Zotero Overides Google Scholar OpenURL Link Resolver Preferences

[tjowens]

When you ingest information from Google scholar Zotero changes the preferences in Google scholar. Zotero sets Google Scholar's OpenURL settings to OhioLink, OCLC and GMU. This bumps out whatever local resolver the user had specified in Google Scholar was using.

[stakats]

I suspect this problem arises from our GS scholar translator setting the GS cookie to include the Endnote (RIS) citation links. We're probably also overwriting the OpenURL parameter.

[cartesian]

Edits to translator by cartesian (SR)

[cartesian]

I've edited the translator to solve a related problem with EZProxy (couldn't set preferences with cookie). This version sets preferences to show EndNote links via URL, instead of cookie.

Bonus random bugs for this translator:
1) Do a google scholar search for "url.replace". This page crashes both the rewritten and existing translators. Don't know why.
2) Also there seem to be timing issues: scaffold often shows "translation successful" a while before it's done. This bug is also carried over from existing translator.

[cartesian]

This code replaces my earlier fix and fixes a bug that arises when Google interface language wasn't set to English

[cartesian]

Bonus random bug 3) I get a busy hourglass on my Firefox when I mouse over various parts of the browser at any time after this translator runs the Zotero.Utilities.loadDocument() code, even after items are successfully added to my library. The hourglass goes away if I hit refresh for the tab that was scraped. I didn't alter that code except to change the URL it uses; this bug is also a holdover from the current translator.

[dstillman]

I think that may be related to the hidden browser, either in translate code or my attachments code. (Is the translator saving a PDF? Another type of attachment? Does it happen if the PDF-saving pref is disabled?)

[cartesian]

Fixed typo in "itemGrabLink"; this fixes my random bonus bug #1

[cartesian]

Replying to dstillman:

I think that may be related to the hidden browser, either in translate code or my attachments code. (Is the translator saving a PDF? Another type of attachment? Does it happen if the PDF-saving pref is disabled?)

I don't think it's related to attachments - it's the same even if the pdf save pref is off in Options. I think it might be because the translator doesn't wait for the call to loadDocument() to finish properly. But I don't understand enough about the Zotero API to fix it myself.

[dstillman]

Replying to cartesian:

Replying to dstillman:

I think that may be related to the hidden browser, either in translate code or my attachments code. (Is the translator saving a PDF? Another type of attachment? Does it happen if the PDF-saving pref is disabled?)

I don't think it's related to attachments - it's the same even if the pdf save pref is off in Options. I think it might be because the translator doesn't wait for the call to loadDocument() to finish properly. But I don't understand enough about the Zotero API to fix it myself.

A link item is still an attachment (and Zotero still loads the page in a hidden browser to index the page content). If I just return at the beginning of Zotero.Attachments.importFromURL(), it doesn't happen.

Created #750 for this problem. I'll see what I can do.

[cartesian]

Places Zotero.wait()s and done()s better. Also handles additional language prefs in the url. Does NOT fix #750. This will be my last contribution for a while!

[stakats]

(In [1741]) Closes #622 by updating Google Scholar translator with cartesian's code.

[dstillman]

Pushed to repo.