Incorrect attached pdf

Hi Paperpile team,

I have been seeing an increasing (and concerning) amount of my entries, old and new, with incorrect PDF attachments. The attached PDF is from my archive, but it is attached to an incorrect entry.

The most serious one just occurred, in which I added a paper from GScholar just to see that the attached paper is an older version of a paper I previously downloaded into Paperpile.

I don’t know how this has happened, but I’m afraid even deleting the incorrect PDF would delete it from my archive.

@neuromechanist could I trouble you for details of some of the affected papers (author, year, title) so we can take a look in backend and maybe try to reproduce as well? By which via did you import the others? Any further details would be helpful to investigate. Thanks in advance.

Yes, there seem to be many. But I can find them one by one when I want to check the PDF, and I usually correct them on the spot. However, here is one that I just found this morning (I am not sure if you can look into my backend database or not):

I added this paper from Google Scholar, with Safari extension, using the add to paperpile add-on button and things worked smoothly (on the surface at least), added PDF easily:

Horien, Corey, Stephanie Noble, Abigail S. Greene, Kangjoo Lee, Daniel S. Barron, Siyuan Gao, David O’Connor, et al. 2021. “A Hitchhiker’s Guide to Working with Large, Open-Source Neuroimaging Datasets.” Nature Human Behaviour 5 (2): 185–93.

Then, looking at the PDF, it is on my account is replaced by the PDF I already have for another paper (Event in Context by Robbins and Makeig 2024)

Thanks for the revert, @neuromechanist — the team is looking into this now. A similar issue we were working on last week was that a wrong PDF was displayed in the import notification pop-up, but then the actual attached file was correct. From your description it sounds like this is not the case for you — just wanted to double check you’ve actually opened the PDFs to ascertain they are the wrong ones. Let me know.

Correct. I opened the PDF and saw that it was not the one that should’ve been attached (it is more than one incident).

Interestingly, the file name shown under the entry is correct, but the link to PDF is not.

More interestingly, when I looked up the file name under my G Drive, I could find the correct PDF already there!

There is something serious going on. I just added Automatic quality control of the numerical accuracy of EEG Lead fields to Paperpile and started reading the PDF (which was automatically added and it was correct).

All of a sudden, the PDF changed to another PDF from another paper I have in my archive! This is making tracking papers very difficult!

@neuromechanist did you try multiple imports at the same time? If possible, please select one or some of the affected refs, press Cmd/Ctrl+J and share the resulting output with us (either here or via messenger/email so the team can investigate further. Thanks for your patience and help troubleshooting this.

Thanks, no, I did not.

However, I found that if I load the PDF under Edge, everything is fine. But when I use Safari, I consistently get a PDF from another paper. It seems that it does not matter if I add the paper from Safari or Edge, the problem is Safari holds. BTW, I already tried clearing the paperpile cache in Safari and setting up everything again, but no luck.

I got two JSONs for Safari and Edge. It seems that Safari JSON lacks the owner field. I have emailed it to the support already and will follow up from that route since it seems to be an isolated issue.


1 Like