Close
Related: forum #18869

Checkmate: any chance to fix them this time? Since we have the all the uncorrupted files now.

By the way, does sankakucomplex literally grab every single image from here?
I've replaced the ones with replacement. Still missing for post #32333 though it doesn't seem to be different compared to its parent? (nuke from orbit?)
edogawaconan said:
I've replaced the ones with replacement. Still missing for post #32333 though it doesn't seem to be different compared to its parent? (nuke from orbit?)
And that file doesn't even look like corrupted..
edogawaconan said:
I've replaced the ones with replacement. Still missing for post #32333 though it doesn't seem to be different compared to its parent? (nuke from orbit?)
That one is also available on sankaku complex https://chan.sankakucomplex.com/post/show/475679
jkdfahlkdhfjgakdhfg said:
That one is also available on sankaku complex https://chan.sankakucomplex.com/post/show/475679
not the same one. That's post #31450. But like what edogawa said, there is no point to keep post #32333 since it's a duplicate of post #31450. And we can just nuke it for sanity's sake.
I've changed the Type of md5_mismatch to Fault, as it's related to something wrong within an image.
fireattack said:
By the way, does sankakucomplex literally grab every single image from here?
I have lists of MD5s from both Yande.re and Sankaku Complex. Between the two, there's an overlap of 262287 images (out of Yande.re's 346087, so 75%). When I filter Yande.re's images to exclude deleted (299052 images), the overlap is 246554, or ~82% of Yande.re's corpus.

They don't take literally everything but they take most of it. I checked against images uploaded since 2015 as well and Sankaku has 29499 out of 39151 so it looks like they mainly have old stuff.
Nice work there. Well done.