Correct fields on DOI import...

hughp3

Greetings, I am noticing with springer publication such as the following that certain metadata seems to be misinterpreted and needs correcting with each import. Does this happen to other users? Is there a general fix? -- Note that I am using the DOI to import the metadata, not the Zotero connector in the browser.

Example:: https://link.springer.com/10.1007/978-3-031-28032-0_7

The extra field gets populated with:

> Series Title: Lecture Notes in Computer Science
> DOI: 10.1007/978-3-031-28032-0_7

The number in the "volume" field should actually be the series number not a volume. It gets put in the wrong slot... this happens with many springer publications.

The "series" field is not populated but it does appear in the the extra field. This seems funky, if the series title can be detected, then why can't it be placed in the series field directly?

poettli

I don't see this. The item is imported as a conference paper.

hughp3

@poettli

If you import from the webpage with springerlink it imports as a conference paper. But with the DOI import it comes in as a book section. I'm specifically talking about the DOI import method.

poettli

OK, you should say it in your first message, because the DOI import is not the recommended method, as you can see.

hughp3

Sorry the DOI qualification was in the post title.

hughp3

I'm not sure where the assumption comes from that the prefered option is to import via springerlink/conference paper. If one has the DOI in the Zotero app then that would be the fastest way to import. Note that the conference paper import method does not import the series volume number at all, but does import the series name to the correct field.

adamsmith

Generally import from the publisher tends to give best results (including, e.g., more reliable abstracts).

For the questions here:
1. Item type is specifically book/chapter in CrossRef -- I don't think we can do much about that
2. The series number (i.e. 13972) is specifically tagged with <volume> in CrossRef -- it's potentially guess-able that this pertains to the series and should thus be series number, but I'm not terribly inclined to touch this -- chances are this will just cause issues in the other direction.
3. Series and Series Title are separate fields in Zotero (check the journal article item type), which IIRC was done to parallel the PubMed metadata schema way back. I think Series Title is generally a bad idea (both fields are mapped to the same CSL collection-title, e.g.), so we can change that to import into Series, which exists across more item types.

hughp3

@adamsmith

Re: 1:
Thanks for pointing out that there is an upstream cross-file data link declaration. (One PDF is declared to be equivalent to another...) Which sources does Zotero rely on? and where is this in the code base?

adamsmith

No, that's a misunderstanding -- I'm just pointing out what metadata CrossRef returns for the above DOI -- are you confusing this with bwiernik's answer in your other thread?

(We get UnixRef XML via DOI content negation, equivalent to
curl -LH "Accept: application/vnd.crossref.unixref+xml" https://doi.org/10.1007/978-3-031-28032-0_7 )

hughp3

Thanks @adamsmith Yes. I was getting this issue mixed up with my other question here.