Hacker News Reader

Comment by buildbuildbuild | Hacker News Reader

Caveat: the last full Kiwix English Wikipedia archive was made in 2018. They could use some help with automating their build process if anyone here has the time.

View on HN

From a cursory glance at the site and source code, it's really hard to see who/what is involved with building an archive. There's automated builds set up for the Pi image itself.

Agreed. I can see that other Wikipedia languages are crawled - https://wiki.kiwix.org/wiki/Content_in_all_languages shows dozens of updates this week - but the best leads I have involve poking around the openZIM Github org, https://github.com/openzim . There might be a running "zimfarm" somewhere?

Looks like you are right, you may be able to join the farm to help ( i have not tested as I am away from my computer at the moment )

https://github.com/openzim/zimfarm/blob/master/worker/README...

Someone should ask drone.io or packet.net for an Epyc machine to do this.

You can build your own ZIMs from any MediaWiki instance using this tool: https://github.com/openzim/mwoffliner.

Maybe it would be worth putting together another zimfarm that is constantly updating.

The last time I checked, it was more a problem of lacking servers with sufficient resources: https://phabricator.wikimedia.org/T124960 https://phabricator.wikimedia.org/T219078

It sure doesn't harm if someone creates their own ZIM files and reports on their results (and/or shares the resulting files).

Reply on news.ycombinator.com