we are writing a shared document with typst in the webui. As one user has pro, we use the sync with zotero feature. Unfortunately we have a of of sources with no author and date in out zotero library.
When syncing , they all get the citekey noauthor_notitle_nodate- and then an increasing number. This is a problem itself, as the title is not in the citekey and whe can therefor not see, which document is which.
But the biggest problem is, that we have more than 100 items like this and the number only goes to 99, after that it starts again at 1 and therefore we have many duplicate cite keys and typst cant build our document.
Zotero has an export option called “BibLaTeX,” which generates the .bib file and the citation keys for the sync feature. You can overwrite the citation key by adding citation key: ... to the Extra field in Zotero.
I know about the better bibtex standard and for local projects I even use the betterbibtex plugin for Zotero, unfortunately as I state in the original post this is a shared project and we use the „Sync with Zotero“ feature of typst. I don’t see how I can make use of betterbibtex there.
Furthermore we would need to manually edit over 300 sources, which is not feasible.
I don’t mean the better bibtex plugin. Zotero has a build-in export option (“BibLaTeX”) which can also be used with the API (biblatex) and it seems like the „Sync with Zotero“ feature uses that export option to generate the .bib file.
The biblatex export format generates the cite keys (with no option to change the used fields) and supports a maximum export of 100 items at once (that’s why the cite keys restart at one).
So this is more of an limitation or issue on Zotero’s side.
This issue is discussed below, @reknih gives a pretty succinct answer
There are two problems here: That Zotero returns us a bunch of empty references and that the disambiguation number resets after 100 entries.
First of all, we would like to propose the workaround of only syncing a collection in which you have moved the items you need explicitly, this reduces the amount of “tombstone” empty references flying around and limits the amount of disambiguation happening. If that is impossible, you can unlink the BibTeX file from Zotero to edit it yourself.
The root cause for the failure to disambiguate more than 100 items is that for large libraries, we have to request their contents from Zotero with multiple loads, each loading a chunk of 100 items in the library. However, > Zotero only disambiguates within that chunk, not across the whole collection.
To fix the empty entry problem, we will just filter empty entries out of the file.
To fix the disambiguation problem, we will need to find items that still conflict and add disambiguation ourselves.
We recognize that the workarounds above are annoying and are prioritizing these fixes internally.
Thank you both, I understand the problem and your proposed workarounds now. I will discuss it with the team and we will choose one of the workarounds. Coming from overleaf we are happy that at least each person in a project is able to reload the sources.
We opted for the workaround where we create a new shared zotero group. Unfortunately, this only partially fixed the problem.
We know have much less items in this zotero group than before and manually added a author and date to every entry which was missing it. But there are still 80 “noauthor__notitle_notdate” entries in the bibfile, like this:
