It is very easy to use Wikipedia data with MoinMoin (not more difficult than with MediaWiki).

You can guess how flexible MoinMoin is - I wrote 30 lines of code (the database adapter) to pull the articles. Try that with any other Wiki.

the MediaWiki parser is not officially supported and will block some functionality like InterWiki linking and smileys in your wiki.

• What do you mean by "block" here?

How to set up such a wiki

• Get MoinMoin 1.3 - at least 1.3.4

• Apply the Virtual Pages patch from MoinMoinPatch

• Install the MediaWiki parser (see ParserMarket or Merten-home.de)

  • Set it up as the default parser (see HelpOnConfiguration)

• Install the Monobook theme (see ThemeMarket)

• Install MySQL and MySQLforPython

• Load the official database dump into MySQL (see http://download.wikimedia.org/)

  • If you want to speed up the import, replace all keys in the table definition by:

       UNIQUE KEY `name_title_dup_prevention` (`cur_title`, `cur_namespace`)

• Add 30 lines of code that access that database. You have to put it into data/plugin/virtual.

  •    1 import MySQLdb
       2 import traceback, sets
       3 
       4 regex = r".*"
       5 dbpages = sets.Set()
       6 hdpages = sets.Set()
       7 
       8 conn = MySQLdb.connect(host = "localhost", user = "root", passwd = "", db = "wp")
       9 
      10 def convPagename(pn):
      11     return conn.escape(pn.encode("iso-8859-1", "replace").replace(" ", "_"))
      12     
      13 def filter(pagename, mo): # set frontpage FIXME
      14     if pagename in ("", "FrontPage", "StartSeite") or pagename in hdpages:
      15         return False
      16     if pagename in dbpages:
      17         return True
      18     cursor = conn.cursor()
      19     rows = cursor.execute("SELECT 1 FROM cur WHERE cur_title = " + convPagename(pagename))
      20     cursor.close()
      21     if (rows > 0):
      22         dbpages.add(pagename)
      23         return True
      24     else:
      25         hdpages.add(pagename)
      26         return False
      27 
      28 def execute(request, match):
      29     print "Fetching " + request.page.page_name
      30     #traceback.print_stack()
      31     cursor = conn.cursor()
      32     cursor.execute("SELECT cur_text FROM cur WHERE cur_title = " + convPagename(request.page.page_name))
      33     row = cursor.fetchone()
      34     if row is not None:
      35         text = row[0]
      36     else:
      37         text = "Page not found."
      38     cursor.close()
      39     return text
    

    dbadapter.py

How big is it

• the English Wikipedia media data is 76GB
• texts are about 3 GB (all revisions)

• for downloading also take a look on Wikipedia:Database_download for uptodate information.

Limitations

The main feature of MoinPedia is currently viewing pages. The following things do not work:

• Editing pages (Virtual pages plugin limitation)

• Searching for text

• Numerous markup things are not understood by the MediaWiki parser for MoinMoin

• Pages with particular Unicode pagenames cannot be accessed correctly if MySQL is not configured correctly

Discussion

1. What about a solution that pulls the articles and puts them into Moin Wiki Pages?
   • That won't work with the current storage system.

     • I thought a thing like this would bei nice: Put only this in a page:  [[WikiPedia(MoinMoin Wiki)]]  This should fetch the page and convert it into Moin Wiki code. So this could work together with InterWikI!

       • Yeah, you could write a macro that fetches the page and outputs a link. The mediawiki parser will handle the markup then.
2. I suppose the patch does not work in 1.5 ? If we had (1.) we could make this version independend.
   • It should work with minor changes.
3. Does anybody have a Wikipedia mirror with Moin online? Could we link to it? -- Thilo