Translator for new ADS site

Gabriel_p

The popular Astrophysics Data System (ADS) publications search engine has a new interface:

https://ui.adsabs.harvard.edu/

Could the current Zotero translator for the old ADS (http://adsabs.harvard.edu/abstract_service.html) site be made to work on this one too?

adamsmith

it's a complete rewrite. Won't happen quickly, but we'll put it on the list.

Gabriel_p

Thanks Adam!

aaccomazzi

Is there anything that we (the ADS project) can do to facilitate this? Is an "export to zotero" button an option? What if we populate HTML content via the javascript app?

Sorry, we don't have a local expert so any pointers on how to best embed the necessary information in our application for use with Zotero would be appreciated.

Thanks,
-- Alberto

adamsmith

thanks for reaching out. There's two different parts you could consider:
1. You could implement any of the open standards mentioned here: https://www.zotero.org/support/dev/exposing_metadata#using_an_open_standard_for_exposing_metadata
many of them will help with other reference managers, too

2. Specifically for making a Zotero translator easier -- what we'd do is to try to grab the Endnote export and it looks like that's unnecessarily(?) complicated. So if I'm say on
https://ui.adsabs.harvard.edu/#abs/2016MNRAS.455.3628R/abstract
I'd like to be able to make a Get or Post request and get the Endnote export .txt returned (and we'd then import that into Zotero automatically). Is that possible? Maybe through an API that, ideally, I don't need a key for (because Zotero translators are open, so keys are a bit problematic)?

aaccomazzi

Hi Adam, thanks for the info. I created an issue about this, feel free to add or modify to it: https://github.com/adsabs/bumblebee/issues/752

One question regarding approach 1 (exposing metadata) in record pages: would this work if it's done by the javascript app in the page or does it need to be in the source HTML? If the javascript approach work, we would let js populate the appropriate <META> tags in the page.

adamsmith

Looking at how the site is built now, I think using the js app should work, yes, but I can't make promises (it's a bit hard to predict when the site will register as loaded in the browser, which is when the translator kicks in to check for the meta tags -- if that happens too early, it won't see them).

How about my questions re 2? Is there an easier way for us to get bibtex or endnote export that I'm maybe not seeing?

aaccomazzi

Thanks.

Regarding the second point, requests to the API need to carry an OAuth token. The token is available in the application itself, so if Zotero had a way to grab and send the token to the API the request would work, otherwise it wouldn't. I don't know enough about how this works but can get one of our developers in touch with you if you think that this is even a possibility.

adamsmith

We do that on other sites, so I think that might work, yes. I'll contact you via github if I need more help.

Gabriel_p

Great to see this is moving forward!

Currently I can save an article from the new Bumblebee ADS interface, but not the PDF. Hopefully this will be added down the road.

Thanks to everyone working to make this possible!

Gabriel_p

I see the ADS guys closed the issue on their end (https://github.com/adsabs/bumblebee/issues/752), so I guess the Zotero guys have taken over.

Any new developments on this feature? Cheers.

Gabriel_p

Quick query about this issue.

The translator works for the new ADS site, except PDFs are not automatically stored. Is this feature planned?