Robowiki Mirrors

List of ways to have back up or mirror of our beloved Robowiki.

To the best of my knowledge there is no easy and transparent way to run just a Wikimedia based wiki unless you have an administrative access to the underline database, i.e. you need administrative privileges on the hosting server.

One can obtain a full wiki dump (with use of dumpgenerator from wikiteam) and then reexport it to the local Wikimedia based instance but there is no way to make this dump incremental. So you would have to redownload about 700 MB (for 2018 Robowiki version) every time.

Working method

So far I succeeded (see Robowiki with Git-Mediawiki) using git-mediawiki to have a local replica of Robowiki. The main advantage that it is incremental so we do not need to redownload the full wiki. There is a dark spot: apparently attachments had to be redownloaded (if you want to keep track of them) every time. The good news it takes about 5 minutes to have up-to-date replica if you already made the initial import (which takes several hours even with fast connection). Additional advantage you can edit local files and push changes back to Robowiki. So if you are command line lover, this is the way. Ideally, we need to host this git repo somewhere (GitHub) so we can pull/push from it directly without stressing Robowiki with multiple api request thus reducing hosting expenses for the consumed traffic.

git-mediawiki gives local copy which currently takes about 530 MB and requires proficiency with command line tools.

Web browser accessible mirror

It would be nice to mirror this back to a web accessible interface for those who want to browse from the web and do not have git skills.

I have partial success to converting git-mediawiki to ikiwiki: some expected rendering problems of the templates and more importantly missing discussions export.

Other possibilities

• it would be super cool to do something along the distributed web projects
  • IPFS
  • Dat Project then we can even host the mirror from our browsers like Beaker

The distributed web seems to be good for read only, and I am not sure how to put updates in. IPFS seems to be designed for a static content.