Wikidata:Property proposal/EU Publication office ID

EU Publications office authority table ID

Originally proposed at Wikidata:Property proposal/Authority control

   Not done

Description	MISSING
Represents	Publications Office of the European Union (Q480222)
Data type	URL
Domain	SKOS Concepts
Allowed values	URIs
Example 1	Andorra (Q228) → http://publications.europa.eu/resource/authority/country/AND
Example 2	Vienna (Q1741) → http://publications.europa.eu/resource/authority/place/AUT_VIE
Example 3	Euro (Q4916) → http://publications.europa.eu/resource/authority/currency/EUR
Planned use	Mapping countries to their corresponding concept in the 'Country list' controlled vocabulary.
Expected completeness	eventually complete (Q21873974)
See also	https://publications.europa.eu/en/web/eu-vocabularies/controlled-vocabularies

Motivation

The Publications Office of the European Union manages controlled vocabularies that are used by public administrations in Europe as reference data. By mapping these controlled vocabularies to wikidata entities will give the advantage of being able to better integrate data sources.

Discussion

•  Comment @Sander VD: you forgot related items on the examples. Nomen ad hoc (talk) 18:49, 14 July 2019 (UTC).
•  Comment @Sander VD: If you want to use the full europa URL, there's no need for a separate prop, just use exact match (P2888). If you want to have a separate prop, what are the different authorities (vocabularies) and number of entries in each? --Vladimir Alexiev (talk) 13:34, 15 July 2019 (UTC)

  Hi Vladimir,

  I see that you have found the list. There are a lot more now. Starting from lists of countries, currencies, languages, browsers for institutional use, to lists of products and services administered by EuroStat.

  Speaking just about thesauries, taxonomies, glossaries, code lists and authority lists, all of them skos based, the catalogue includes around 250 vocabularies. If we add all other the number goes over 400.

  The catalogue can be (and always been) accessed by a sparql endpoint here https://publications.europa.eu/webapi/rdf/sparql and here https://op.europa.eu/en/advanced-sparql-query-editor

  The number of concepts varies between them. In total there are over 1.2 million. Mpaunescu (talk) 08:51, 15 September 2023 (UTC)

•  Comment @Vladimir Alexiev: The full list of authority tables can be found here, the country list I'm interested in mapping now contains 253 concepts, but the file types and languages taxonomies are also on my wish list. Since the publications office lists are not yet available in a SPARQL endpoint getting a (meaningful) count of concepts for all of them is a bit cumbersome at the moment.
•  Comment If the prefix is always http://publications.europa.eu/resource/authority/, then the data type should be "external identifier" with values "country/AND", "place/AUT_VIE", and "currency/EUR". The URL formatter for a human-friendly page appears to be https://publications.europa.eu/en/web/eu-vocabularies/at-concept/-/resource/$1. Triples containing the linked data URI can the be automatically added into WDQS by specifying formatter URI for RDF resource (P1921). With the above changes I am prepared to support (useful for the translations if nothing else; and far better to have an external-id than to clutter up above the line with exact match (P2888). Otherwise this is an  Oppose. Jheald (talk) 19:43, 15 July 2019 (UTC)

  Thank you for your comment @Jhead.

  Indeed the structure for a concept URI is: http://publications.europa.eu/resource/authority/ {id of the vocabulary} / {id of the concept}.

  The URL for the human-friendly pages is using the following pattern: https://op.europa.eu/en/web/eu-vocabularies/concept/-/resource?uri= {URI of the concept}

  That makes everything simpler and I think we are compliant with the request. Mpaunescu (talk) 09:48, 15 September 2023 (UTC)

•  Comment All examples result in "Page not found. Sorry but the page you are looking for was moved or doesn’t exist anymore." --Jneubert (talk) 13:24, 20 July 2019 (UTC)

  Hi @Jneubert. The URIs are persistent so you should be able to access them (unless there was an accident at that moment). There are systems dependent on these lists so they have to be there all the time. Mpaunescu (talk) 08:53, 15 September 2023 (UTC)

  Can confirm that the URIs resolve. Cheers, Jneubert (talk) 18:05, 15 September 2023 (UTC)

•  Comment @Jheald: Sorry for the delay, I was off for vacation, just got back to work. Thanks for the tip Jheald, that does indeed seem like the proper solution. In this case however, wouldn't it be more appropriate to name the property 'EU Publications office authority table id'? The publications office has other types of identifiers (for referring to legislation for instance), which would be out of scope with the URI prefix.  – The preceding unsigned comment was added by Sander VD (talk • contribs) at 12:10, 6 August 2019‎ (UTC).

  Fully agree that we should avoid confusions with the other IDs. Yet speaking just about 'Authority tables' would limit the scope. There are some other types of vocabularies that could/should enter under the umbrella.

  Maybe something like: 'EU Publications Office vocabularies id' would work better? Mpaunescu (talk) 09:52, 15 September 2023 (UTC)

•  Comment @Jneubert: The examples are not URL, but URIs that are not necessarily dereferenceable. The first example for instance can be found in this controlled vocabulary: https://publications.europa.eu/en/web/eu-vocabularies/at-dataset/-/resource/dataset/country  – The preceding unsigned comment was added by Sander VD (talk • contribs) at 12:10, 6 August 2019‎ (UTC).

  Indeed the URLs take the form : https://op.europa.eu/en/web/eu-vocabularies/concept/-/resource?uri= {URI of the concept} Mpaunescu (talk) 09:53, 15 September 2023 (UTC)