hbz / lobid

Linking Open Bibliographic Data
https://lobid.org/
Eclipse Public License 2.0
16 stars 4 forks source link

Some GND labels not correctly built from Aleph XML #179

Closed acka47 closed 8 years ago

acka47 commented 9 years ago

E.g. , the label of GND entity http://d-nb.info/gnd/36467-8 isn't correctly transformed to RDF, see http://lobid.org/resource?author=36467-8. lt should read "Rheinland-Pfalz. Statistisches Landesamt" or similar and not only "Rheinland Pfalz".

Example source data (snippet):

<datafield tag="205" ind1="-" ind2="1">
   <subfield code="k">Statistisches Landesamt</subfield>
   <subfield code="h">Rheinland-Pfalz</subfield>
</datafield>

In RDF, subfields h and kshould be concatenated, resulting in:

{
    "@id" : "http://d-nb.info/gnd/36467-8",
    "preferredName" : "Rheinland Pfalz. Statistisches Landesamt",
    "preferredNameForTheCorporateBody" : "Rheinland Pfalz. Statistisches Landesamt"
}
acka47 commented 9 years ago

@jschnasse Let us know if this has to be prioritized.

acka47 commented 9 years ago

Similar to lobid/lodmill#735 but for corporations.

acka47 commented 9 years ago

Another example:

Source:

<datafield ind2="1" ind1="-" tag="200">
            <subfield code="k">Deutschland</subfield>
            <subfield code="b">Bundesministerium für Gesundheit</subfield>
            <subfield code="9">(DE-588)2117859-8</subfield>
</datafield>
<datafield ind2="1" ind1="a" tag="204">
            <subfield code="k">Deutschland</subfield>
            <subfield code="b">Umweltbundesamt</subfield>
            <subfield code="9">(DE-588)2116834-9</subfield>
</datafield>

Current JSON:

{
    "@id" : "http://d-nb.info/gnd/2117859-8",
    "preferredName" : "Deutschland",
    "preferredNameForThePerson" : "Deutschland"
 }, {
    "@id" : "http://d-nb.info/gnd/2116834-9",
    "preferredName" : "Deutschland",
    "preferredNameForThePerson" : "Deutschland"
  }

Expected JSON:

{
    "@id" : "http://d-nb.info/gnd/2117859-8",
    "preferredName" : "Deutschland. Bundesministerium für Gesundheit",
    "preferredNameForThePerson" : "Deutschland. Bundesministerium für Gesundheit"
 }, {
    "@id" : "http://d-nb.info/gnd/2116834-9",
    "preferredName" : "Deutschland. Umweltbundesamt",
    "preferredNameForThePerson" : "Deutschland. Umweltbundesamt"
  }
dr0i commented 9 years ago

Data will be deployed tomorrow for testing. Because I don't work tomorrow: If you find staging good, go on and merge the open pull request. Data will then be ready on monday and I can go to deploy them to production then. (if you feel bold or @fsteeg can help you, you may also want to switch staging index to production tomorrow. Hm. Ok this only takes 2 minutes, I will probably have a look tomorrow myself and will switch myself :) )

acka47 commented 8 years ago

Looks great from my side: +1 @jschnasse What do you think?

jschnasse commented 8 years ago

http://test.lobid.org/resource?author=36467-8 looks still wrong?

dr0i commented 8 years ago

The test instance is automatically switched to the newest index which is build weekly at the weekend, not having the commit merged thus not having the new data in it. I switched the index alias back to index which reflects the commit. So, please review anew.

jschnasse commented 8 years ago

+1

dr0i commented 8 years ago

Deployed to production, closing.