Open Source @lemmy.ml asudox @lemmy.world 1 yr. ago

GitHub - Asudox/LemmyCTTS: A program to convert Lemmy comments to a video

github.com GitHub - Asudox/LemmyCTTS: A program to convert Lemmy comments to a video

A program to convert Lemmy comments to a video. Contribute to Asudox/LemmyCTTS development by creating an account on GitHub.

I basically made this program to convert Lemmy comments to videos just like those Reddit videos on YouTube. I know that there are a few bugs here and there, I will fix them later. For now, the most obvious one is the transition effect video not showing up during transitions, only the audio does. There also might be issues when you run it, I didn't test if it works when the github version gets cloned so tell me if something is wrong, thanks. I also just got it to work as I wanted even though there are improvements I could make if it weren't 10pm for me. I will do the improvements when I have the time, for now, take this one as a alpha version.

Anyways, what do you guys think?

Edit: Sorry about the previous error, made the program remove the .gitkeep files beforehand as they break the program

Edit 2: Added a config.json for easy configuration. I also made a quick project blog page explaining how it works: https://asudox.dev/projects/lemmyctts

15 comments

So, it seems I'm really out of the loop. Is this some AI thing generating a video from my comment? Or does this scroll through all comments while "recording" the screen?
- It firstly takes the posts sorted by the SORT_TYPE (e.g. Hot) from a community, for example, asklemmy@lemmy.world
  
  It then gets the comments from that random post sorted by Top and selects as much as the COMMENT_LIMIT is. Then the function screenshotLemmy() from screenshotLemmy.py gets called. This function uses the selenium library to simulate the browser of choice (e.g. Firefox) and screenshots the random comments and the random post selected earlier. After that, createTTS() inside createTTS.py gets called which converts the post and the comments to speech using the gTTS library. Then, createLemmyVideo() inside createLemmyVideo.py gets called. This function firstly creates parts of the final video that will be put together later using moviepy. After all that, cleanup() inside finalCleanup.py gets called, which deletes everything inside the tts, video/parts and screenshots folder. The put together video (final video) will be inside the video folder.
- God no. Please, we dont need more anon-style YouTube videos that just read low quality text content that was never intended to he in a video format
  
  Wait for when someone makes a 4chan to video converter :)
- Here's a more detailed version: https://asudox.dev/projects/lemmyctts
Can you link to one of these videos? Would be good to have an example in the github readme somewhere. I have no idea what this is about. What reddit videos?
- For example this one: https://youtu.be/Rz6RDJNwe7Y?si=gVG2SLMh6vGIjZVX
  
  Mine isn't as advanced as the current ones though.
  
  Here is an alternative Piped link(s):
  
  https://piped.video/Rz6RDJNwe7Y?si=gVG2SLMh6vGIjZVX
  
  Piped is a privacy-respecting open-source alternative frontend to YouTube.
  
  I'm open-source; check me out at GitHub.
  
  Wow, that has 43k views.
I don't like these videos, but it could boost Lemmy's popularity if used right.

You've viewed 15 comments.