Hi all
I have enabled a new version of WikiProxy which uses the text databases on clematis to fetch page content. It has a couple of advantages over river's version:
* you don't need to know the revision id to fetch the latest version of a page (where "latest" is the latest version according to the toolserver's db, so it is consistent if the wiki's database is lagged) * you can use prefixed page names, or alternatively a namespace id separate from the page title * it automatically falls back to pulling the page content via HTTP should the text database fall behind * it provides some meta-info about the page in the http header * I'm also planning an alternative access path to avoid the overhead php-startup induces - maybe a basic web server written in php, maybe something involving fifos... not sure yet. Also don't know when it will come about.
@river: does your version prevent access from outside the toolserver? that would be important.
To use WikiProxy, you can go to http://tools.wikimedia.de/~daniel/WikiSense/WikiProxy.php and fill in the form. This will only work from the toolserver, though - for access from the outside, you need an access token, available on request.
To use it programmatically, use a URL like the following:
http://tools.wikimedia.de/~daniel/WikiSense/WikiProxy.php?wiki=de.wikipedia....
additional parameters: * rev=12345 if you want a specific revision * ns=5 if title is an un-prefixed page name (you can also use canonical or local namespace names)
The wiki parameter accepts names like "de", "dewiki", or "de.wikipedia.org".
Please try it and tell me if you experience any problems.
Enjoy, Daniel