Add @type to GND entities

[acka47]

In API 1.0 we had some reason not to include the type of a linked GND resource. With API 2.0 we can do this. Example:

{
   "@context":"http://lobid.org/download/contextTmp.json",
   "@id":"http://lobid.org/resources/HT018612706#!",
   "contributor":[
      {
         "@id":"http://d-nb.info/gnd/10040152-1",
         "@type": "http://d-nb.info/standards/elementset/gnd#CorporateBody"
         "preferredName":"Zentralarchiv des Internationalen Kunsthandels",
         "preferredNameForTheCorporateBody":"Zentralarchiv des Internationalen Kunsthandels"
      },
      {
         "@id": "http://d-nb.info/gnd/1069643262",
         "@type": "http://d-nb.info/standards/elementset/gnd#ConferenceOrEvent"
         "preferredName":"49. Art Cologne, 2015, Köln",
         "preferredNameForTheConferenceOrEvent": "49. Art Cologne, 2015, Köln",

      },
      {
         "@id":"http://d-nb.info/gnd/1050748387",
         "@type": "http://d-nb.info/standards/elementset/gnd#DifferentiatedPerson",
         "preferredName":"Herzog, Günter",
         "preferredNameForThePerson":"Herzog, Günter"
      }
   ],
   "contributorOrder":[
      "http://d-nb.info/gnd/1050748387 | http://d-nb.info/gnd/1069643262 | http://d-nb.info/gnd/10040152-1"
   ],

   "subject":[
      {
         "@id":"http://d-nb.info/gnd/4011889-7",
         "@type": [ "http://d-nb.info/standards/elementset/gnd#Country", "http://d-nb.info/standards/elementset/gnd#TerritorialCorporateBodyOrAdministrativeUnit" ]
         "preferredName":"Deutschland, Bundesrepublik",
         "preferredNameForThePlaceOrGeographicName":"Deutschland"
      },
      {
         "@id":"http://dewey.info/class/700/"
      },
      {
         "@id":"http://d-nb.info/gnd/4013255-9",
         "@type": "http://d-nb.info/standards/elementset/gnd#TerritorialCorporateBodyOrAdministrativeUnit",
         "preferredName":"Düsseldorf",
         "preferredNameForThePlaceOrGeographicName":"Düsseldorf"
      },
      {
         "@id":"http://d-nb.info/gnd/4076259-2",
         "@type": "http://d-nb.info/standards/elementset/gnd#SubjectHeadingSensoStricto",
         "preferredName":"Pop-Art",
         "preferredNameForTheSubjectHeading":"Pop-Art"
      },
      {
         "@id":"http://d-nb.info/gnd/4031483-2",
         "@type": "http://d-nb.info/standards/elementset/gnd#TerritorialCorporateBodyOrAdministrativeUnit",
         "preferredName":"Köln",
         "preferredNameForThePlaceOrGeographicName":"Köln"
      }
   ],
   "subjectOrder":[
      "http://d-nb.info/gnd/4011889-7, http://d-nb.info/gnd/4076259-2, Geschichte 1959-1972, Ausstellung, Köln <2015>",
      "http://d-nb.info/gnd/4013255-9, http://d-nb.info/gnd/4031483-2, http://d-nb.info/gnd/4076259-2, Geschichte 1959-1972, Ausstellung, Köln <2015>"
   ],
   "title":[
      "Wie die Pop Art nach Deutschland kam"
   ],
   "type":[
      {
         "@id":"http://purl.org/ontology/bibo/Proceedings"
      },
      {
         "@id":"http://purl.org/dc/terms/BibliographicResource"
      },
      {
         "@id":"http://purl.org/ontology/bibo/Book"
      }
   ]
}

When we have the type, we could probably leave out the specific preferredNameForThe... properties. (As soon as we add variant names we should do so anyways to not inflate the documents to much.) I will make another issue for this.

[acka47]

Here is a list of all types we will use:

• "http://d-nb.info/standards/elementset/gnd#ConferenceOrEvent":"ConferenceOrEvent"
• "http://d-nb.info/standards/elementset/gnd#Family":"Family",
• "http://d-nb.info/standards/elementset/gnd#CorporateBody":"CorporateBody",
• "http://d-nb.info/standards/elementset/gnd#Person":"Person",
• "http://d-nb.info/standards/elementset/gnd#PlaceOrGeographicName":"PlaceOrGeographicName",
• "http://d-nb.info/standards/elementset/gnd#SubjectHeading":"SubjectHeading",
• "http://d-nb.info/standards/elementset/gnd#Work":"Work"

[acka47]

We shouldn't use the URIs but alias them in the @context. See #26.

[acka47]

@dr0i asked offline how to identify in the MAB source that a GND entry is of type "Family". Let's have a look at some example files with GND families:

HT017472134:

<datafield tag="902" ind1="-" ind2="1">
  <subfield code="p">Dorhoff</subfield>
  <subfield code="c">Familie</subfield>
  <subfield code="9">(DE-588)102851445X</subfield>
</datafield>

BT000038827:

<datafield tag="902" ind1="-" ind2="1">
  <subfield code="p">Heinsberg, von</subfield>
  <subfield code="c">Familie</subfield>
  <subfield code="9">(DE-588)189407115</subfield>
</datafield>

HT016255059:

<datafield tag="902" ind1="-" ind2="1">
  <subfield code="p">Wittgenstein</subfield>
  <subfield code="c">Familie</subfield>
  <subfield code="d">ca. 19./20. Jh.</subfield>
  <subfield code="9">(DE-588)118769618</subfield>
</datafield>

I had a look at several more examples. It looks like the GND references of type "Family" are identified by the string "Familie" in subfield c.

[dr0i]

If an ID is subject AND creator(like for http://lobid.org/resources/HT018771475 and HT014681992, HT014319164 ) the type is only added to one of them. Created the ticket #40.

[dr0i]

@acka47 odd - couldn't find a CorporateBody-subject. You know where to look?

[acka47]

@dr0i for example http://lobid.org/resources/HT014376582 or http://lobid.org/resources/HT019035373 (with hbz as subject).

[dr0i]

Deployed test set, see e.g. http://staging.lobid.org/resources/HT018612706#!.

[acka47]

Something strange happened to the subject array in http://staging.lobid.org/resources/HT008357364#!. There's an empty object:

"subject":[
   {

   },
   {
      "dateOfBirth":"1921",
      "dateOfDeath":"1995",
      "id":"http://d-nb.info/gnd/11855087X",
      "preferredName":"Highsmith, Patricia",
      "preferredNameForThePerson":"Highsmith, Patricia",
      "type":[
         "http://d-nb.info/standards/elementset/gnd#Person"
      ]
   }
]

Seems to be a problem in many resources, e.g. 8 occurences of "subject":[{} in http://gaia.hbz-nrw.de:9200/resources-staging/resource/_search?q=*&size=50.

[acka47]

+1