Danbooru

Import posts from Sankaku

Posted under General

incubatedveg said:

is there any quick way to upload posts from Sankaku and keep all of their tags? Seems like a good idea to do so, but maybe there's a reason why its not.

Sankaku has a habit of being extremely hostile to anyone who tries scraping "their" content with things like takedown requests. In spite of this, most of "their" content that is both good enough quality to be on Danbooru and has sufficient tagging is just scraped from us anyway. Because of Danbooru will likely never natively support uploading from Sankaku. Best you could do is write some Python script to aide you, assuming there's anything worthwhile on there.

Basically what Talulah said, though there is still a substantial amount of content over there that would be good enough to have here. The issue then comes from having to identify what wasn't scraped from elsewhere, whether it is of decent quality, whether it is actually even sourced, etc. The effort needed to basically scrape Sankaku Channel of everything it is worth is too immense and time-consuming.

If a day were to come where we finally achieved something of that sort, then I could only say one thing: Sankaku Channel delenda est! Burn that garbage website to the ground!

Talulah said:

Sankaku has a habit of being extremely hostile to anyone who tries scraping "their" content with things like takedown requests. In spite of this, most of "their" content that is both good enough quality to be on Danbooru and has sufficient tagging is just scraped from us anyway. Because of Danbooru will likely never natively support uploading from Sankaku. Best you could do is write some Python script to aide you, assuming there's anything worthwhile on there.

thats weird, so both sites have chosen to work against each other to achieve optimal image tagging? lame. You say danbooru already scrapes some sankaku data, but ive just uploaded a bunch of images from sankaku that werent here apparently, and used the sankaku tags.

incubatedveg said:

thats weird, so both sites have chosen to work against each other to achieve optimal image tagging?

We can't support Sankaku uploading because they actively sabotage any effort of having a functional api. Even their direct image links have an expiration key to avoid reposting.

Especially now that they're pushing the garbage that is their beta site, any attempt at supporting Sankaku in the upload page would be a never ending cat and mouse game.

Also, you uploaded image samples. You should get the original source and upload that instead.

incubatedveg said:

thats weird, so both sites have chosen to work against each other to achieve optimal image tagging? lame. You say danbooru already scrapes some sankaku data, but ive just uploaded a bunch of images from sankaku that werent here apparently, and used the sankaku tags.

You got that backwards. Sankaku scrapes a fair bit of their content from us, and the stuff that isn't is mostly mass uploaded by bots with no quality control or tagging. There are outliers that are well-tagged, sufficient quality, and aren't paid rewards, but finding those is looking for a needle in a haystack.

I'm not against it, but as mentioned it's not necessarily easy because Sankaku actively resists scraping by others. Which is frustrating because they themselves scrape large volumes of content from other sites.

incubatedveg said:

oh shit. is there a guide somewhere on how to replace them? thanks for the response :)

Use SauceNao to try to find the original source before uploading. The Image Search Options browser extension makes this a simple right click->search operation.

Dolmatov said:

Error. The child post #5262678. The image cannot be replaced (with the replace function) because the original is already loaded. We are waiting for the end of the verification period. "This post is pending approval."

Again, please stop replying in the forums if you can't speak english without google translate. You're doing more harm than good with this kind of gibberish.

incubatedveg said:

i dont understand? whats the problem lol

You uploaded a sample of a picture we already have.

Updated

1