There is no bot that covers all (or even many) kinds of bad id. I do a bad pixiv id sweep at the start of each month. BrokenEagleBot seems to have bad twitter id covered.
Posted under General
There is no bot that covers all (or even many) kinds of bad id. I do a bad pixiv id sweep at the start of each month. BrokenEagleBot seems to have bad twitter id covered.
Etou said:
I've noticed that bad_id posts don't seem to be automatically getting tagged as such. Is there an issue with the bot that usually applies this tag? Is this a known problem?
Post sources here, and then we can tell you if they're covered or not.
For myself, the sources that I do cover only regularly get checked at upload time, at 1 week, then at 1 month. Apart from that, I do a full check every few months or so.
bad_nijie_id doesn't seem to be covered, and I've seen several Twitter posts that have been deleted for several months not being tagged with bad_twitter_id, such as kawasemi27's art.
Etou said:
bad_nijie_id doesn't seem to be covered, and I've seen several Twitter posts that have been deleted for several months not being tagged with bad_twitter_id, such as kawasemi27's art.
It's been a while since I've done a full sweep of Nijie. I'll start one tomorrow.
As to the Twitter one, I'll pay close attention to that artist the next time I run a Twitter full sweep.
BrokenEagle98 said:
I just finished a full sweep of Nijie.
Sorry to bug you about this, but I think that some posts were missed. The source for post #2500861 looks like it was deleted, yet it wasn't marked as bad id. Same for post #2432233. I'm assuming it's probably an issue caused by several sources pointing to direct image links instead of the actual artwork pages.
It seems like there were also false positives, such as post #2520753.
Updated
Etou said:
Sorry to bug you about this, but I think that some posts were missed. The source for post #2500861 looks like it was deleted, yet it wasn't marked as bad id. Same for post #2432233. I'm assuming it's probably an issue caused by several sources pointing to direct image links instead of the actual artwork pages.
My script wasn't set up to check HTTPS image links, since back when I created the script, they were all HTTP. I've added HTTPS in and did a recheck.
It seems like there were also false positives, such as post #2520753.
That's not a false positive. It's pretty much damn near impossible to find the original post link if all you have is the image link*. It's not like Pixiv where the image link contains the post ID in the filename itself. Therefore, the script only checks the actual image link itself, and if that fails, then it gets marked as bad nijie id.
* Unless you scan every single image link from an artist, which I'm not going to do.