Yes, the keywords field is interpreted as a comma-separated list. I don't think there currently is a way around that, other than using a different character, e.g., a semicolon. Alternatively, but somewhat hacky, you could use a different comma from https://www.compart.com/en/unicode/category/Po, e.g., a full-width comma or an arabic comma.

Hi Iago,

I forced a refetch from Zotero by adding `nocache=1` to the URL once (link). The C-reactive paper is gone now:
https://bibbase.org/show?bib=https://bibbase.org/zotero-group/iagogv/4865242

The Zotero API is too slow for us to fetch from it every time the page refreshes, so we need to cache the results. The cache get updated regularly (I think currently once a day or so), so you could probably also have just waited. But if you want to force a refresh, you can always open the above link with nocache once manually.

My apologies. There was a bug in a recent patch we deployed. It's fixed now. This only affected users that do not use a complete URL (with http://) as their "bib" parameter, which is why we didn't notice this right away.

Thanks for reporting this.

Another good idea!

We didn't add the date, but downloaded files are now either named after the input file, or, if none is given, e.g., when using Mendeley as a data source, it will be call "all.bib".

Thanks for the suggestion.

Good idea! This is done now via a new option "titleLink" (see updated documentation).

Hi Paolo,

Right now only user libraries ("My Library") and group libraries are fully supported. The issue with the collections is that Zotero will only export 100 items at a time and the required logic to make a sequence of requests of 100 until all publications are received is not yet implemented for collections.

If it's an option for you I recommend downloading the complete bibtex file from Zotero, and uploading it directly into your account on BibBase. From there you'll be able to serve the publications list in full and this is always the best supported option.

Christian

OK, it worked. I've deleted all files except the two bib-files.

Hi Adam,

I'll try and manually delete them. I'll report back here.

Christian

According to https://developer.mozilla.org/en-US/docs/Web/HTML/Element/base#attr-target you can add

<base target="_blank" />

to the head of your HTML document to make all links open in a new tab. This would then apply to links created by BibBase as well. Does that work for you?

Hi Adam,

I think we really need to find a way to avoid this workflow. Deleting and then re-uploading over a gigabyte each time there is a change is not good for a number of reasons, not least of which being what you saw, namely that it will slow things down. I believe this is not really a bug, but rather an effect of the pricing structure of our cloud provider (AWS): reads and writes can burst (go really fast) but only for so long. Once that burst budget has been used up things can get very slow, and I think that's what's happening here.

Would it help if we added a de-duplication feature, where, when dropping files that already exist in the account, there is a popup asking whether ti skip or replace them? If I understand correctly, the only reason you are reuploading everything is because it's hard for you to know which files were added and deleted, right? This would not handle the deleted ones, but would give us a viable way to add new ones without needing to reupload those that are unchanged.

Would that work