On Thu, Jan 21, 2016 at 3:12 PM Marco Fossati <[email protected]> wrote:

> Thanks Magnus for the pointers, the mix'n'match data are like jewels for
> StrepHit.
> Does the database you mentioned also contain the **body** of the
> catalogs items, i.e., the raw text of biographies?
> If so, I can avoid scraping all those sources, and it would be just
> perfect.
>

Sorry, no text body. Just these brief one-line descriptions alone are
touching a legally grey area (especially in Europe)...


> As a side note, I'm currently outreaching GLAM people to ask for more
> biographical sources: links are coming, and I'll definitely import them
> into mix'n'match as well.
>

Excellent!


>
> I was aware of the Sourcerer tool: I'm concerned with those references
> coming from Wikipedia articles though, since they stem from inside a
> Wikimedia project, and I want to make sure that everything comes from
> the outside.
>

The Sourcerer references do NOT come from Wikipedia! I am using third-party
sites for which we already have IDs (e.g. GND) to auto-validate values, and
add the appropriate reference if identical. Basically, what you want to do,
on the cheap ;-)



> I'm open for discussion with the community about this.
>
> What do you think?
>
> Cheers,
>
> Marco
>
> On 1/21/16 13:00, [email protected] wrote:
> > Date: Wed, 20 Jan 2016 16:44:46 +0000
> > From: Magnus Manske<[email protected]>
> > To: "Discussion list for the Wikidata project."
> >       <[email protected]>
> > Subject: Re: [Wikidata] Mix'n'match tool catalogues data
> > Message-ID:
> >       <
> caghuetalaz3ofx93olv45fy4sgffst+oslz9fyt5x1yvbbb...@mail.gmail.com>
> > Content-Type: text/plain; charset="utf-8"
> >
> > I also have a bot that can add references from various web sources:
> >
> https://bitbucket.org/magnusmanske/wikidata-todo/src/f56dfdaaaee053abaadefb584fcb4f714bc82545/scripts/autosource/botsource.php?at=master&fileviewer=file-view-default
> >
> > Edits so far:
> > https://www.wikidata.org/wiki/Special:Contributions/SourcererBot
> >
> >
> > On Wed, Jan 20, 2016 at 4:42 PM Magnus Manske<
> [email protected]>
> > wrote:
> >
> >> >Hi Marco,
> >> >
> >> >I run this tool. Quick answers:
> >> >
> >> >1. Yes. If you have a Labs account, you can see everything in database
> >> >s51434__mixnmatch_p . You can also get most of the data via the API
> >> >(undocumented; ask me for specifics, check out the requests of the
> >> >interface in the browser, or try the source code at
> >> >
> https://bitbucket.org/magnusmanske/mixnmatch/src/63c9ba58dd236e0aeb5a7ad12315047d787530f0/public_html/api.php?at=master&fileviewer=file-view-default
> >> >)
> >> >
> >> >2. Anyone can match entries to Wikidata items. I added most of the
> >> >catalogs, but you can also do that yourself at
> >> >https://tools.wmflabs.org/mix-n-match/import.php  .
> >> >
> >> >Cheers,
> >> >Magnus
> >> >
> >> >On Wed, Jan 20, 2016 at 4:31 PM Marco Fossati<[email protected]>
> >> >wrote:
> >> >
> >>> >>Hi everyone,
> >>> >>
> >>> >>The mix'n'match tool [1] provides a list of catalogues from different
> >>> >>sources with lots of biographical data.
> >>> >>
> >>> >>The list seems like a great starting point for the selection of
> reliable
> >>> >>sources that would feed the StrepHit pipeline [2].
> >>> >>
> >>> >>I was wondering 2 things:
> >>> >>1. Is it possible to directly access those datasets?
> >>> >>2. Who are the contributors that maintain the list?
> >>> >>
> >>> >>If you are involved into this effort, please get in touch with me.
> >>> >>Cheers,
> >>> >>
> >>> >>Marco
> >>> >>
> >>> >>[1]https://tools.wmflabs.org/mix-n-match/
> >>> >>[2]
> >>> >>
> >>> >>
> https://meta.wikimedia.org/wiki/Grants:IEG/StrepHit:_Wikidata_Statements_Validation_via_References
> >>> >>
> >>> >>_______________________________________________
> >>> >>Wikidata mailing list
> >>> >>[email protected]
> >>> >>https://lists.wikimedia.org/mailman/listinfo/wikidata
> >>> >>
> >> >
>
> _______________________________________________
> Wikidata mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikidata
>
_______________________________________________
Wikidata mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikidata

Reply via email to