Adding Zotero metadata to website

beely

Hi,

I've been asked to add Zotero metadata to my website CMS, and I'm finding the documentation not that helpful to enable this.

From: https://www.zotero.org/support/dev/exposing_metadata I can see I can use meta tags, which is great and easy to implement. However, there is no listing I can find of *what* tag names Zotero expects. From Googling around and looking at a few sources, I can see some example Highwire and Dublin Core tags which do work and import ok, but these seem to only cover a subset of the metadata in Zotero.

How, for example, do I set the "Item Type" from the default "Website" that the web connector adds the webpage as, to "Magazine Article"? I've tried variations on names but nothing I have tried maps to this.

Is there just a 1:1 list of meta tags the web connector parses and maps to the Zotero tags, so I can just read through the ones I need to implement and easily add them in to my site?

Many thanks,

beely

I'm still crawling around the docs to find out how to influence the behaviour to get my content recognised. So:

> Some webpages don't provide any information that Zotero can
> recognize. On these pages, the save button will show a gray
> webpage icon. If you click the save button on these pages,
> Zotero will import the page as a “Web Page” item with a title,
> URL, and access date.

So, which meta data (ideally using Highwire or Dublin Core meta tags) do I need to include for Zotero to display the save icon as a "Magazine Article" so users can save an article from my site with the correct type and metadata?

Or if that's not possible using the above metadata systems (it's not clear to me yet), is it possible if I implement a different system of tagging my articles, using the generic translators?

Or do I have to be a "supported site" for the web connector to show anything other than a grey web page icon?

Edit: I've been playing around with Highwire tags and I can now get citations to show up as "Journal Article", which is some progress, but it seems Zotero is deciding this based on the available metadata in the page. I'd still like to programatically decide to set this to "Magazine Article", so if anyone can help setting this directly, or letting me know which metadata is affecting Zotero's decision to choose a type of "Magazine Article".

Thanks,

beely

Well, it seems this is either the wrong place to make such an enquiry about adding Zotero support for my archive (in which case, I'd be happy to be directed to the proper place - the development list doesn't seem appropriate), or maybe just no-one here knows...

I'm updating this thread with my discoveries so far just in case anyone else in a similar position stumbles across this thread in the future, to hopefully save them from some of my frustration.

I've not been able to find any way so far to use Google/Highwire metatags to change the type to "Magazine Article" when using the web connector to save a web page.

I then investigated the other methods of providing metadata, to see if any of those could do the job. I wasn't too keen on setting up the unAPI method, as it was a bunch of work and didn't seem to be able to provide much more than I can already do with simple metatags.

I did try the COinS method, despite the site/documentation being gone (I managed to access the docs via the Internet Archive Wayback Machine).

I set Zotero's export format to COinS, and copied a well-populated example to see the format (which is fairly ugly). This *does* include a field for type, and using this method I can indeed set the type to "Magazine Article" for my web articles. However, using this method, there are other fields that don't come across, meaning once again I can't get all the metadata I need across.

I tried combining both methods on the page, and the Zotero web connector sees both sets of data - the meta tags, and the COinS data, but it's one or the other when saving the page, they don't seem to be able to be combined.

(Setting the export format to be JSON, all the Zotero fields are there, and I wonder why they just don't implement a JSON parser in the web connector, so you can just add metadata for any available Zotero field, without all these tedious guessing games about which fields may or may not be supported in any method you choose to implement...)

So for now, I've stuck to the meta tags method (Google/Highwire tags), as having all the fields I need to get across, but leaving the type inferred as "Journal Article", and the high readability and low development effort is preferable than having the web connector (using COinS) setting the type correctly to "Magazine Article", but then not being able to get other necessary fields I want to transfer across.

I still welcome any input, and I wonder why adding Zotero support for my archive is such a blackhole of documention, information, and support.

I will add any more info should I discover any,

Thanks for reading.

adamsmith

Sorry about that. Generally zotero-dev is the better place for dev questions:
https://groups.google.com/forum/#!topic/zotero-dev/

You can infer the embedded metadata tags from the import code:
https://github.com/zotero/translators/blob/master/Embedded Metadata.js
Creating a magazine item type is indeed not possible (there's just no vocabulary for that)

JSON-LD support using schema.org metadata is planned, though no ETA. See https://github.com/zotero/translators/issues/917 and various tickets linked therein.

beely

Thanks Adam, and thanks for the code pointer. I did have a look around the source code looking for clues but there's a lot of it and I didn't find that, so that's useful. Especially as I can now see how the type is inferred from the incoming tags.

I'll keep an eye on JSON support, and if the situation changes in the future I'll look to implementing that.

Many thanks!