submitted 1 year ago by Lionir@beehaw.org to c/support@beehaw.org

34 comments fedilink hide all child comments

After some discussions in !chat, we came up with the conclusion we should adopt rules surrounding bots.

We'll ban bots which we are aware of that currently don't follow these rules and contact their creators. Please report bots that don't follow these.

you are viewing a single comment's thread
view the rest of the comments

[-] admin@beehaw.org 3 points 1 year ago

I'm just trying to understand, generally speaking, how the bot works. It appears to me that the bot is looking for posted articles that exceed a certain word count threshold. If it finds these, then it creates a summary and posts this as a comment. Am I understanding this correctly?

[-] rikudou@lemmings.world 4 points 1 year ago

It has support for specific news sites, I don't want to rely on some automatic text extraction because those are prone to breaking. Here are the content extractors themselves, each for one site. If a post that contains a link to any of the supported sites is found across all of Lemmy (that the bot can see), it extracts the text and then summarizes it using this. It takes 6 sentences directly from the article that look most important to the machine learning model it uses. Then it posts it as a comment.