Open acka47 opened 8 months ago
These changes are now coming in and start to make problems so that we received an alert mail "Alert GND: found not compacted field(s)":
Am 20.02.24 um 12:41 schrieb sol:
gnd_20221209.mappings.authority.properties.https://d-nb.properties.info/standards/elementset/gnd#originalScriptNameOfThePerson.properties.@language.type gnd_20221209.mappings.authority.properties.https://d-nb.properties.info/standards/elementset/gnd#originalScriptNameOfThePerson.properties.@language.fields.keyword.type gnd_20221209.mappings.authority.properties.https://d-nb.properties.info/standards/elementset/gnd#originalScriptNameOfThePerson.properties.@value.type gnd_20221209.mappings.authority.properties.https://d-nb.properties.info/standards/elementset/gnd#originalScriptNameOfThePerson.properties.@value.fields.keyword.type
These changes are now coming in and start to make problems [...]
This was just a temporary config issue on the server, statements leading to non-compacted fields are now skipped during processing (see https://github.com/hbz/lobid-gnd/commit/b0c88f6), the alert was for the wrong index (which still contains the non-compacted field).
Regearding
Addition of personal relations to nodes without GND ID
This originates from issue GND-57 we opened at DNB' s JIRA in 2018, see https://github.com/hbz/lobid-gnd/issues/124. I am not sure whether this will directly work for lobid so that we have it in the index (and the UI).
Example has GND ID 124537359:
The turtle now contains the information:
<https://d-nb.info/gnd/124537359> gndo:acquaintanceshipOrFriendship <https://d-nb.info/gnd/11968568X>,
<https://d-nb.info/gnd/102831963>, <https://d-nb.info/gnd/119603578>, <https://d-nb.info/gnd/119616572>,
<https://d-nb.info/gnd/118677446>, <https://d-nb.info/gnd/11968599X>, <https://d-nb.info/gnd/119522403>,
_:node1hn2nbfoqx3718620 .
_:node1hn2nbfoqx3718620 gndo:forname "Johannes Adam";
gndo:surname "Lonicer" .
However, this resource was not updated in lobid as the updates are apparently triggered by the modification date in the MARC record which is (as changes have only happened in the transformation to RDF): 2022-03-23
<controlfield tag="005">20220323003024.0</controlfield>
The RDF/turtle just uses the MARC modification date in dct:modified
:
<https://d-nb.info/gnd/124537359/about> dcterms:license <http://creativecommons.org/publicdomain/zero/1.0/>;
dcterms:modified "2022-03-23T00:30:24.000"^^xsd:dateTime;
gndo:descriptionLevel <https://d-nb.info/standards/vocab/gnd/description-level#1> .
Thus, we will have to reindex the whole GND to get in those changes and to check whether this works.
However, with this query I found some records that have recently changed and contain blank nodes in acquaintanceshipOrFriendship
:
describedBy.dateModified:2024-02-19 AND _exists_:acquaintanceshipOrFriendship
.
MARC:
<datafield tag="500" ind1="1" ind2=" ">
<subfield code="a">Melanchthon, Philipp</subfield>
<subfield code="4">beza</subfield>
<subfield code="4">https://d-nb.info/standards/elementset/gnd#acquaintanceshipOrFriendship</subfield>
<subfield code="w">r</subfield>
<subfield code="i">Bekanntschaft</subfield>
<subfield code="e">Bekanntschaft</subfield>
<subfield code="9">v:VD-16 Mitverf.</subfield>
</datafield>
<datafield tag="500" ind1="1" ind2=" ">
<subfield code="a">Frenzel, Salomon</subfield>
<subfield code="4">beza</subfield>
<subfield code="4">https://d-nb.info/standards/elementset/gnd#acquaintanceshipOrFriendship</subfield>
<subfield code="w">r</subfield>
<subfield code="i">Bekanntschaft</subfield>
<subfield code="e">Bekanntschaft</subfield>
<subfield code="9">v:VD-16 Mitverf.</subfield>
</datafield>
<datafield tag="500" ind1="1" ind2=" ">
<subfield code="a">Sigfrid, Johannes</subfield>
<subfield code="4">beza</subfield>
<subfield code="4">https://d-nb.info/standards/elementset/gnd#acquaintanceshipOrFriendship</subfield>
<subfield code="w">r</subfield>
<subfield code="i">Bekanntschaft</subfield>
<subfield code="e">Bekanntschaft</subfield>
<subfield code="9">v:VD-16 Mitverf.</subfield>
</datafield>
<datafield tag="500" ind1="1" ind2=" ">
<subfield code="a">Dornkreil, Tobias</subfield>
<subfield code="4">beza</subfield>
<subfield code="4">https://d-nb.info/standards/elementset/gnd#acquaintanceshipOrFriendship</subfield>
<subfield code="w">r</subfield>
<subfield code="i">Bekanntschaft</subfield>
<subfield code="e">Bekanntschaft</subfield>
<subfield code="9">v:VD-16 Mitverf.</subfield>
</datafield>
<https://d-nb.info/gnd/12490582X> gndo:acquaintanceshipOrFriendship <https://d-nb.info/gnd/116617446>,
<https://d-nb.info/gnd/121563650>, <https://d-nb.info/gnd/116974788>, _:node1hn2nbfoqx3728988 .
_:node1hn2nbfoqx3728988 gndo:forname "Philipp";
gndo:surname "Melanchthon" .
<https://d-nb.info/gnd/12490582X> gndo:acquaintanceshipOrFriendship _:node1hn2nbfoqx3728989 .
_:node1hn2nbfoqx3728989 gndo:forname "Salomon";
gndo:surname "Frenzel" .
<https://d-nb.info/gnd/12490582X> gndo:acquaintanceshipOrFriendship _:node1hn2nbfoqx3728990 .
_:node1hn2nbfoqx3728990 gndo:forname "Johannes";
gndo:surname "Sigfrid" .
<https://d-nb.info/gnd/12490582X> gndo:acquaintanceshipOrFriendship <https://d-nb.info/gnd/119763931>,
<https://d-nb.info/gnd/103084681>, _:node1hn2nbfoqx3728991 .
_:node1hn2nbfoqx3728991 gndo:forname "Tobias";
gndo:surname "Dornkreil" .
So, the GND turtle looks fine (at least on first sight) but not the [lobid-gnd JSON]():
"acquaintanceshipOrFriendship": [
{
"surname": [
"Sigfrid"
]
},
{
"surname": [
"Frenzel"
]
},
{
"surname": [
"Melanchthon"
]
},
{
"surname": [
"Dornkreil"
]
}
]
The forename is lost and this is apparently because of a typo in the transformation as it is delivererd as gndo:forname
which does not exist instead of the correct gnd:forename
. I will comment accordingly in the GND JIRA.
Regarding
Für die GND-Entität Person werden Namensformen in Originalschriftund -sprache gekennzeichnet und ausgeliefert
We will have to add https://d-nb.info/standards/elementset/gnd#originalScriptNameOfThePerson
to the JSON-LD context.
However, there remain some questions:
Example turtle:
<https://d-nb.info/gnd/118839810> gndo:originalScriptNameOfThePerson "מרקה"@he .
It seems to me that it makes most sense in this case to use the value object approach (which is the default way the JSON-LD library handles this, anyway, see the alert quoted in https://github.com/hbz/lobid-gnd/issues/362#issuecomment-1954082993):
{
"id":"https://d-nb.info/gnd/118839810",
"originalScriptNameOfThePerson":{
"@value":"מרקה",
"@language":"he"
}
}
Why not use another string internationalization approach in JSON-LD?
originalScriptNameOfThePerson
.originalScriptNameOfThePerson.@value
in the JSON to be shown in an application.Right?
As published by the DNB.
Additional property in GND ONtology:
Additional information in the GND RDF:
Änderungen in den GND-Konversionen
Examples
Addition of personal relations to nodes without GND ID
SKOS Relations to other controlled vocabs
Test data