hbz / lobid-gnd

UI and API to the Integrated Authority File (Gemeinsame Normdatei, GND)
http://lobid.org/gnd
Eclipse Public License 2.0
25 stars 5 forks source link

"Reconciliation" of GND identifiers #247

Closed wetneb closed 4 years ago

wetneb commented 4 years ago

Just making sure you had seen this, from OpenRefine's mailing list: https://groups.google.com/d/msgid/openrefine/0b677083-2be3-437c-8f8c-f61b7234330b%40googlegroups.com

A reconciliation query containing a GND identifier as text (such as 132390299) does not give any results, whereas a suggest query for the same text returns the expected result.

acka47 commented 4 years ago

Thanks for the hint. No one from the lobid team was a subscriber until now. I will subscribe to the list today.

Regarding the conrete problem, we should probably add gndIdentifier to the list of fields a reconciliation query is matched against:

https://github.com/hbz/lobid-gnd/blob/cace3e55211aba9976fec081414a9738ddeff98b/app/controllers/Reconcile.java#L446-L453

acka47 commented 4 years ago

We should probably also add ~sameAs.collection.id~ sameAs.id where the corresponding identifiers from other sources are located. People might even be more interested in using external identifiers for reconciliation than GND IDs.

fsteeg commented 4 years ago

Deployed to test:

http://test.lobid.org/gnd/reconcile/?queries={"q1":{"query":"118624822"}} http://test.lobid.org/gnd/reconcile/?queries={"q1":{"query":"http://viaf.org/viaf/50566653"}}

fsteeg commented 4 years ago

There's an issue when the IDs contain a dash:

http://test.lobid.org/gnd/reconcile/?queries={"q1":{"query":"4463847-4"}}

fsteeg commented 4 years ago

Redeployed to test:

http://test.lobid.org/gnd/reconcile/?queries={"q1":{"query":"118624822"}} http://test.lobid.org/gnd/reconcile/?queries={"q1":{"query":"http://viaf.org/viaf/50566653"}} http://test.lobid.org/gnd/reconcile/?queries={"q1":{"query":"4463847-4"}} http://test.lobid.org/gnd/reconcile/?queries={"q1":{"query":"https://d-nb.info/gnd/4463847-4"}}

acka47 commented 4 years ago

Works mostly fine but it doesn't work with ORCIDs, e.g. http://orcid.org/0000-0003-0232-7085.

acka47 commented 4 years ago

My bad, I took non-http ORCID URIs from the production system. On staging, there are https-URIs. So everything works fine. +1