I haven’t had a look at the kwakk.info logs in a while, but I had a peek today and saw something really odd. The pattern is this:
GET /res/search/eng/BODOI/NONE/the.json GET /res/search/eng/BODOI/NONE/and.json GET /res/search/eng/BODOI/NONE/of.json GET /res/search/eng/BODOI/NONE/in.json GET /res/search/eng/BODOI/NONE/is.json GET /res/search/eng/BODOI/NONE/for.json GET /res/search/eng/BODOI/NONE/with.json GET /res/search/eng/BODOI/NONE/was.json GET /res/search/eng/BODOI/NONE/that.json GET /res/search/eng/BODOI/NONE/on.json ...
That is, this bot fires off searches (kwakk.info is a search engine for magazines about comice) for common words, about a dozen per second. Curious!
But then! It switches around and does:
GET /res/words/0be3bae149351808ab66622eb39d801667d495ccb0b0b85c4766e5b17591408c.json GET /res/image/0be3bae149351808ab66622eb39d801667d495ccb0b0b85c4766e5b17591408c.jpg GET /res/words/aa1dff35f83db8355ab7d69f5e29ef39bbce74774f9387b7d3144e7fc9032f85.json GET /res/image/aa1dff35f83db8355ab7d69f5e29ef39bbce74774f9387b7d3144e7fc9032f85.jpg GET /res/words/a9b634e0b4a1a26702fde0706d670864749282b4ddc55f665b233efb89a01a66.json GET /res/image/a9b634e0b4a1a26702fde0706d670864749282b4ddc55f665b233efb89a01a66.jpg GET /res/words/44de1fb7aa7c3ce59d35ec16315609935f80f9c5ff3d650e7c743f6698cf3626.json ...
I.e., it downloads a bunch of scanned pages and the OCR text for those pages.
Aha! Because for each kwakk.info search result, you can view five pages (i.e., you get the hashed IDs). So this bot is firing off searches for common words to get as many results as possible, and then for every result it gets, it downloads the five scanned magazine pages it now has access to. So per search you get 200 hits, times 5 magazine pages, so *counts on fingers* 1000 magazine pages per search.
L33t hax0rs!
I guess what’s going on is that there’s an AI bot that wants to train on magazines about comics?
So that’s fun, and I certainly don’t object or anything… It’s just kinda sad, because all the magazines on kwakk.info are readily available in convenient formats on archive.org, libgen.li or Anna’s Archive, really. (Well, OK, some are only available using eMule, but…)
And there you get the complete magazines with the pages in correct order and everything.
www@quimby:~$ grep "Mozilla/5.0 (Macintosh; Intel Mac OS X 15.7; rv:151.0) Gecko/20100101 Firefox/151.0" /var/log/apache2/access.log | wc -l 178764
Just a couple hundred thousand hits from this bot today. I guess it’s going to have to do millions of searches to get all the pages… 🤷
But… like… sure? You figured wanted to download stuff, and then you figured out how kwakk.info works (or had Claude do it for you), so… Congratulations?