Import from Brage does not parse records properly

ninsbl · September 8, 2020

Hi,

I am writing a kind of manual on how to use Zotero most efficiently in my institute.
We are using an archive called BRAGE, where scientific reports are published. BRAGE exposes Dublin Core metadata and records do import into Zotero. Unfortunately not entirely correct.

Here is an example:
https://brage.nina.no/nina-xmlui/handle/11250/2671391?show=full

In Zotero, the reports is registered as book, and report number and series title are missing. Page number gets registered as container title (under "extra").

I am using Zotero 5.0.89 on Windows 10 and import with "Embedded metadata" from Google Chrome. From MS Edge, results are the same, just the PDF import fails in addition...

Any idea how to fix that? Or is this an issue on the BRAGE side? I see for example that the publisher-place is missing...

ninsbl · September 9, 2020

Would we need to write a custom Translator for BRAGE?

Like:
https://github.com/zotero/translators/blob/master/EPA National Library Catalog.js
or
https://github.com/zotero/translators/blob/master/Library Catalog (PICA).js
or
https://github.com/zotero/translators/blob/master/Embedded Metadata.js

adamsmith · September 9, 2020

A custom translator would get you best results, but if they do want to fix this in the metadata:
the presence of a
citation_technical_report_institution
metatag (that's Highwire, not DC) will lead Zotero to recognize this as a report, as would setting dc.type to "report" instead of "research report".

citation_series_title and citation_technical_report_number will work for series title and report number.

Putting the number of pages in dc.source is probably not a useful thing to do...

ninsbl · September 9, 2020

Hi adamsmith,

And thanks for the reply!
How much effort would such a custom translator be for someone who knows the code base (and probably knows which existing translator to start from), roughly?

Just in case I can convince someone to finance a development effort...

adamsmith · September 9, 2020

given that the structure of the repo is pretty straightforward, with the help of a set of sample entries, this is <3hs effort, so a budget of $500 would be plenty, with a bit of wiggle room.

ninsbl · September 14, 2020

Hi again,

It seems colleagues would first try to fix metadata provided at catalog side.

This: https://hdl.handle.net/11250/2673082 new addition comes as report and is understood by Zotero as Report.

However, Zotero does not pick up series title and number, cause it is coded as "dc.relation.ispartofseries".

Any hints what metadata that should be used for:
- serie-title
- report-number
- publisher-place

I could not find any obvious dc tags for that here: https://www.dublincore.org/specifications/dublin-core/dcmi-terms/.

My colleague told me that a customized translator is probably more relevant for ORIA (https://bibsys-almaprimo.hosted.exlibrisgroup.com/primo-explore/search?vid=BIBSYS) cause that is where people rather search than in Brage and Brage is envisioned to be replaced in the next years.

Still, ORIA depends on good metadata in the associated repos, so, any hints on the issues above would be highly appreciated!

Cheers
Stefan

adamsmith · September 14, 2020

See above on metadata:

citation_series_title and citation_technical_report_number will work for series title and report number.

place of publication is difficult with standard metatags. The only metatag that I'm aware of that does this is eprint.place_of_pub

hmbjerkan · November 11, 2020

The issue with page numbers ending up in Extra-field is not only related to the Brage archives, but also to the Oria discovery service in general. Also, this information is now (wrongly) included in the bibliography when using APA 7th, which I believe is a recent change in behaviour. An example reference to illustrate: https://bibsys-almaprimo.hosted.exlibrisgroup.com/permalink/f/38k55/BIBSYS_ILS71534608300002201

adamsmith · November 11, 2020

(due to the same basic new-ish behavior in Zotero, but entirely separate issue; we'll fix)

adamsmith · November 12, 2020

This particular issue is now fixed in Primo/BibSys there may be others.
Note that the fact that this is included in the APA 7 citations in Bibsys itself is due to a somewhat related (but completely independent of Zotero) ExLibris bug that I'm working convincing ExLibris on with one of the BibSys admins.