search

psu-libraries / researcher-metadata

Penn State University's faculty and research metadata repository
https://metadata.libraries.psu.edu/
MIT License
7 stars 0 forks source link

OA Button importer is filling in wrong DOIs #962

Open anaelizabethenriquez opened 7 months ago

anaelizabethenriquez commented 7 months ago

I thought we had an issue for this, but I can't find it.

@nmg110 and I have heard several times now from a developer in the College of Ag who is using the RMD API. Publication 538574 is an Extension Publication (Vaccination for Highly Pathogenic Avian Influenza (HPAI)) that does not have a DOI. The DOI field should be blank. Instead, one of the importers keeps filling in an incorrect DOI, for a journal article that has a similar title: "Rational approach to vaccination against highly pathogenic avian influenza in Nigeria: a scientific perspective and global best practice" from Archives of Virology (doi: https://doi.org/10.1007/s00705-023-05888-2)

As @ajkiessl shared in an email back in December, querying OA Button with the correct title leads to this record, which has the title for the PSU Extension publication but the DOI and other metadata for the Archives of Virology article: https://bg.api.oa.works/find?id=Vaccination%20for%20Highly%20Pathogenic%20Avian%20Influenza

I have reported this error to OA Button but not heard back. Is there anything we can do on the Penn State side to stop this happening? I thought about changing the "authorships" in RMD on this publication to "unconfirmed," but the authors do want the article to show up in their data. How difficult/how much of a bad idea would it be to exclude the publication type "Extension Publication" from being updated by OA Button and Unpaywall? I don't think any of them (at least published by PSU) have DOIs (at the moment; could change in the future). Any other ideas?

anaelizabethenriquez commented 7 months ago

I just reported the bug again, this time by sending an email to hello@oa.works so I have more of a record of it (the other times I used the webform).

ajkiessl commented 7 months ago

@anaelizabethenriquez I don't think there's anything we can do that wouldn't involve changing the code. Filtering out Extension Publications from the OAB and Unpaywall searches wouldn't be too difficult, though. I also don't think that would cause any issues, unless (as you mentioned) Extension Publications start getting DOIs.

I'm pretty booked right now development-wise, so I'm not sure I'll be able to get to that change in the short term.

nmg110 commented 7 months ago

@ajkiessl & @anaelizabethenriquez I think this can and should to wait until our next RMD development cycle as this is only one Faculty member out of thousands and she can contact me if she has complaints. I will let Tim know we won't be able to fix this until later this year. I also think it would be OK to exclude Extension Publications from the OAB and Unpaywall.

anaelizabethenriquez commented 7 months ago

Thanks, @nmg110 and @ajkiessl. Let's save this for the next development cycle.

anaelizabethenriquez commented 5 months ago

Tagging this as low priority just to indicate that it is relatively low priority among the issues being considered for the next dev cycle (given the number of users impacted).

nmg110 commented 2 months ago

I added help wanted to this, because it really is becoming a nuisance. I have had to go in every week this month to fix a couple publications for this one user (mul132). I am sure it is not just this user having an issue and is university wide issue. For mul132's publications (Avian metapneumovirus, 2024) they reported the title "Fowl Cholera" to me 3 times this month. Each time I have gone in and removed the DOI and it keeps coming back! Last week I decided I would delete the open access location information from the records and added the url to the publication hoping that would over ride it and it didn't. Is there any way we add just an override button that stops the record from being checked?