No PDF Attachments from science.sciencemag.org with EZProxy

KrisJ

Hi guys,

My Setup:

Zotero Standalone
Safari 10.0.3

I wanted to save following paper using the Safari Extension with Zotero Standalone running:

http://science.sciencemag.org/content/sci/347/6221/543.full

When I do this from within my university network, all is well:

1) HighWire 2.0 translator is loaded and it seems to automatically find the URL for the PDF: url: "http://science.sciencemag.org/content/sci/347/6221/543.full.pdf"
2) Everything is saved properly

Now, outside of campus, our university offers EZProxy, which I activate using following bookmarklet:

javascript:%7Bproxy='kuleuven.ezproxy.kuleuven.be';document.location.href=('http://'+proxy+'/login?url='+document.location.href);%7D

The URL for the paper then obviously changes:

http://science.sciencemag.org.kuleuven.ezproxy.kuleuven.be/content/347/6221/543.full

The Highwire 2.0 translator gets loaded but saving the PDF fails. This is because this translator internally calls 'Embedded Metadata.js' which in turn reads a meta key from the page html:

'meta name="citation_pdf_url" content="http://science.sciencemag.org/content/sci/347/6221/543.full.pdf"'

This value is used as is and is not amended with the EZProxy prefix and thus, the PDF will not get downloaded...

https://github.com/zotero/translators/blob/master/Embedded Metadata.js#L478

UPDATE:

It is also broken in Chrome 55.0.2883.95

adomasven

This will be fixed in the upcoming version of Zotero Connectors for both Safari and Chrome. At the moment the connector is out for Firefox as a Beta version, if you want to try it.

KrisJ

@adomasven Thanks for the update!

I usually like to run betas of just about anything but with Zotero, the thing holding me back a bit is that I seem to have gathered that it will update the DB permanently.

Because of this, can I expect other tools such as the Papership app (http://www.papershipapp.com) to stop functioning or are the major web API's unchanged?

That app is a major part of my workflow...

adamsmith

Thanks Adomas for letting use know, that's excellent news.

No changes to the web API at all in connection with the 5.0 update, so I wouldn't expect any problems with PaperShip (or other important web apps like ZotPress)

KrisJ

@adomasven I just gave it a try.

The new Firefox add-on indeed works nicely.

However, when Standalone (also the new beta) is running, automated proxy detection did not fire. When I close standalone, things immediately work as advertised: I get a notification that a proxy redirection was detected and this proxy gets added.

However then, the Add-on throws up a message that standalone is not running.

Given that both these observations are a bit contradictory this is perhaps not the intended behaviour?

EDIT:

The above only seems to happen the the very first time when a proxy was never previously used. I checked by deleting all configured proxies.

When first using the proxy afterwards, it is not detected when standalone is running.

Once a proxy was properly detected, restarting Firefox with Standalone running, the proxy is always detected.

This might be a bit confusing for users.

adomasven

Running Standalone does not have any effect on proxy detection. To detect a proxy for the first time, you need to be taken via EZProxy authorization page, where you sign in with your credentials. The login credentials time-out after a while, which was maybe what you were observing, coinciding with Standalone running or not running.

Once the Connector detects your institutional proxy, it will automatically add other proxied resources by just matching the URL pattern and without needing to go through the authorization page again.