Installing can be done with:
Run the CUDA installer
pip install -r requirements
Then running the Magick installer and specifying a path for it
Then run the espeak installer and add C:\Program Files\eSpeak NG to PATH
Downloading ffmpeg, place in C:, and add ffmpeg bin to PATH
Finish by rebooting the system

Probably isn't needed, but could try running what's below
Before running commands run the line below:
Start-Process -FilePath "C:\Program Files\eSpeak NG\espeak-ng.exe"

Future Upgrades:
-Train own TTS model
-Run TTS on GPU with CUDA
-Get TTS to pause a little for ()
-Allow mixing of original video's audio with TTS generated audio
-Add slideshow format videos
-Add more acronyms
-Remove pause in TTS when text pages split mid-sentence. Could generate a setcond audio clip that reads both pages, and use the original audio to tell when to turn the page
-Stylize story title page to indicate title page vs body page
-Choose videos to use in lower half of screen that are snappier and pull in attention. Speed up Tom and Jerry clips?

Text Splitter Upgrades:
-Add () and ""
-Account for sets of multiple periods, especially ...
-Account for periods between () and ""
-Try testing multiple periods in a row
-Change () to have higher importance than .
-Test the below piece next!
Read title and body in 1 go each. For the body, take the total time of each text_page and get the ratio between it and the audio duration. Multiply this ratio by each individual text page's duration to get the duration to use for each text page.
-Remove excess period created when splitting on . and adding back in a . (Might be complete, have to test) (extra ; should also be fixed, have to test) also test acronym replacement and Reddit post title

Bug Fixes:
-Fix vertical lyric vid's text slides

Next Steps:
-Implement manual lyric splitter to determine page durations
-Auto Upload videos to YouTube and TikTok
-Generate Icon and Banner for YouTube and TikTok accounts
-Create YouTube and TikTok accounts
-Wikipedia Scraper or other sites
-Add image scraper from Google Images or elsewhere
-Integrate AI content generation instead of scraping Using Google Gemini
-Create webpage to serve as UI for process



Video Formats:

Story Format:
Includes a banner at bottom of video with stock video footage filling the lower half of screen.
Top half of screen includes text that's being read by TTS
Vertical.

Lyric Format:
Includes lyrics on screen in pages with music in background. Pages flip according to music
First page has song title followed by Artist
Can be created in Vertical or Horizontal Format. Pages will differ between the two since Horizontal allows more text on screen at once

Slideshow Format:
Features images and video displayed using the entire screen. Images move slightly to catch the eye.
Text is placed on top of slideshow and read by TTS.
Vertical.