Re: [Wikidata-l] Bot request: 250+ thousands person data
Hoi,
When Wikipedia has an approach to specific articles that are not compatible
with Wikidata, we can create items that fit our need and keep the original
item for what it is .. for instance a list of people (in the case of the
Wright brothers).
The notion that Wikidata defers to Wikipedia is not one can keep because
there are bound to be Wikipedias who differ in their approach and have an
article for both Wilbur and Orville Wright..
Yes, it is good to have a hope for algorithms in the future, in the mean
time consider what percentage is wrong and that quite often not having data
is more damaging than having data that can be manipulated with queries,
tools. No data is no grip at all. We do have queries in WDQ/Autolist and we
have tools in ToolScript and pywikipedia.
IMHO the most important thing we should do to get better quality is report
on differences. This helps all projects involved in an import / export /
comparison.
Thanks,
Gerard
On 29 April 2014 09:15, John Mark Vandenberg jay...@gmail.com wrote:
On Sun, Apr 27, 2014 at 8:28 PM, Amir Ladsgroup ladsgr...@gmail.com
wrote:
there are some problems in using bio template for example they used it
for a
group of people
https://it.wikipedia.org/wiki/Fratelli_Wright
This is quite a difficult problem. Also look for infoboxes not at the
top of a page, because the Wikipedia page contains two concepts. Here
is an example with {{Bio}}:
https://it.wikipedia.org/wiki/Slashdot
In the journals area, I faced this many times with the article about a
society not having an infobox for the society, but including an
infobox in a section for their primary journal .
My bot has some very hacky code to detect the infobox type in a few
languages
https://www.wikidata.org/wiki/User:JVbot/periodicalbot.py
(the first function)
It would be good if we can create an algorithm that detects all these
anomalies, or a special hidden parameter added to the invocation, to
exclude those templates from automated parsing, but also lists all
pages like this so that those pages can be split on the Wikipedias
(unless notability rules prevent the split).
--
John Vandenberg
___
Wikidata-l mailing list
Wikidata-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-l
___
Wikidata-l mailing list
Wikidata-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-l
Re: [Wikidata-l] Bot request: 250+ thousands person data
2014-04-28 17:08 GMT+02:00 David Cuenca dacu...@gmail.com: On Mon, Apr 28, 2014 at 3:10 PM, Luca Martinelli martinellil...@gmail.com wrote: I recalled the fact quite correctly: https://it.wikipedia.org/wiki/Modulo:Bio takes dates of birth and death from Wikidata. I think we can talk to extend the possibility to gender, and later to other fields. That's perfect, because that means that the bot can just delete the text on import. I would say -1 for the moment. We first need to talk about it and create hidden categories in order to control the retrievals. There's time to delete. :) -- Luca Sannita Martinelli http://it.wikipedia.org/wiki/Utente:Sannita ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
Re: [Wikidata-l] Autocomplete API for a search field in an external app
Hi, the autocomplete box is using the wbsearchentities api. You could try to copy the jquery.wikibase.entityselector.js from wikibase/lib/resources/jquery.wikibase it looks like it works independently from wikibase. It needs the jqueryui suggester. best regards, Christian 2014-04-28 18:40 GMT+02:00 Maxime Lathuilière gro...@maxlath.eu: Hi, I'm looking for a way to reproduce the autocomplete on wikidata search field in an external app (to tag resources with wikidata entities ids), but so far I couldn't find what could allow me to do this in the API doc. In a little experiment https://github.com/maxlath/wikidata-autocomplete, I tried to use the API action *wbsearchentities*, but it appear to be inefficient until I reach the full name... Is there already a way to query such an autocomplete API or any plans to implement one? Or even better, is there a way to easily re-use the wikidata search field widget? in a non-PHP project? :D Any clue welcome :) Best regards, Max -- Maxime Lathuilière maxlath.eu @maxlath Zorglub27 https://www.wikidata.org/wiki/User:Zorglub27 Contributionshttps://www.wikidata.org/wiki/Special:Contributions/Zorglub27 ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
Re: [Wikidata-l] Autocomplete API for a search field in an external app
Hi Maxime, After some quick reverse engineering of the site with the Chrome Developer Tools, here's the API it is using: http://www.wikidata.org/w/api.php ?callback=[YOUR CALLBACK NAME] action=wbsearchentities format=json language=en type=item continue=0 _=[TIMESTAMP (as a cache buster)] search=[YOUR QUERY] If you don't need the callback, then the API is as follows (note the missing cache buster): http://www.wikidata.org/w/api.php action=wbsearchentities format=json language=en type=item continue=0 search=[YOUR QUERY] Hope this helps. Best, Tom -- Thomas Steiner, Employee, Google Inc. http://blog.tomayac.com, http://twitter.com/tomayac -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.12 (GNU/Linux) iFy0uwAntT0bE3xtRa5AfeCheCkthAtTh3reSabiGbl0ck0fjumBl3DCharaCTersAttH3b0ttom.hTtP5://xKcd.c0m/1181/ -END PGP SIGNATURE- ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
Re: [Wikidata-l] Autocomplete API for a search field in an external app
Small correction: If you don't need the callback, then the API is as follows (note the missing cache buster): http://www.wikidata.org/w/api.php action=wbsearchentities ?action=wbsearchentities (replace '' with '?', copy and paste oversight, sorry). -- Thomas Steiner, Employee, Google Inc. http://blog.tomayac.com, http://twitter.com/tomayac -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.12 (GNU/Linux) iFy0uwAntT0bE3xtRa5AfeCheCkthAtTh3reSabiGbl0ck0fjumBl3DCharaCTersAttH3b0ttom.hTtP5://xKcd.c0m/1181/ -END PGP SIGNATURE- ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
Re: [Wikidata-l] Bot request: 250+ thousands person data
Am 29.04.2014 03:53, schrieb Amir Ladsgroup: It's not a big deal, parsing it would be no problem, I can use it in parsing data from Bio template in Italian Wikipedia but I have to use precision argument in snak. Am I right? Yes, exactly. what value have to set for precision if I have just year (and no month and day)? If you just have they yea, the precision value would be 9. This is arbitrary and obscure, sorry. I have filed a bug to fix this: https://bugzilla.wikimedia.org/show_bug.cgi?id=64593 For reference, here is the table of precisions to be used for time values, as defined in the TimeValue class: const PRECISION_Ga = 0; // Gigayear const PRECISION_100Ma = 1; // 100 Megayears const PRECISION_10Ma = 2; // 10 Megayears const PRECISION_Ma = 3; // Megayear const PRECISION_100ka = 4; // 100 Kiloyears const PRECISION_10ka = 5; // 10 Kiloyears const PRECISION_ka = 6; // Kiloyear const PRECISION_100a = 7; // 100 years const PRECISION_10a = 8; // 10 years const PRECISION_YEAR = 9; const PRECISION_MONTH = 10; const PRECISION_DAY = 11; const PRECISION_HOUR = 12; const PRECISION_MINUTE = 13; const PRECISION_SECOND = 14; If you have something like between 1846 and 1855, you can use the before and after fields of the time value: time: +0001850-00-00T00:00:00Z, precision: 9, before: 4, after: 5 This means the main value is 1850, given as a year, with a lower bound four years before and an upper bound 5 years after the main value (before and after are given in the unit specified by the precision value). The main value is what is going to be displayed per default; it will also be used for sorting query results (once we have queries). This is a bit complicated, but should allow you to actually represent uncertain dates. We made it so you can be precise about the uncertainty :) HTH Daniel -- Daniel Kinzler Senior Software Developer Wikimedia Deutschland Gesellschaft zur Förderung Freien Wissens e.V. ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
Re: [Wikidata-l] Bot request: 250+ thousands person data
Il 29/apr/2014 09:31 Gerard Meijssen gerard.meijs...@gmail.com ha scritto: Hoi, When Wikipedia has an approach to specific articles that are not compatible with Wikidata, we can create items that fit our need and keep the original item for what it is .. for instance a list of people (in the case of the Wright brothers). The notion that Wikidata defers to Wikipedia is not one can keep because there are bound to be Wikipedias who differ in their approach and have an article for both Wilbur and Orville Wright.. Exactly, I kinda had the same problem with Sacco and Vanzetti when I was uploading Italian authority codes. They have two different codes in the Italian national library system, but have a joint article on Wikipedia. L. ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
Re: [Wikidata-l] Bot request: 250+ thousands person data
On Tue, Apr 29, 2014 at 12:48 PM, Daniel Kinzler daniel.kinz...@wikimedia.de wrote: If you have something like between 1846 and 1855, you can use the before and after fields of the time value: time: +0001850-00-00T00:00:00Z, precision: 9, before: 4, after: 5 This means the main value is 1850, given as a year, with a lower bound four years before and an upper bound 5 years after the main value (before and after are given in the unit specified by the precision value). The main value is what is going to be displayed per default; it will also be used for sorting query results (once we have queries). Is it possible to have just an lower bond, leaving the upper one open? I am thinking of uses like https://www.wikidata.org/wiki/Wikidata:Property_proposal/Generic#earliest_date For things like circa I don't see any clear solution other than inventing some ranges... Micru ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
