Right, baker changes are a problem then.
They would have to get formatted in a certain way, so I can detect them.
Tripcodes are blacklisted except for Q on Qresearch? That sucks.
>BTW, do you know if there is a way to archive threads so that the mp4's are still playable?
No.
I doubt that's even possible like that.
Would have to look into it on archive.org.
I'm not even sure if archive.org will archive .mp4s.
Test Video
Okay, I just tried and archived this manually.
http://web.archive.org/web/20200517203014/https://8kun.top/comms/res/13590.html
I archived the last pic too "muh fren" and it seems to work even on archive.org. I can click on it, and it changes to a larger pic.
For me it works, even for videos.
Check this out:
http://web.archive.org/web/20200517203014/https://8kun.top/comms/res/13590.html
It doesn't play embedded, but when you right click on the first URL above the pic/movie (which points to media.8kun.top), it even plays the video.
I also just found a simple API to trigger archiving of websites. I will have to look into it, I guess sending too many requests at once will make it reject the requests.
So at least in theory this could all work somewhat, but the one problem remains - how to detect baker changes.
Even if it was possible, I wouldn't want to archive all photos/videos of every bread. Doing that would be silly and the risk of saving illegal content would also be high. I have no idea if that could even get me into legal trouble.
I would prefer being able to identify baker and archive all posts tagged by baker(s).
Will now just first of all write code to read the catalog and archive threads into my local database, including doing incremental updates of the threads.
Another question also remains: how to identify Qresearch generals.
I also wonder when I shall start archiving breads. The General never leaves the catalog unless it's full. But for other threads that may happen, and I think the pictures get lost as soon as it falls off the catalog. This means I should archive it way earlier than that.
Test multiple pics
Hmmm, another idea would be to always identify baker by first post, but then also go reverse from the next bread, looking for post numbers mentioned on there by baker until dough, and identifying baker changes that way.
This wouldn't require a fixed protocol.
The only problem would be shill breads.
Do you have any experience how fast shill breads are removed? Probably never?
Does it even happen that often?
Dough post is fixed, right?
At least it should be a link to pastebin.