Citation recognition issues..? #464
-
|
Hello, and Thank you for the amazing work for building this add-on I've been using this pulgin just yesterday and I am a big fan of the citation popup that was inspired by Google Scholar pdf reader but I found out sometime pdf++ reads the citation in a bit weired way here is some screen shots issue paper : https://arxiv.org/pdf/2504.20468 // citation number 23 issue paper : https://arxiv.org/pdf/2301.12597 // citation Radford et al., 2021 there is more.. I think the way that it finds citation breaks on some circumstance. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
|
Yeah, the current implementation is pretty much in a proof-of-concept phase. It does not use any fancy machine learning algorithms, it's just made of very naive hand-crafted rules... Line 267 in 27667b6 So it's not surprising to see it breaks on some papers. However the example PDFs you provided will be very useful for future improvements of the extraction algorithm. Thank you! |
Beta Was this translation helpful? Give feedback.
Yeah, the current implementation is pretty much in a proof-of-concept phase. It does not use any fancy machine learning algorithms, it's just made of very naive hand-crafted rules...
obsidian-pdf-plus/src/bib.ts
Line 267 in 27667b6
So it's not surprising to see it breaks on some papers. However the example PDFs you provided will be very useful for future improvements of the extraction algorithm. Thank you!
Also I'll welcome your contribution if you're interested:)