Hi,

   I have a mapping dump handy from ~ last February, I think it makes
   no difference but if you want a fresh one I can generate one for you.

That's great, an old dump is fine, we're just using it for test purposes. It would be great to also have the mediawiki code in case it has some changes or it's an old version.

   The template mapping is also a good idea, the Airpedia guys already
   generated a list of mappings, Maybe your script reuses their output?

Yes, Including the Airpedia mappings in the wiki is a great idea. We will build some intermediate pages however, where people can revise the mappings and commit them in bulk. Including them directly without editor revision might confuse things.

We're close to being done with the bulk label updates, inserting the Airpedia mappings would be more or less similar to the labels. Right now we're looking into assessing the workload of implementing things as mediawiki plugins, especially displaying more info at mapping creation. However, I fear we'll have to do it as separate tools first and use API calls for most things.

Cheers,
Alexandru






On 02/21/2014 09:38 AM, Dimitris Kontokostas wrote:
Hi Alexandru,

I have a mapping dump handy from ~ last February, I think it makes no difference but if you want a fresh one I can generate one for you.

The template mapping is also a good idea, the Airpedia guys already generated a list of mappings, Maybe your script reuses their output?

btw, great to have you in GSoC ;)
You can add your self as a mentor here http://wiki.dbpedia.org/gsoc2014/
add yourself next to the ideas you like here: http://wiki.dbpedia.org/gsoc2014/ideas and subscribe to the gsoc list : https://lists.sourceforge.net/lists/listinfo/dbpedia-gsoc



On Mon, Feb 17, 2014 at 4:04 PM, Alexandru Todor <to...@inf.fu-berlin.de <mailto:to...@inf.fu-berlin.de>> wrote:

    Hi Dimitris,


    Having worked intensively on the mappings in the last couple of
    weeks, this is a feature I definitely want to implement asap.
    I've looked at the mediawiki API and it seems straight forward,
    although very sparsely documented [1],[2],[3] (if anyone knows
    more about the mediawiki api, please reply)

    What I would really love to have from you is a backup of the
    mappings wiki code and database. The database can be anonymized by
    excluding all password hashes except for the user ThermoS. I don't
    want to mess around with the mappings wiki directly, working with
    a backup is safer.

    Furthermore I think it would be good to extend this approach for
    the template mappings themselves and not only ontology labels.

    The workflow for mappings right now looks like this:
    - loop at mappings statistics and pick out a template that hasn't
    been mapped
    - look at the template in Wikipedia
    - look if there is an english version of the template
    - look if the template has been mapped in the english namespace
    - create a mapping that maps to the same class and matching properties
    - if there isn't a matching template / mapping do it your own way
    - in case there is a mapping in another namespace that mapps to a
    semantic equivalent class, now you have multiple classes/
    properties in the ontology for the same template

    I started implementing a script that goes trough the mapping wiki,
    checks all templates that have been mapped in  one namespacace,
    and if there is an equivalent template in another language, that
    hasn't been mapped yet, it posts a link to the other mapping
    namespaces to look at how it has been mapped in the other
    languages. Furthermore I want to show all the information need to
    make a mapping on one page, that is: local wikipedia entry for the
    template, wikipedia entry for the template that has been already
    mapped, the existing mapping in a different language, the dbpedia
    ontology and the mapping edit box.
    Of course if you do this you can also semi-automatically create
    mappings with the help of google translate and some statistical
    analysis, but in the end an editor should still look over if it
    makes sense.

    I would also like to apply to be a CO-Mentor or advisor for the
    GSOC points 4.4 and 4.9 (if it's not too late) since it would be
    good to have any tools that eventually get developed there
    integrated with this functionality.


    Cheers,
    Alexandru

    [1] http://www.mediawiki.org/wiki/API:Main_page
    [2] http://www.mediawiki.org/wiki/API:Changing_wiki_content
    [3] http://www.mediawiki.org/wiki/API:Tutorial





    On 02/15/2014 09:20 AM, Dimitris Kontokostas wrote:
    Hi Alexandru,

    That would be a very cool feature actually. I don't have any
    experience with the mw api but I can tweak [1] to get things in
    the format you want.
    I can also give you bot access in the mappings wiki - after we do
    some testing first of course;)
    btw, [1] is a page generated from the extraction framework
    (Server module) after we parse all the mappings & the ontology.

    Cheers,
    Dimitris


    On Fri, Feb 14, 2014 at 8:18 PM, Alexandru Todor
    <to...@inf.fu-berlin.de <mailto:to...@inf.fu-berlin.de>> wrote:

        Hi,

        We're currently improving the German mappings, and adding
        missing labels is proving really boring since you basically
        have to make hundreds of edits for each
        class/property/datatype that is missing one. This wastes a
        considerable amount of time. I was thinking of digging a bit
        trough the Wikimedia API, printing a list of all the
        classes/properties/datatype contents (english labels mostly)
        that are missing labels, translating them all in one go and
        then committing everything back trough the API. I know that
        some people allready have some experience with this since we
        have pages like in [1] . If someone has any good tips on
        where to start to implement such functionality, or if there
        already is a way to do add mappings via an API, I would be
        very grateful for the help.


        [1]
        http://mappings.dbpedia.org/server/ontology/labels/missing/de/


        Kind Regards,
        Alexandru Todor

        Alexandru-Aurelian Todor
        Freie Universität Berlin
        Department of Mathematics and Computer Science
        Institute of Computer Science
        AG Corporate Semantic Web
        Königin-Luise-Str. 24/26, room 116
        14195 Berlin
        Germany



        
------------------------------------------------------------------------------
        Android apps run on BlackBerry 10
        Introducing the new BlackBerry 10.2.1 Runtime for Android apps.
        Now with support for Jelly Bean, Bluetooth, Mapview and more.
        Get your Android app in front of a whole new audience.  Start
        now.
        
http://pubads.g.doubleclick.net/gampad/clk?id=124407151&iu=/4140/ostg.clktrk
        _______________________________________________
        Dbpedia-discussion mailing list
        Dbpedia-discussion@lists.sourceforge.net
        <mailto:Dbpedia-discussion@lists.sourceforge.net>
        https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion




-- Kontokostas Dimitris


    
------------------------------------------------------------------------------
    Android apps run on BlackBerry 10
    Introducing the new BlackBerry 10.2.1 Runtime for Android apps.
    Now with support for Jelly Bean, Bluetooth, Mapview and more.
    Get your Android app in front of a whole new audience.  Start now.
    http://pubads.g.doubleclick.net/gampad/clk?id=124407151&iu=/4140/ostg.clktrk
    _______________________________________________
    Dbpedia-discussion mailing list
    Dbpedia-discussion@lists.sourceforge.net
    <mailto:Dbpedia-discussion@lists.sourceforge.net>
    https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion




--
Kontokostas Dimitris

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Dbpedia-discussion mailing list
Dbpedia-discussion@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion

Reply via email to