Fix non-latin folder and file names

nashid · Oct 20, 2021 · 40622d8 · 40622d8
1 parent 3f33393
commit 40622d8
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 4 deletions.
diff --git a/gdown/download.py b/gdown/download.py
@@ -183,7 +183,10 @@ def download(
             return
 
     if gdrive_file_id and is_gdrive_download_link:
-        m = re.search('filename="(.*)"', res.headers["Content-Disposition"])
+        content_disposition = six.moves.urllib_parse.unquote(
+            res.headers["Content-Disposition"]
+        )
+        m = re.search(r"filename\*=UTF-8''(.*)", content_disposition)
         filename_from_url = m.groups()[0]
     else:
         filename_from_url = osp.basename(url)

diff --git a/gdown/download_folder.py b/gdown/download_folder.py
@@ -116,12 +116,15 @@ def parse_google_drive_file(folder, content, use_cookies=True):
     folder_contents = [] if folder_arr[0] is None else folder_arr[0]
 
     gdrive_file = GoogleDriveFile(
-        id=folder[39:],
-        name=folder_soup.title.contents[0][:-15],
+        id=folder.split("/")[-1],
+        name=" - ".join(folder_soup.title.contents[0].split(" - ")[:-1]),
         type=folder_type,
     )
 
-    id_name_type_iter = ((e[0], e[2], e[3]) for e in folder_contents)
+    id_name_type_iter = [
+        (e[0], e[2].encode("raw_unicode_escape").decode("utf-8"), e[3])
+        for e in folder_contents
+    ]
 
     return gdrive_file, id_name_type_iter