Bibliontology Journal Article import does not properly parse journal title

jimkreft

I am generating Bibliontology via unAPI for journal articles.

Example page here:

http://pubs-test.er.usgs.gov/publication/70141759

Example unapi call here:

http://pubs-test.er.usgs.gov/unapi?id=http://pubs.er.usgs.gov/publication/70141759&format=rdf_bibliontology

The Bibliontology RDF document that the site generates matches what Zotero's Biblioontology output generates.

The issue that I am running into is that the Journal title is not mapped correctly, and on import, the paper's title is imported instead of the Journal's title.

This can be replicated in Scaffold as well- here is the relevant snippet:

...
"title": "Chilled frogs are hot: hibernation and reproduction of the Endangered mountain yellow-legged frog <i>Rana muscosa</i>"
...
"publicationTitle": "Chilled frogs are hot: hibernation and reproduction of the Endangered mountain yellow-legged frog <i>Rana muscosa</i>"
- "code": "Chilled frogs are hot: hibernation and reproduction of the Endangered mountain yellow-legged frog <i>Rana muscosa</i>"
- "reporter": "Chilled frogs are hot: hibernation and reproduction of the Endangered mountain yellow-legged frog <i>Rana muscosa</i>"
- "publisher": "Inter-Research"
}
17:12:17 Translation successful

It looks like the issn mapping is also not quite working.

aurimas

The ISSN isn't in the data, but journal title looks right. Will take a look.

jimkreft

Just curious if you have an idea for solving this problem.


It is definitely something up with lines 212 and 213 of the translator

"publicationTitle": [CONTAINER, n.dcterms+"title"],
"ISSN": [CONTAINER, n.bibo+"issn"],

CONTAINER is not the right reference for the journal title (or issn) which is nested pretty deep within the RDF structure. If I switch container to SUBCONTAINER I can get it to parse the <dcterms:isPartOf> section, but I'm not quite grokking how to construct a bracketed statement that gets me to the right level.

<bibo:AcademicArticle>
<dcterms:isPartOf>
<bibo:Issue>
<bibo:Journal>
<dcterms:title>Endangered Species Research</dcterms:title>
<bibo:issn></bibo:issn>
</bibo:Journal>
</bibo:Issue>
</dcterms:isPartOf>
</bibo:AcademicArticle>

If I switch CONTAINER to SUBCONTAINER I can get it to parse the <dcterms:isPartOf> section, but I'm not quite grokking how to construct a bracketed statement that gets me to the right level of the rdf.

aurimas

That's an export only translator. RDF import is handled separately. Are you interested in import or export?

jimkreft

I am interested in import (or rather, facilitating the best possible import to Zotero for others, as I am a dev for http://pubs.er.usgs.gov). Bibliontology RDF was mentioned in other forum posts as the best way to send data to all the Zotero fields, and it is working great, for example, for reports.

jimkreft

Based on your comments in this zotero-dev thread(from two years ago, but it doesn't look like things have changed),

https://groups.google.com/forum/#!searchin/zotero-dev/bibliontology/zotero-dev/dYm5FXF5MfU/Dlw8Y80TD3UJ

The Bibliontology import translator does not use the RDF translator. If I should be changing the data on export, that's fine too- I'm doing this just for Zotero so I can set the extra field, etc- but it does seem like the Bibliontology import and export data models should match.

aurimas

Ah, you're right. And based on those comments, I remember how I disliked debugging the Bibliontology RDF translator. I don't have time to look at it right now unfortunately.

zuphilip

(Maybe unrelated, but any reason your RDF/XML is not starting with the xml declaration? See also the RDF Validator.)

jimkreft

@zuphillip Ah, no, I just didn't put it in for no good reason- oXygen wasn't yelling at me, so I didn't notice it. The output matches what Zotero generates as output, and I was using that as a guide, even if it wasn't perfect RDF.

It looks like, then, using Bibliontology RDF is not a valid option if I want to control the data fully in Zotero, since things are pretty wonky, and the previous issues with books, etc did not appear to be fully resolved either. I will look into using MODs (which seems better tested and a little easier to make sense of as someone who is not a javascript pro) before I go down the road of an actual site-specific translator.

It also seems that another option could be to just match the Zotero json data model exactly using unapi and just "translate" that.

aurimas

It also seems that another option could be to just match the Zotero json data model exactly using unapi and just "translate" that.

If you want to make sure that everything goes in the right place, that's the way to go. I don't think we have a translator for that atm, but it would be in our best interest to make it work, so it shouldn't be a problem.

adamsmith

theoretically, of course, the appeal of using standards rather than Zotero JSON (for lack of a better name) is that other tools can read them, too, though I'm not aware of anyone else reading unAPI in the first place. (this is not to say that we shouldn't make that easily workable)

I'll try to take a look at Bibliontology soon-ish. aurimas is right that it's a pest, but it's also good practice. Remind me if you haven't heard back in 10 days.

jimkreft

@adamsmith, thanks for looking into getting Bibliontology working better, and let me know if there is anything I can do on my end to make things easier. I am a big fan of standards too, though a standard that is not really adopted anywhere is somewhat questionably a standard. In the land of standards, I might also look into building a schema.org JSON-LD parser for embedded metadata, since that seems like it might be an approach that other parts of the internet are using, and the data model for bibliodata is actually getting pretty mature.

jimkreft

@adamsmith it's been 12 days. Any update on Bibliontology import?

jimkreft

@adamsmith it's been almost a month. No pressure, but could you please let me know if I should not display the bibliontology RDF for journal articles, or if a fix for this translator may be on the way and I should hold off.

Thanks,

Jim

zuphilip

<bibo:Issue>
<bibo:volume>27</bibo:volume>
<bibo:issue>1</bibo:issue>
<dcterms:date>2015</dcterms:date>
<bibo:Journal>
<dcterms:title>Endangered Species Research</dcterms:title>

Shouldn't you have here another isPartOf relation, i.e. bibo:Issue dcterms:isPartOf bibo:Journal?

jimkreft

@zuphillip

It appears that you are, in fact, correct. I could have sworn that I matched the Zotero Output exactly, but it appears that I did not. I will get the fix in on my end today and see if that fixes things- it almost certainly will, based on some testing with Scaffold. I will get back to you all shortly, and thanks for the critical eye.

jimkreft

Success! Thanks to all, sometimes a second pair of eyes is critical. Now to generate bibliontology rdf for the rest of the publication types, and enable it for search.

zuphilip

Cool. It is good to see that this works out.

adamsmith

sorry for being awol on this and thanks for zuphilip for taking a closer look.