Merge pull request #29 from nitrix/fix/issue-28

Fixes issues #28, invalid filenames on some OSes

.gitignore

@@ -1,2 +1,3 @@
+/.idea
+/.vscode
 youtubedr
-.vscode/launch.json

youtube.go

@@ -7,6 +7,7 @@ import (
  "io"
  "io/ioutil"
  "log"
+ "mime"
  "net/http"
  "net/url"
  "os"
@@ -109,23 +110,85 @@ func (y *Youtube) StartDownloadWithQuality(destFile string, quality string) erro
 func (y *Youtube) StartDownloadFile() error {
  //download highest resolution on [0]
  err := errors.New("Empty stream list")
- for _, v := range y.StreamList {
- url := v["url"]
- y.log(fmt.Sprintln("Download url=", url))
+ for _, stream := range y.StreamList {
+ streamUrl := stream["url"]
+ y.log(fmt.Sprintln("Download url=", streamUrl))
+
+ // Find out what the file name should be.
+ fileName := sanitizeFilename(stream["title"])
+ fileName += pickIdealFileExtension(stream["type"])
 
  usr, _ := user.Current()
- fileName := fmt.Sprintf("%s.mov", v["title"])
  destFile := filepath.Join(filepath.Join(usr.HomeDir, "Movies", "youtubedr"), fileName)
  y.log(fmt.Sprintln("Download to file=", destFile))
 
- err = y.videoDLWorker(destFile, url)
+ err = y.videoDLWorker(destFile, streamUrl)
  if err == nil {
- break
+ return nil
  }
  }
  return err
 }
 
+func pickIdealFileExtension(mediaType string) string {
+ defaultExtension := ".mov"
+
+ mediaType, _, err := mime.ParseMediaType(mediaType)
+ if err != nil {
+ return defaultExtension
+ }
+
+ // Rely on hardcoded canonical mime types, as the ones provided by Go aren't exhaustive [1].
+ // This seems to be a recurring problem for youtube downloaders, see [2].
+ // The implementation is based on mozilla's list [3], IANA [4] and Youtube's support [5].
+ // [1] https://github.com/golang/go/blob/ed7888aea6021e25b0ea58bcad3f26da2b139432/src/mime/type.go#L60
+ // [2] https://github.com/ZiTAL/youtube-dl/blob/master/mime.types
+ // [3] https://developer.mozilla.org/en-US/docs/Web/HTTP/Basics_of_HTTP/MIME_types/Common_types
+ // [4] https://www.iana.org/assignments/media-types/media-types.xhtml#video
+ // [5] https://support.google.com/youtube/troubleshooter/2888402?hl=en
+ canonicals := map[string]string {
+ "video/quicktime": ".mov",
+ "video/x-msvideo": ".avi",
+ "video/x-matroska": ".mkv",
+ "video/mpeg": ".mpeg",
+ "video/webm": ".webm",
+ "video/3gpp2": ".3g2",
+ "video/x-flv": ".flv",
+ "video/3gpp": ".3gp",
+ "video/mp4": ".mp4",
+ "video/ogg": ".ogv",
+ "video/mp2t": ".ts",
+ }
+
+ if extension, ok := canonicals[mediaType]; ok {
+ return extension
+ }
+
+ // Our last resort is to ask the operating system, but these give multiple results and are rarely canonical.
+ extensions, err := mime.ExtensionsByType(mediaType)
+ if err != nil || extensions == nil {
+ return defaultExtension
+ }
+
+ return extensions[0]
+}
+
+func sanitizeFilename(fileName string) string {
+ // Characters not allowed on mac
+ // :/
+ // Characters not allowed on linux
+ // /
+ // Characters not allowed on windows
+ // <>:"/\|?*
+
+ // Ref https://docs.microsoft.com/en-us/windows/win32/fileio/naming-a-file#naming-conventions
+
+ fileName = regexp.MustCompile(`[:/<>\:"\\|?*]`).ReplaceAllString(fileName, "")
+ fileName = regexp.MustCompile(`\s+`).ReplaceAllString(fileName, " ")
+
+ return fileName
+}
+
 func (y *Youtube) parseVideoInfo() error {
  answer, err := url.ParseQuery(y.videoInfo)
  if err != nil {

youtube_test.go

@@ -63,3 +63,23 @@ func TestParseVideo(t *testing.T) {
  return
  }
 }
+
+func TestSanitizeFilename(t *testing.T) {
+ fileName := "a<b>c:d\\e\"f/g\\h|i?j*k"
+ sanitized := sanitizeFilename(fileName)
+ if sanitized != "abcdefghijk" {
+ t.Error("Invalid characters must get stripped")
+ }
+
+ fileName = "aB Cd"
+ sanitized = sanitizeFilename(fileName)
+ if sanitized != "aB Cd" {
+ t.Error("Casing and whitespaces must be preserved")
+ }
+
+ fileName = "~!@#$%^&()[].,"
+ sanitized = sanitizeFilename(fileName)
+ if sanitized != "~!@#$%^&()[].," {
+ t.Error("The common harmless symbols should remain valid")
+ }
+}