DOI is parsed from PDF with preceeding `.org/`

[Turun]

JabRef version

5.15 (latest release)

Operating system

Windows

Details on version and operating system

No response

Checked with the latest development build (copy version output from About dialog)

• I made a backup of my libraries before testing the latest development version.
• I have tested the latest development version and the problem persists

Steps to reproduce the behaviour

Thank you for this awesome program and it's continued maintenance and feature expansion. It's hard to find good library management software and I really like JabRef, I'm so glad it exists.

When creating an entry by passing in a PDF (either by dropping it onto the open application or via Menu>Lookup>Seach for unlinked local files (Shift+F7) sometimes the doi is parsed incorrectly, resulting in .org/AABB1122 instead of AABB1122. An example of this is 10.1063/1.2168253.

In another instance the doi was left with a trailing dot (.org/AABB1122.), though I can't find the exact paper/pdf for which this happened.

When testing with the .exe from https://builds.jabref.org/main/windows-amd64/ [1] the issue is no longer present, because

• when loading the mentioned PDF alone, the doi is not parsed at all. The window to merge the data from the metadata of the pdf with the data extracted from the content of the pdf blinks up briefly before disappearing again.
• when I drag and drop multiple PDFs at once, they all get recognized correctly.

(and the PDFs get moved next to the bib file. I get why you would want to copy them there, but moving is way too harsh, especially since they get renamed as well. Now I need to move them all back manually. Also, clicking "undo" in JabRef after that removes the latest addition to the library, but does not undo moving the file, nor can I undo more than one addition)

[1]: This is from the about page
JabRef 6.0-alpha.346--2026-02-15--72f8ced
Windows 11 10.0 amd64
Java 25.0.2
JavaFX 25.0.2+4

Appendix

Running JabRef.exe from the terminal (Windows Terminal, Shell: Powershell) does not produce any output.

[Siedlerchr]

Thanks for the kind words and the report. We will look into this.
Regarding the "move" vs. "copy" behavior and the directories, if you want to copy the files, you need to press ctrl while dragging. https://docs.jabref.org/collect/findunlinkedfiles (The screenshots are outdated, but the functionality is still there)
JabRef has a couple of file directories you can configure to your needs: https://docs.jabref.org/finding-sorting-and-cleaning-entries/filelinks#directories-for-files
And you can, of course, adjust the renaming.

[koppor]

An example of this is 10.1063/1.2168253.

This is behind a paywall - and the PDF google found does not have the issue. Do you have a PDF with the example availble somewhere in the net? It would be nice open access, but IEEE and SpringerLink should also work.

[Turun]

Regarding the "move" vs. "copy" behavior and the directories, if you want to copy the files, you need to press ctrl while dragging. https://docs.jabref.org/collect/findunlinkedfiles (The screenshots are outdated, but the functionality is still there) JabRef has a couple of file directories you can configure to your needs: https://docs.jabref.org/finding-sorting-and-cleaning-entries/filelinks#directories-for-files And you can, of course, adjust the renaming.

Thanks, 5.15 does not move the files unexpectedly - be that due to my settings or just the general program behavior. I'll be sure to test after upgrading.
I can definitely see the use case of moving the files and giving them a coherent name. Download "204506_1_5.0271368.pdf" (publishers, sigh) into the downloads folder, drag it onto Jabref and boom, it's moved to the configured folder and named by author/title/etc. But it was unexpected as a default behavior.

Edit: After working some more with JabRef I noticed that 5.15 also moves the files automatically, but I now have it setup in a way that makes sense, so while initially very unexpected I think the feature itself is a useful one.

An example of this is 10.1063/1.2168253.

This is behind a paywall - and the PDF google found does not have the issue. Do you have a PDF with the example availble somewhere in the net? It would be nice open access, but IEEE and SpringerLink should also work.

The following dois get parsed with a suffix .: 10.1109/LED.2024.3424396, 10.1109/TED.2024.3367314, 10.1109/JMEMS.2022.3172766. None of them are Open Access as far as I can tell. Sorry about that. The first one of those contains text like: "Color versions of one or more figures in this letter are available at https://doi.org/10.1109/LED.2024.3424396. Digital Object Identifier 10.1109/LED.2024.3424396", which might confuse the parsing algorithm to focus on the first mention of "doi" that it can find.

The following Papers are open Access or freely available and get a doi like .org/AABB1122 parsed from the PDF: 10.1063/5.0086978 and 10.1063/5.0266356.

[Mahakbajpai]

I would like to work on improving the DOI parsing logic to correctly handle prefixes like https://doi.org/
and trailing punctuation. I will investigate the current extraction implementation in the external-files component and propose a cleanup or regex-based fix.

[jabref-machine]

👋 Hey @Mahakbajpai, looks like you’re eager to work on this issue — great! 🎉 It also looks like you don't know how to assign issues to yourself. Please read our CONTRIBUTING.md to find out how. It will also guide you in other aspects of contributing to JabRef.

[Mahakbajpai]

/assign-me

[jabref-machine]

👋 Hey @Mahakbajpai, thank you for your interest in this issue! 🎉

We're excited to have you on board. Start by exploring our Contributing guidelines, and set up your local development workspace by following the steps in our workspace setup guidelines to get started smoothly.

For questions on JabRef functionality, you can consult the JabRef Guru. For questions related to the codebase, please start using DeepWiki.

In case you encounter failing tests during development, please check our developer FAQs!

Still facing issues or having more questions? Feel free to ask here on GitHub or on JabRef's Gitter chat. Please don't hesitate to open a (draft) pull request early on to show the direction you are heading towards if unsure.

Happy coding! 🚀