Hi,
I have a quick question: What is the best way to integrate wikipedia
into other applications without using the html interface? I would like
to integrate the search functionality and the content docs into my app.
Is there an xml interface to the wikipedia? I would like to bypass
screen scraping. Ideal would be a REST style interface.
The application is a public project where I try to connect webservices
in a random way:
http://konnexion.sourceforge.net/
I would love to integrate wikipedia.
Thanks for any feedback,
Stan Wiechers
This was helpfull. Do you know of any java framwework which transforms the raw wiki format into a document tree? Would that be usefull?
Thanks for the feedback,
Stan
-----Original Message-----
From: wikitech-l-bounces(a)wikimedia.org on behalf of Rowan Collins
Sent: Tue 1/11/2005 1:30 AM
To: Wikimedia developers
Subject: Re: [Wikitech-l] Wiki Webservices
On Mon, 10 Jan 2005 12:00:07 -0500, Stan Wiechers <Stan(a)rga.com> wrote:
> I have a quick question: What is the best way to integrate wikipedia
> into other applications without using the html interface? I would like
> to integrate the search functionality and the content docs into my app.
> Is there an xml interface to the wikipedia? I would like to bypass
> screen scraping. Ideal would be a REST style interface.
Well, there's "&action=raw" to get the raw wikitext of a page. Or look
at http://pywikipediabot.sf.net for a Python bot framework for
interacting with the site. Or download the database from
http://download.wikimedia.org and do really crazy stuff to it without
loading the wikimedia servers any.
But no, no XML interface. There's been discussion of code to make bot
access easier, but I can't find any of it, and I don't think it
resulted in much in the way of new code, I'm afraid.
--
Rowan Collins BSc
[IMSoP]
_______________________________________________
Wikitech-l mailing list
Wikitech-l(a)wikimedia.org
http://mail.wikipedia.org/mailman/listinfo/wikitech-l
> >create a "pdf icon" only when the implementation of the external.png is
>How do you want to detect whether an URL points to a PDF-file (i. e. not
>something that Internet Exlporer would executeable but is called
>"something.pdf")
>Hendrik
Hendrik,
as the subject says: I only propose to show an icon, I did not propose to scan file types of something extern. If that file is internal i.e. linked via the [[media:filename.pdf]] notation, then the server has potentially the chance to check the "magic" byte of that file; if it is external, i.e. [http://server/filename.pdf|othername], you can only argue from the filename.pdf that it pretends to be "PDF" . If you see a security problem, what does the (PDF.png vs. external.png ?) icon change in this respect ?
Nothing; but with my proposal you could even have ***two*** different PDF icons, a green "secure" icon for internal (filetype scanned) [[media:filename.pdf]] files and a red "insecure" alerting one ("extern; probably PDF") for the others. This is just meant as an ad-hoc proposal to show you the positive possibilities of my proposal.
Does this answer satisfy you ?
Tom
Hi,
Thank you Jakob for added me on this page
http://meta.wikimedia.org/wiki/Research, I will try to get some information on
it. Normally I will finish the redaction of the first and little part of my
work till middle february. Mainly in french but I will translate it in English
too.
Thank also for the link download.mediawiki.org, I take (firstly) the french
Wikipedia, cause 80G whoa !! I understand now why France is a little bit in
late for the IT society :/
But I have still a problem. Ok I have the articles and all history of them in
the database, but I don't see the "User:" pages. Maybe it's not the same with
en.wikipedia. Does it really contains the user pages? If this is not the case,
can we have access to these datas without being connected?
One more time ... thank you for your enthousiasm.
Julien Levrel
With a lot of experimental code that doesn't get looked over as closely
or as quickly as we might all like, I'm not willing to run a CVS
HEAD-based test.wikipedia.org on our main servers anymore. I've
discovered too many holes in code weeks after it went up on test to be
comfortable leaving unvetted code running on our production servers.
At some point we may set up a separate test wiki on one of the
non-main-cluster "test" boxes.
-- brion vibber (brion @ pobox.com)
Dear Gerard,
your wrote
>The current "external.png" implementation is horribly broken in
>languages that are right to left like Farsi, Hebrew, Arabic. So let's
>create a "pdf icon" only when the implementation of the external.png is
>fixed. Let's not add more garbage in a manner that we know is broken.
however, it doesn't matter, where that icon actually is (left or right
-- this can be programmable, dependent on the direction of writing).
My point is, do we want this or not ?
Tom
Frank v Waveren a écrit :
> On Thu, Jan 06, 2005 at 10:38:49AM +0100, Grégoire Colbert wrote:
>
>> $alh = trim( $_SERVER["HTTP_ACCEPT_LANGUAGE"] );
>
> This would be a horrible abuse of HTTP. You should only present the
> same content in different languages based on the Accept-Language
> field, never to different content. Having links point to different
> things based on browser settings would be a nightmare.
>
The fact that www.wikipedia.com does not care about the visitor's
language does not seem an "horrible abuse" to you? Only HTTP matters? If
so, you've got a strange vision of what a website is made for.
Grégoire
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hello,
I set up the nagios monitoring tool on larousse (thanks to brion and
innocence for the help).
The setup is somehow described on http://wp.wikidev.net/Nagios .
Interface access (need account):
http://noc.wikimedia.org/nagios/
To gain access you need to have root access, then read:
/h/w/doc/nagios-password
Also to complete a bit the installation I would need:
_ sudo configured to allow users in group wikidev (or nagios) to run as
root the commands:
/etc/init.d/nrpe (start|stop|restart|status)
_ I can not find the gd library development files on larousse, would
also need: png-devel and jpeg-devel libraries.
cheers,
- --
Ashar Voultoiz - WP++++
http://en.wikipedia.org/wiki/User:Hashar
Servers in trouble ? noc (at) wikimedia (dot) org
"This signature is a virus. Copy me in yours to spread it."
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFB4aJvpmyHQ2O4INERAoqxAKCpKcV+lKWgVJu9hbEKohb/MFuyzwCfTlys
2DRCZaeplmqEtHKjl4zy8aE=
=Fppf
-----END PGP SIGNATURE-----