Some Unicode / Persian filename causes FileNotFoundError in tempfile.mkstemp during download

[esmaeelE]

Describe the bug

Downloading videos with Persian (Farsi) titles fails with a FileNotFoundError in Python's tempfile.mkstemp() when using the wget module.

Consider download this file by apyrat

https://www.aparat.com/v/u75537l?playlist=289610

To Reproduce

1. Install latest version:

   pip install -U apyrat

2. Run:

   apyrat https://www.aparat.com/v/u75537l?playlist=289610

3. Choose quality (e.g. 720p)

Expected behavior

The video should download successfully with the correct Persian filename.

Actual behavior

لینوکس از کیلومتر صفر (ویدیوی دوم: انتخاب توزیع گنو/لینوکس)
https://as9.cdn.asset.aparat.com/...
Traceback (most recent call last):
  ...
  File "/.../wget.py", line 506, in download
    (fd, tmpfile) = tempfile.mkstemp(".tmp", prefix=prefix, dir=".")
  ...
FileNotFoundError: [Errno 2] No such file or directory: '/home/user/apyrat/entekhab/لینوکس از کیلومتر صفر (ویدیوی دوم: انتخاب توزیع گنو/لینوکس).mp4xcjpfmtx.tmp'

Environment

• OS: Debian 13 (also likely affects other Linux distributions)
• Python: 3.11
• apyrat version: latest (as of June 2026)
• Shell: bash
• Locale: UTF-8 (Persian)

Additional context

• The bug occurs because wget.download() creates a temporary file using the full output filename (which contains Persian characters) as a prefix in tempfile.mkstemp().
• The combination of Persian directory + Persian filename breaks the temporary file creation on some systems.
• yt-dlp downloads the same video without any issues.

Suggested fix by AI

In src/apyrat.py (around line 158):

1. Sanitize the filename (remove/replace invalid characters for the filesystem).
2. Use a safe ASCII temporary filename.
3. Consider replacing the wget module with requests + proper streaming, or use tempfile.mkstemp(dir="/tmp") explicitly.