MediaWiki/archive/embedding

Overview

This page is about embedding MediaWiki content within non-MediaWiki sites, or as standalone pages without the standard navigation and control links provided by most MediaWiki skins.

non-PHP

This is easiest to do if the other application is also written in PHP, but it is possible to do this using other languages as well. Perl has a library, PHP::Interpreter, that allows it to execute PHP code; other languages may have similar capabilities.

If they do not, then one can still accomplish a certain amount by loading MediaWiki pages using http and the "?action=render" directive, which displays the page's bare contents without any of the navigation aids. (This method is not recommended if you have any alternatives, as it is slow and wasteful of CPU resources.)

PHP

This works with MediaWiki v1.16:

• /1.16/index.php

These have been tested with MediaWiki v1.14.

• /1.14/index.php is a usage example; as given, it provides a standalone page, but it also shows the code needed to retrieve page content
• /1.14/extract.php is the code which actually does the work of extraction; the function GetWikiPage($iTitle) returns the page's contents as a string.

Configuration info is in index.php. Note that this has not been tested very much; I got it working on my server, then posted it.

Works on MW 1.10:

• /old/extract.php/v2

Based on MW somewhere around 1.7:

• /old/extract.php/v1

To Do

• I need to work out the code for detecting if a given page exists in the wiki, so the main application can display default values other than the wiki's default ("This page doesn't exist yet..."). Detecting this with the method below may or may not be possible; if MediaWiki returns a 404 code for non-existent articles, then that could be used, but I don't yet know if it does. --Woozle 18:52, 15 October 2006 (EDT)
• We don't yet know how to get MediaWiki to emit fully-qualified URLs; it is currently necessary to do a search-and-replace to fix links targeting the wiki. (It's probably a configuration thing -- $wgArticlePath, I'd guess. Don't know if link URLs get cached, though, so that might not work reliably. --Woozle 19:29, 6 June 2009 (EDT))