Support Multilang Subtitles for Videos using AI Models

[swve]

Verified issue

• Someone from the team allowed me to create an issue here

Issue Content

Hi Everyone,
Thanks for your interest in LearnHouse, this is a good first issue for anyone that want to contribute.

Issue Description:

Ability to get Transcriptions (CC) from hosted videos on the Backend.

Difficulty Level: Easy/Intermediate/Advanced

Advanced

Expected Benefits:

• Being able to get multilang subtitles for Videos hosted through the Backend

Implementation

• Any AI model would expect .mp3 files to be uploaded with a prompt to detect audio, for that reason the mp3 sound should be extracted from the video before upload, use FFmpeg for this.
• Endpoints would be part of /ai
• Multiple subtitles would be produced from an mp3, for at least 4 languages.
• Video upload should be the priority in the operation, you can use BackgroundTasks for this.
• Use langchain and OpenAI's Speech To Text for now
• This Issue will serve as a discussion to work together on this feature

Future Implementations

• Get the subtitles to show on the upcoming video player