Migration is quite slow, and we need to improve that.

• Investigate extract metadata task and propose solution. It's copying the file and we can save some time with better solution than running the task
  • Dont run the task, use duration and quality if they're exist, task will run after migration.
• Investigate extract frames task and how frames are used on player.
  • If frames are exists, we can move the files, but we'll not run the task if they're missing
• Design how we can use more frames (we should use slides as frames) and how they can be uploaded with REST API
  • New celery task implemented to extract more frames only for chapters