In the past I've been a big fan of Pandoc for converting text from
one format to another, including HTML->Mediawiki.
The University of the Highlands and Islands (just had to spell it out,
because that seems like such a *great* name for a university) had the
start of a small course in HTML format. Each page was short and
reasonably regular (with the exception of video). I decided to use a
quick iojs ES6 script to scrape the pages and create a 1-for-1
wikipage mapping (minus the non-free images).
This was NOT intended as a general solution, but I think the code is
simple enough it might serve as a model for others wanting to do
something similar. I used cheerio to mangle the HTML into usable
The code is in the wikieducator project on gitorious.
 Pandoc is the Swiss Army knife of text conversion tools
 cheerio is a subset of jQuery for HTML manipulation on the server
 gitorious yesterday announced they are shutting down at the end of
May, so expect our project to move
Jim Tittsler Tokyo