RayVentura
Split the content engines into multiple classes that will take care of specific actions, such as Script generation, Image / video search, ect..
The current content engine is very general, and takes care of the following steps of content creation: 1- script generation 2- audio voiceover creation 3- audio speeding 4- caption timing from voiceover (with whisper) 5- If wants images: generate image search terms and time them 6- Scrape the images 7- Choose background music 8- Choose background video 9- Prepare custom assets if any 10- Edit and render video or short 11- Add metadata
These steps are pretty good for a few types of content, but it restricts all other types of contents that are much more volatile in their video editing ability. (Imagine playing clips of other audios in the middle of the video, or playing a highlight, ect..) As such, we must separate the tasks of the Content engine into many smaller modules that can handle these separate tasks, and have higher abstractions binding them.
