. "archive/zip"
  4. "bytes"
  5. "compress/flate"
  6. "errors"
  7. "io"
  8. "io/fs"
  9. "os"
  10. "path/filepath"
  11. "strings"
  12. "unicode/utf8"
  13. "github.com/ChineseSubFinder/ChineseSubFinder/pkg"
  14. "github.com/bodgit/sevenzip"
  15. "github.com/mholt/archiver/v3"
  16. "golang.org/x/text/encoding/simplifiedchinese"
  17. "golang.org/x/text/transform"
  18. )
  19. // UnArchiveFileEx 发现打包的字幕内部还有一层压缩包···所以···
  20. func UnArchiveFileEx(fileFullPath, desRootPath string) error {
  21. // 先进行一次解压
  22. err := UnArchiveFile(fileFullPath, desRootPath)
  23. if err != nil {
  24. return err
  25. }
  26. // --------------------------------------------------
  27. doUnzipFun := func() error {
  28. // 判断一次
  29. needUnzipFileFPaths := make([]string, 0)
  30. err = filepath.Walk(desRootPath, func(path string, info fs.FileInfo, err error) error {
  31. if info.IsDir() == true {
  32. return nil
  33. }
  34. nowExt := filepath.Ext(path)
  35. // 然后对于解压的内容再次进行判断
  36. if nowExt != ".zip" && nowExt != ".tar" && nowExt != ".rar" && nowExt != ".7z" {
  37. return nil
  38. } else {
  39. needUnzipFileFPaths = append(needUnzipFileFPaths, path)
  40. }
  41. return nil
  42. })
  43. if err != nil {
  44. return err
  45. }
  46. // 如果有压缩包,那么就继续解压,然后删除压缩包
  47. for _, needUnzipFileFPath := range needUnzipFileFPaths {
  48. err = UnArchiveFile(needUnzipFileFPath, desRootPath)
  49. if err != nil {
  50. return err
  51. }
  52. err = os.Remove(needUnzipFileFPath)
  53. if err != nil {
  54. return err
  55. }
  56. }
  57. return nil
  58. }
  59. // 第二次解压
  60. err = doUnzipFun()
  61. if err != nil {
  62. return err
  63. }
  64. // 第三次解压
  65. err = doUnzipFun()
  66. if err != nil {
  67. return err
  68. }
  69. // 第四次解压
  70. err = doUnzipFun()
  71. if err != nil {
  72. return err
  73. }
  74. return nil
  75. }
  76. // UnArchiveFile 7z 以外的都能搞定中文编码的问题,但是 7z 有梗,需要单独的库去解析,且编码是解决不了的,以后他们搞定了再测试
  77. // 所以效果就是,7z 外的压缩包文件解压ok,字幕可以正常从名称解析出是简体还是繁体,但是7z就没办法了,一定乱码
  78. func UnArchiveFile(fileFullPath, desRootPath string) error {
  79. switch filepath.Ext(strings.ToLower(fileFullPath)) {
  80. case ".zip":
  81. z := archiver.Zip{
  82. CompressionLevel: flate.DefaultCompression,
  83. MkdirAll: true,
  84. SelectiveCompression: true,
  85. ContinueOnError: false,
  86. OverwriteExisting: true,
  87. ImplicitTopLevelFolder: false,
  88. }
  89. err := z.Walk(fileFullPath, func(f archiver.File) error {
  90. if f.IsDir() == true {
  91. return nil
  92. }
  93. zfh, ok := f.Header.(zip.FileHeader)
  94. if ok {
  95. err := processOneFile(f, zfh.NonUTF8, desRootPath)
  96. if err != nil {
  97. return err
  98. }
  99. } else {
  100. // 需要检测文件名是否是乱码
  101. err := processOneFile(f, !utf8.ValidString(f.Name()), desRootPath)
  102. if err != nil {
  103. return err
  104. }
  105. }
  106. return nil
  107. })
  108. if err != nil {
  109. return err
  110. }
  111. case ".tar":
  112. z := archiver.Tar{
  113. MkdirAll: true,
  114. ContinueOnError: false,
  115. OverwriteExisting: true,
  116. ImplicitTopLevelFolder: false,
  117. StripComponents: 1,
  118. }
  119. err := z.Walk(fileFullPath, func(f archiver.File) error {
  120. if f.IsDir() == true {
  121. return nil
  122. }
  123. err := processOneFile(f, false, desRootPath)
  124. if err != nil {
  125. return err
  126. }
  127. return nil
  128. })
  129. if err != nil {
  130. return err
  131. }
  132. case ".rar":
  133. z := archiver.Rar{
  134. MkdirAll: true,
  135. ContinueOnError: false,
  136. OverwriteExisting: true,
  137. ImplicitTopLevelFolder: false,
  138. StripComponents: 1,
  139. }
  140. err := z.Walk(fileFullPath, func(f archiver.File) error {
  141. if f.IsDir() == true {
  142. return nil
  143. }
  144. err := processOneFile(f, false, desRootPath)
  145. if err != nil {
  146. return err
  147. }
  148. return nil
  149. })
  150. if err != nil {
  151. return err
  152. }
  153. case ".7z":
  154. return unArr7z(fileFullPath, desRootPath)
  155. default:
  156. return errors.New("not support un archive file ext")
  157. }
  158. return nil
  159. }
  160. func processOneFile(f archiver.File, notUTF8 bool, desRootPath string) error {
  161. decodeName := f.Name()
  162. if notUTF8 == true {
  163. //ouBytes, err := ChangeFileCoding2UTF8([]byte(f.Name()))
  164. //if err != nil {
  165. // return err
  166. //}
  167. i := bytes.NewReader([]byte(f.Name()))
  168. decoder := transform.NewReader(i, simplifiedchinese.GB18030.NewDecoder())
  169. content, _ := io.ReadAll(decoder)
  170. decodeName = string(content)
  171. //decodeName = string(ouBytes)
  172. }
  173. var chunk []byte
  174. buf := make([]byte, 1024)
  175. for {
  176. n, err := f.Read(buf)
  177. if err != nil && err != io.EOF {
  178. return err
  179. }
  180. //说明读取结束
  181. if n == 0 {
  182. break
  183. }
  184. //读取到最终的缓冲区中
  185. chunk = append(chunk, buf[:n]...)
  186. }
  187. err := pkg.WriteFile(filepath.Join(desRootPath, decodeName), chunk)
  188. if err != nil {
  189. return err
  190. }
  191. return nil
  192. }
  193. func unArr7z(fileFullPath, desRootPath string) error {
  194. r, err := sevenzip.OpenReader(fileFullPath)
  195. if err != nil {
  196. return err
  197. }
  198. defer r.Close()
  199. for _, file := range r.File {
  200. err = un7zOneFile(file, desRootPath)
  201. if err != nil {
  202. return err
  203. }
  204. }
  205. return nil
  206. }
  207. func un7zOneFile(file *sevenzip.File, desRootPath string) error {
  208. rc, err := file.Open()
  209. if err != nil {
  210. return err
  211. }
  212. defer rc.Close()
  213. data, err := io.ReadAll(rc)
  214. if err != nil {
  215. return err
  216. }
  217. decodeName := file.Name
  218. decodeName = filepath.Base(decodeName)
  219. err = pkg.WriteFile(filepath.Join(desRootPath, decodeName), data)
  220. if err != nil {
  221. return err
  222. }
  223. return nil
  224. }
  225. func IsWantedArchiveExtName(fileName string) bool {
  226. switch strings.ToLower(filepath.Ext(fileName)) {
  227. case ".zip", ".tar", ".rar", "7z":
  228. return true
  229. default:
  230. return false
  231. }
  232. }