Coders, I have a question. Or, a request, etc. I spent some time with xmc discussing something we could do to make things easier around here. What we came up with is a trigger for a bot, which can be triggered by people with ops. You tell it a website. It crawls it. WARC. Uploads it to archive.org. Boom. I can supply machine as needed. Obviously there’s some sanitation issues, and it is root all the way down or nothing. I think that would help a lot for smaller sites Sites where it’s 100 pages or 1000 pages even, pretty simple. And just being able to go “bot, get a sanity dump”
ArchiveBot has two major backend components: the control node, which runs the IRC interface and bookkeeping programs, and the crawlers, which do all the Web crawling. ArchiveBot users communicate with ArchiveBot by issuing commands in an IRC channel.
User’s guide: http://archivebot.readthedocs.org/en/latest/ Control node installation guide: INSTALL.backend Crawler installation guide: INSTALL.pipeline
Copyright 2013 David Yip; made available under the MIT license. See LICENSE for details.
Thanks to Alard (@alard), who added WARC generation and Lua scripting to GNU Wget. Wget+lua was the first web crawler used by ArchiveBot.
Thanks to Christopher Foo (@chfoo) for wpull, ArchiveBot’s current web crawler.
Thanks to Ivan Kozik (@ivan) for maintaining ignore patterns and tracking down performance problems at scale.
Other thanks go to the following projects:
Dragonette, Barnaby Bright, Vienna Teng, NONONO.
The memory hole of the Web has gone too far. Don’t look down, never look away; ArchiveBot’s like the wind.