Translating wikimedia

From Apertium
Revision as of 08:33, 21 February 2011 by Unhammer (talk | contribs)
Jump to navigation Jump to search

To translate wikimedia formatted text, like Wikipedia articles, there are several options.

  • apertium-desmediawiki/-remediawiki in the apertium package is installed by default, but only handles the basics
    • Usage: cat es-article.wiki | apertium-desmediawiki | apertium -f none es-ca | apertium-remediawiki > ca-article.wiki
  • apertium-mediawiki -- a newer work-in-progress format handler that handles links better (see below)
  • Mediawiki Extension:Translate -- if you run a wiki, you can install this to use Apertium through it
but can you actually translate wikimedia-formatted articles with it, or just po-files and such?
  • you can run a clean-up script such as wikifix.pl after just using the regular format handler


apertium-mediawiki

This is a work-in-progress full format handler for mediawiki.

Download and compile like this:

   svn co https://apertium.svn.sourceforge.net/svnroot/apertium/trunk/apertium-mediawiki
   sh make.sh

There's no real makefile yet, so just put that directory in your $PATH, e.g. by adding PATH=/path/to/apertium-mediawiki:$PATH into your ~/.bash_profile.

Open a new terminal and you should be able to do

   cat es-article.wiki | demediawiki | apertium -f none es-ca | remediawiki > ca-article.wiki


See also