Hey @rikudou@lemmings.world, the bot seems to be having some difficulties with correctly parsing articles from the ABC. It’s been doing it on a fair few posts (see below examples as well). As far as I can tell, it’s only occuring on articles from the ABC and I’m not entirely sure what’s causing it.
It looks like ABC must have changed the internal layout of their pages for whatever reason. It seems like the bot is just selecting the first block quote as the entire article.
On The Register for example it selects the div with the id #body. For ABC it seems that it looks for the class Article_Body which I can’t find on that article. I might have a closer look later if I’ve got some time and try to get a PR in if it doesn’t get fixed.
That’s the case, they removed one level of nesting from the html. Anyway, it doesn’t look for Article_Body class, but any class that starts with Article_Body. They’re using randomized class names with the prefix being constant, that’s why I have to do it that way. I’ve updated it to this horrible looking selector: div[class*="Article_body"] > div > p, div[class*="Article_body"] > div > ul:not([class*="ShareUtility"]) > li.
Thanks! I thought it might’ve been a wildcard thing but wasn’t sure. They really don’t want their articles summarised do they (or they’re probably trying to discourage AI scrapers)
Hey @rikudou@lemmings.world, the bot seems to be having some difficulties with correctly parsing articles from the ABC. It’s been doing it on a fair few posts (see below examples as well). As far as I can tell, it’s only occuring on articles from the ABC and I’m not entirely sure what’s causing it.
Other examples:
https://lemmings.world/comment/8105800
https://lemmings.world/comment/8196693
It looks like ABC must have changed the internal layout of their pages for whatever reason. It seems like the bot is just selecting the first block quote as the entire article.
On The Register for example it selects the div with the id
#body. For ABC it seems that it looks for the classArticle_Bodywhich I can’t find on that article. I might have a closer look later if I’ve got some time and try to get a PR in if it doesn’t get fixed.That’s the case, they removed one level of nesting from the html. Anyway, it doesn’t look for
Article_Bodyclass, but any class that starts withArticle_Body. They’re using randomized class names with the prefix being constant, that’s why I have to do it that way. I’ve updated it to this horrible looking selector:div[class*="Article_body"] > div > p, div[class*="Article_body"] > div > ul:not([class*="ShareUtility"]) > li.Thanks! I thought it might’ve been a wildcard thing but wasn’t sure. They really don’t want their articles summarised do they (or they’re probably trying to discourage AI scrapers)
Thanks for the report! It’s fixed now.