Convert a wiki from MediaWiki to Gollum and Markdown, importing all metadata.

README.md

This will convert a wiki from MediaWiki to Gollum and Markdown (or any other format supported by Pandoc).

1. Install dependencies:

   brew install pandoc icu4c
   gem install --no-ri --no-rdoc hpricot gollum gollum-lib pandoc-ruby

2. Perform a Special:Export

3. Edit the constants at the top of the .rb for your particular needs.

4. ruby mw-to-gollum.rb

All inner links between pages with diacritics and spaces will be broken. You can list all links to be reviewed by searching for the string wikilink in all pages.

mw-to-gollum.rb

#!/usr/bin/env ruby
require 'rubygems'
require 'hpricot'
require 'gollum'
require 'gollum-lib'
require 'i18n'
require 'pandoc-ruby'


SOURCE = 'wiki.openfisca.fr-20160414143040.xml'
TARGET = '.'
DESTINATION_FORMAT = :markdown

I18n.config.available_locales = :en  # just to make it shut up, we want to use a static method anyway

wiki = Gollum::Wiki.new(TARGET)

file = File.open(SOURCE, 'r')

doc = Hpricot(file)

doc.search('/mediawiki/page').each do |page|
  title = I18n.transliterate(page.at('title').inner_text)

  page.search('revision').each do |revision|
    revision_id = revision.at('id').inner_text
    author = revision.at('contributor/username').inner_text
    original_time = revision.at('timestamp').inner_text
    comment = revision.at('comment') && revision.at('comment').inner_text

    begin
      content = PandocRuby.convert(revision.at('text').inner_text, :from => :mediawiki, :to => DESTINATION_FORMAT)
    rescue
      print '!'
      next
    end

    commit = {
      message: "#{comment}

Import from #{SOURCE} (rev #{revision_id})
Original modification made at #{original_time}",
      name:    author,
      email:   "#{author.tr(' ', '_')}@#{SOURCE}"
    }

    page = wiki.page(title)

    if page
      wiki.update_page(page, page.name, DESTINATION_FORMAT, content, commit)
      print '.'
    else
      wiki.write_page(title, DESTINATION_FORMAT, content, commit)
      puts
      puts title
    end
  end
end

puts
puts 'Done!'
file.close

awesome thanks for adding the revision support

Thanks! Worked well. I had to change line 27 to

author = revision.at('contributor/username') ? revision.at('contributor/username').inner_text : 'anonymous'