Translator broken for new ProQuest interface?

blauar

ProQuest introduced a new interface for its entire platform yesterday (20 Aug. 2015), and now the only options to save to Zotero that appear in Firefox (40.0) for ProQuest databases are "save to Zotero using embedded metadata" or "save to Zotero as web page." All of these just get a web snapshot, not actual article metadata, and PDF download also doesn't work.

It looks like something similar happened the last time ProQuest introduced a new interface (https://forums.zotero.org/discussion/16034/proquest-translator-not-recognising-their-new-interface/).

adamsmith

Yes, I knew about their interface update and was waiting to hear back on that, figured it'd break.
@zuphilip @aurimas -- either of you have time&access? I don't have PQ at the moment.

aurimas

I have access, but not much time atm. The PME probably has (or should have) the most interest in this. I hope they can pick this up.

zuphilip

Just tested a little: The translator should still work if you switch to the "Abstract/Details" tab.

blauar

Thanks for the "Abstract/Details" tab tip -- I'm getting inconsistent results there. Sometimes I just get an error message, sometimes Zotero downloads *some* bibliographic info (but not author name, and no PDF although it shows the blue dot, and double-clicking on the blue dot gives me an error on a ProQuest page).

adamsmith

(we're working on it, should have something soon).

keithstanger

Thank you adamsmith!

Looks like Zotero is working with the new ProQuest interface. I've been trying out searches on peer reviewed articles in PsycINFO. Articles with attached pdfs do successfully download the pdf. I think I recall that Zotero saved a snapshot of the page when I saved items from PsycINFO. It does NOT seem to do that now, even though in preferences I have checked "Automatically take snapshots..." Do I recall incorrectly?

aurimas

That has changed somewhat. We import most metadata from the ProQuest pages, so saving a snapshot seems unnecessary. For articles that do not have a pdf, but offer full text content on the page, we save the full text as snapshot. If that's not working, or you think that saving snapshots of the catalog entries is a good idea, let us know. I don't have access to PsycInfo through ProQuest, so I can't explore this myself unfortunately.

keithstanger

I have never found a need for the page snapshots from our bibliographic academic databases. I will not miss them. I was just observing what I thought was a change in Zotero's behavior.

keithstanger

Aurimas, you said:

"For articles that do not have a pdf, but offer full text content on the page, we save the full text as snapshot."

I just ran a search in ProQuest and for an article available from Newsweek, with only html full text, Zotero DID include a link to the full text snapshot.

adamsmith

For the record, I had pretty much nothing to do with the fix, this was all zuphilip and aurimas.

aurimas

I just ran a search in ProQuest and for an article available from Newsweek, with only html full text, Zotero DID include a link to the full text snapshot.

So working as intended then, right? (just checking this wasn't a typo).

Also, a new "hidden" feature is that if you check the checkboxes next to ProQuest search results, they will be pre-selected in the Zotero import dialog. We'll probably be adding that to more websites (PubMed does that too now)

keithstanger

Aurimas said: "So working as intended then, right? (just checking this wasn't a typo)."

Right...For a ProQuest record with only html full text Zotero correctly saves a snapshot of the full text!

DWL-SDCA

Also, a new "hidden" feature is that if you check the checkboxes next to ProQuest search results, they will be pre-selected in the Zotero import dialog. We'll probably be adding that to more websites (PubMed does that too now)

Please keep us updated as more database websites are set up with this feature. Is this controlled only through Zotero or is action required of the database managers.

I ask because I would be pleased to have that option for Zotero users who select items resulting from a search of the SafetyLit.org database. (SafetyLit is a free service presented without advertising) All SafetyLit records are Zotero-ready and are recognized by type when a user clicks the Zotero download button or folder icon.

aurimas

Currently, pre-selection has to be coded-in specifically for each translator. I see that SafetyLit takes advantage of unAPI, which could be quite challenging to add support for this feature. We'll see what we can do.

I talked to Dan briefly about this, but I think item selection directly in the page should be something we consider in the future across the board (not as pre-selection, but as replacement for the multiples dialog in general).

Please open up a new thread for any further discussion on these topics.

blauar

Thanks, aurimas, zuphilip, and adamsmith! Your work is very much appreciated.