PubMed Import of Journal Title

meyertet

Hi,

i am a zotero newbe so please apologize my probale stupid question. I recognized that most citation i grab from pubmed pages have the correct journal title with some exceptions. For example citations of the journal "Retina" result in

Retina (Philadelphia, Pa.) (the pubmedxml title)

instead of just

Retina (the pubmedxml MedlineTA).

For most journals pubmedxml title equals pubmedxml MedlineTA. Is there a special filter/patch to use MedlineTA as source for the journal title instead of pubmedxml title ?

One more question - is there an integrated style editor in zotero ? There is the style manager in the preferences section but i do not see a way to edit the styles using the style manager except adding or removing styles.

Thanx in advance for your valuable support

Tobias

zuphilip

Can you give an example of two different pubmedxml formats? Currently, in Zotero we use for example this xml data. This is hardcoded in the Zotero translator. However, "Retina (Philadelphia, Pa.)" looks like an error from pubmed, which they should correct.

There is a simple style editor integrated in Zotero and there is an online visual editor for CSL styles. Some documentation about styles and how to edit/create them.

meyertet

Hello zuphilip,

thank you for your prompt response. As you cited the xml coded title information for the journal retina is - for some reason i do not unterstand - "Retina (Philadelphia, Pa.)".
However there is a second coded title infromation in the section MedlineJournalInfo


<MedlineJournalInfo><Country>United States</Country><MedlineTA>Retina</MedlineTA><NlmUniqueID>8309919</NlmUniqueID><ISSNLinking>0275-004X</ISSNLinking></MedlineJournalInfo>

with the correct title "Retina".

I observed this problem also with the journal "Eye" where "Eye (London, England)" is coded for the <title> information and again the <MedlineTA> information is correct.

Is there a work around that zotero pulls its journal title information from the <MedlineTA> value ?

Could you give me a hint where i can find the simple zotero in build style editor ?

Thanks in advance

Tobias

zuphilip

Okay, I see. Changing the behaviour of zotero for importing pubmed, would need to change the translator code (javascript). However, if this is always an improvement we should update that in the official translator. Is the MedlineJournalInfo data always better?

Try the visual editor for editing styles: http://editor.citationstyles.org/visualEditor/

adamsmith

The built-in editor is described at: https://www.zotero.org/support/dev/citation_styles/reference_test_pane
but note that cleaning up data can't be done in citation styles, so it wouldn't help you with this issue.

We've been wanting to improve the details of journal titles imported by Zotero from PubMed for a while (it also imports the "Journal of the American Association of ABC" subtitles of some journals) , so we'll look at using the MedlineTA field or some other way of removing the geographic information, thanks for bringing it up. (overlapped with zuphilip's response).

meyertet

Thank you for your help - as far as i tested MedlineJournalInfo seems to work for the journal i use.

Best regards

Tobias

adamsmith

@DWL -- do you have any insights on this? Can we expect MedlineTA to be available for all articles?

adamsmith

unfortunately MedlineTA is not an option -- it's the Medline abbreviation, see e.g.
http://www.ncbi.nlm.nih.gov/pubmed/26215967?report=xml&format=text
so just using that won't work.
I'd suggest just chopping off everything after and including " : " and " ("

DWL-SDCA

The NLM uses two mechanisms (both endorsed by international standards bodies) to disambiguate journal titles. One mechanism is to include the city where the journal was first published. The other involves two journals with the same name (generally these have different publication date ranges) that were published in the same city. In the latter case, the name or abbreviation is followed by the year the journal was first published. There are quite a few journals with the same that are currently and simultaneously published. There are many more that have the same name as a journal that is no longer published. (Sometimes this involves a journal that changed to a new name and then changes back again.

Thus, the disambiguation city or year is a true part of the journal name. In the case of the journal named "Retna" I have seen journal articles cited using the disambiguation city in both Vancouver and APA styles.

I don't have time, now, to find the specific examples to post here.

There are many, many journals with titles that are disambiguated this way.

Edit: I agree that it is almost always proper to delete the journal subtitle after the ":" because there are journals that change the subtitle often.

adamsmith

Great. What we should do, though (unless that's already the case, I haven't checked) is to strip the part in parentheses from the auto-abbreviations: The abbreviated form of Retina (Philadelpha, PA) should just be "Retina".

DWL-SDCA

@adamsmith

That is not what I recommended for journals that have a city or a year after the journal name or abbreviation. (i do recommend that subtitles be removed.)

There are several journals with the name Retina:

Retina (Milano)
Retina (Philadelphia, Pa)
The Retina (Toledo, Ohio)
Retina (Cincinnati)
Retina (Monterrey, México)
Retina (Stuttgart)

adamsmith

right, but journals commonly require abbreviations according to Index Medicus/NLM and the NLM Abbreviation for Retina (Philadelphia, PA) is definitely "Retina"
http://www.ncbi.nlm.nih.gov/nlmcatalog/8309919
Is there ever a case where NLM keeps the location for the abbreviated form?

aurimas

Yes, locations are kept in abbreviations. https://forums.zotero.org/discussion/32232/journal-name-is-slightly-incorrect/

adamsmith

that's unfortunate. We do get Retina wrong, though -- it turns into "Retina Phila PA", which is nonsense. Eye (London, England) should turn into Eye (Lond).

So at a minimum, we should keep parentheses when abbreviating. We could consider removing the part after the comma. That still won't get Retina quite right, but I'd say that's as good as we can get without a list match?

aurimas

As far as abbreviations go, I still think that these cases should be handled by ISBN, so the journal title field should be tailored to how it should be cited in unabbreviated form.

adamsmith

agree about the journal title field. I'm thinking about the abbreviation algorithm here.
Question is with what timeframe we could implement ISSN (I assume that's what you mean) based matching.

We actually do seem to have all pertinent data here:
https://github.com/citation-style-language/abbreviations
including a python script that looks like it could be adapted to include ISBNs with relatively little effort.

For journals not on the list, though (and until it's implemented) I think stripping parentheses but leaving their contents is the worst possible option.