[Zotero 7 Beta] Affiliations interpreted as references

mjthoraval

The new automatic extraction of references is confusing the affiliations superscripts as references:
https://s3.amazonaws.com/zotero.org/images/forums/u265723/ykt1nkzjbz4neyqucvtz.png
https://s3.amazonaws.com/zotero.org/images/forums/u265723/64trmxyzlcnsplkzyv8a.png
Craig, M. T., Wohland, J., Stoop, L. P., Kies, A., Pickering, B., Bloomfield, H. C., Browell, J., De Felice, M., Dent, C. J., Deroubaix, A., Frischmuth, F., Gonzalez, P. L. M., Grochowicz, A., Gruber, K., Härtel, P., Kittel, M., Kotzur, L., Labuhn, I., Lundquist, J. K., … Brayshaw, D. J. (2022). Overcoming the disconnect between energy system and climate modeling. Joule, 6(7), 1405–1417. https://doi.org/10.1016/j.joule.2022.05.010

Another example where it extracts all the self-citations of the co-authors:
https://s3.amazonaws.com/zotero.org/images/forums/u265723/81rj5y7ref0hspkr90he.png
Comas-Bru, L., Rehfeld, K., Roesch, C., Amirnezhad-Mozhdehi, S., Harrison, S. P., Atsawawaranunt, K., Ahmad, S. M., Brahim, Y. A., Baker, A., Bosomworth, M., Breitenbach, S. F. M., Burstyn, Y., Columbu, A., Deininger, M., Demény, A., Dixon, B., Fohlmeister, J., Hatvani, I. G., Hu, J., … SISAL Working Group members. (2020). SISALv2: A comprehensive speleothem isotope database with multiple age–depth models. Earth System Science Data, 12(4), 2579–2606. https://doi.org/10.5194/essd-12-2579-2020

Zotero 7.0.0-beta.82+79fa160e4 (64-bit)
Windows 10

martynas_b

The new automatic extraction of references is confusing the affiliations superscripts as references:

We'll improve on that.

Another example where it extracts all the self-citations of the co-authors:

That is expected behavior. There is no way we can avoid that. You have to decide by your self whether the author match makes sense or not.

mjthoraval

Thanks a lot for all the progress continuously coming to Zotero. I report the problems, but the references extraction is really impressive! Even the hyperlinks are active in the popup window, which is nice to go directly to that paper.

Here is another open access paper illustrating some of the problems:
https://doi.org/10.1063/5.0086745

For the papers with numerical style references, it also confuses the content of the equations, and also the references to equation numbers:
https://s3.amazonaws.com/zotero.org/images/forums/u265723/yo54o2wchssmxxza68co.png

The superscript references are correctly identified ("the full drop.^{108}"), but not the in text references ("Ref. 108"):
https://s3.amazonaws.com/zotero.org/images/forums/u265723/vpaxfrx43g681xlvebfi.png

For this specific format of references, the numbers are also in superscript in the references, without any space between the numbers and the references. It would still be nice to add a space when displayed in the popup window:
https://s3.amazonaws.com/zotero.org/images/forums/u265723/fc9b92r1fck5b0796m7r.png
https://s3.amazonaws.com/zotero.org/images/forums/u265723/vijr3t3pp1w582f0m3gz.png

I am curious which papers can support the feature and which ones cannot. For example, recent PRL papers do not seem to get the popup window for references, although they seem to be nicely formatted:
https://doi.org/10.1103/PhysRevLett.131.194002

For the author-year reference styles, it is indeed a nice feature to display all the self-citations.

However, for references in the text, it seems to show all the references that include that author instead of narrowing down the precise reference that is cited. That is fine most of the time, but in some cases the list can become quite long, and I need to scroll down to see the correct reference:
https://doi.org/10.1007/s00348-023-03759-9
https://s3.amazonaws.com/zotero.org/images/forums/u265723/gdjuybmi1w2es6hu53gv.png

warwickmm

I'm also seeing the false positives issue in papers with numerical style references.

Version 7.0.0-beta.83+066eda731 (64-bit) on Linux.

drhere

Same here!

warwickmm

I have a PDF where the equation labels are interpreted as citations, but the actual citations are not.

In this screenshot, the equation label is treated as a citation:

https://s3.amazonaws.com/zotero.org/images/forums/u7554954/iq6tiq9342iy4dg6qcos.png

In this screenshot, there are no preview popups available for the actual citations:

https://s3.amazonaws.com/zotero.org/images/forums/u7554954/msy9fgzxk2llfmx4p02q.png

This is using version 7.0.0-beta.85+c0c00a00e (64-bit) on Linux.

mjthoraval

A small update on the examples above:

https://doi.org/10.1016/j.joule.2022.05.010
→ All solved

https://doi.org/10.5194/essd-12-2579-2020
→ The reference extraction on the list of authors breaks the existing hyperlinks to the affiliation.
Before:
https://s3.amazonaws.com/zotero.org/images/forums/u265723/vd7bj9j5sdkk0xffaox0.png
After:
https://s3.amazonaws.com/zotero.org/images/forums/u265723/h18itap6t9jufelfdtsg.png

https://doi.org/10.1063/5.0086745
→ The wrong references identifications has been improved, with still a few remaining:
https://s3.amazonaws.com/zotero.org/images/forums/u265723/aq8x7a7hatjgi5vapigh.png

https://s3.amazonaws.com/zotero.org/images/forums/u265723/kqo8n3jab9m2uz990grm.png

But the references detection as been completely broken in the whole paper:

https://s3.amazonaws.com/zotero.org/images/forums/u265723/w2f996x45248oa3bfb1i.png

Same for all the AIP journals papers with numerical styles.

https://doi.org/10.1103/PhysRevLett.131.194002
→ References detection is working great in PRL articles, even for old PDF files which do not have hyperlinks:

https://s3.amazonaws.com/zotero.org/images/forums/u265723/enlzxv128j6hkwuxf48d.png

Zotero 7.0.0-beta.87+f59a4da7f (64-bit)
Windows 10

martynas_b

@mjthoraval

I am curious which papers can support the feature and which ones cannot. For example, recent PRL papers do not seem to get the popup window for references, although they seem to be nicely formatted

This specific style, where the number is superscript, is not currently supported, but it will be in the future. Generally, typical numbered lists are usually parsed.