Chrome, Zotero Standalone, and automatic PDF download

eawtrey

I've searched these forums for a while, but can't quite seem to nail down the answer to my issue...

I've been using FF/Zotero 2.x for quite a while. The combination of citation management and automated PDF download is the killer feature for me. I recently switched my main browser to Chrome for a variety of features, but I find that this nixes my ability to automate PDF downloads to my local computer. Can anyone speak to this scenario?

EDIT: I'm attempting to use Zotero 3.0 standalone and the Chrome connector.

adamsmith

The problem most likely is that you're trying to get to gated pdfs.
In Firefox, Zotero uses the built-in proxy redirection to make sure that the URL to the fulltext PDF is redirected through your institutional proxy, making sure you have access.
The Zoter's Chrome/Safari Connectors can't do that. There are extensions for Chrome that will do automatic proxy detection/redirection for you - you'll have to search for them. With those enabled you should be able to get gated pdfs.

To check whether this is the case, try importing the citation here:
To
http://www.advancesindifferenceequations.com/content/2012/1/2/abstract
You should get a pdf (the journal is open access).

eawtrey

Aha! This is true - this link worked through Chrome. I've seen the references to Chrome proxy detectors in other places. I'll track those down, test, and post my final solution here.

lucagiupponi

Did you ever get this figured out? I'd love to be able to automatically save PDFs!

d.bobak

I have the same problem also.
Latest Zotero Standalone (i.e 3.0)
Chrome 17 (beta)

In Chrome I have installed Proxy SwitchySharp extension, which automatically redirects me through a proxy (with login and password) for some sites , like Science Direct. When I try to download pdf directly from Chrome, it works perfectly. When I get article to Zotero by an icon on the address bar, it works for open access sites. But it does not work for sites where I have full access only by a proxy. The reference goes to Zotero, but it cannot grab a pdf.

I have noticed, that there is a settings in about:config like "network.proxy.http", "network.proxy.http_port", but it seems not working for proxies, where I have to authorize me by a login and password. The syntax similar to: user:password@proxy_address does not work. Why there is no proxy tab in Zotero Standalone settings?!

For Science Direct there is also possibility to login to their site, as a authenticated user I can download pdf's... but of course that also does not work for Zotero Standalone.

So, anyone have any idea, how to work with Zotero Standalone? I don't like Firefox so much and prefer Chrome...

eawtrey

@lucagiupponi,

I did not get it figured out. I haven't found a proxy detector that functions like the one in FF did - specifically, I haven't found something that will attach my proxy information to the PDF download portion of Zotero.

This seems to be the same issue that @d.bobak is having. (They just don't have the automatic proxy switcher in Chrome.)

@adamsmith seems to think there is some Chrome extension out there that can do this, but all I can locate is a EZProxy extension that adds the appropriate domain name to an address in Chrome - doesn't work for the Zotero download portion.

Sadness - I, like @d.bobak, much prefer Chrome, but this is the one thing I must fix!

eawtrey

Update: My workaround, at this point, is to manually download the journal articles (PDFs) that I'm interested in, and then drag-and-drop them into SA Zotero. I can then have Zotero look up the metadata for the PDFs and create proper Zotero links.

Not ideal, but better than nothing. Hopefully someone will figure out how to make the Chrome-SA proxy situation work better in the future.

Thanks to the Zotero team! Despite my issues, this is a great product.

adamsmith

As a better workaround - have a look at the Zotfile plugin. You can install plugins in standalone through Tools--> Add-ons --> Tool Icon (top right) --> install add-on from file.

https://addons.mozilla.org/en-US/firefox/addon/zotfile/versions/
(install 2.0)

eawtrey

Interesting - I already use ZotFile, but how would this help with the downloading of PDFs through SA? Or are you suggesting that for the metadata lookup?

My current workflow:
1) Locate article online
2) Download to local directory
3) Drag-drop to Zotero SA
4) Look up metadata via Zotero
5) Rename & move via ZotFile

adamsmith

the suggested workflow would be:
1) locate article online
2) Download article data using connector and Zotero icon.
3) Download pdf to download folder
4) Attach, move, and rename using ZotFile

That way you a) never leave your browser and b) get publisher data rather than PDF lookup. The former is usually much better.

eawtrey

Ah, yes! It never occurred to me to use the attach function of ZotFile. I was only using it to sync with a DropBox folder for remote usage.

Your workaround saves me a step - thanks!

d.bobak

Finally, I have found a solution, that works!

The keyword is local proxy server, in my case it is privoxy (www.privoxy.org).

1. Download and install privoxy. It may be a good idea to install it as a service in Windows.

2. Next step, is to tell privoxy to forward traffic aiming to some sites through "parent proxy". In main config file (config.txt in Windows) put some lines like that:

forward .sciencedirect.com parent_proxy_addres:port
forward .jstor.org parent_proxy_address:port
and so on with other sites we need...

3. Second step is to "force" privoxy to send our parent proxy login and password. We do it in user.action file, put it in the end of the file:

{+add-header{Proxy-Authorization: Basic BASE64ENCODEDUSERANDPASSWORD==}}
.sciencedirect.com
.jstor.org
.nature.com
.sciencemag.org
.scopus.com
.wiley.com
.and so on with other sites we need.

The string BASE64ENCODEDUSERANDPASSWORD== has to be replaced by string:

username:password

encoded by Base 64 algorithm. There are many sites with online encoders, for example: http://www.opinionatedgeek.com/DotNet/Tools/Base64Encode/default.aspx
Properly encoded string ends with two equal signs (==).

4. Next step, is to tell Zotero Standalone (and possibly our browsers and system) to use privoxy as a proxy server. In Zotero, in about:config we have to change following preferences:
- network.proxy.type should be "1"
- network.proxy.http should be "127.0.0.1"
- network.proxy.http_port should be "8118"
- network.proxy.ssl should be "127.0.0.1"
- network.proxy.ssl_port should be "8118"

The address 127.0.0.1 and port number 8118 is privoxy default, if we changed it in config.txt we should also change it here.

5. We can also change system-wide proxy setting as well as in some browsers.

6. It is good to make a restart of Windows, just in case ;)

After all those steps, I have a fully functional and working Zotero Standalone and Chrome cooperation!

Simon

I have a fix for this (provided you are logged into your library proxy in Chrome) that will probably make it into 3.0.2.

adamsmith

We attach cookies sent from the connector to attachment downloads, which should get most proxies working, provided that the user is logged in.

That's a nice way to solve (or at least mostly solve) this - cool!

eawtrey

Excellent! I'll be eagerly watching for this :)

eawtrey

@Simon I've downloaded version 3.0.2 - what should I be looking for? Behavior hasn't changed much.

adamsmith

You should get pdfs attached e.g. in JSTOR.

eawtrey

Looks like it works now - I just wasn't being patient enough with JSTOR :)

EBSCO is sometimes wonky, but I think that may be on my end.

shel

I just installed 3.03 on my windows machine, using Chrome standalone and connector, and am not getting any of the Zotero icons when using jstor from my university library. No problems in EBSCO, though.

adamsmith

You may have to manually accept the terms of conditions once - did you try that?

Simon

What's the URL?

shel

Hi Adam,

I did accept the terms and conditions. The problem, which I'm having on my linux machine now as well, is that in the address bar, there are no icons for me to click to add a citation to my zotero library. Many of my students who have tried the standalone with Chrome or Safari have had the same problem at our institution. I don't know if all of the problems are related to off-campus access or not. (I usually do most of my work with Zotero off-campus--no time when I'm on-campus.) The problem seems to be just with jstor.

From home, I use a proxy. I sign in with my credentials and then, for example, go to the article here:

http://proxy.ulib.csuohio.edu:2092/stable/4486409

I'm wondering if this is a problem with Chrome and the proxy switching problems discussed above. I have not been brave enough to try any of the solutions discussed by d.bobak. I have tried to stop using Firefox for several reasons, but will go back to it if I have to to get full Zotero functionality.

Thanks for any help!

Thanks for any help.

adamsmith

I'm surprised that works with Firefox even.
But yes - since jstor is not in the URL, the connector won't recognize the site as JSTOR and won't display an icon. In the short run there's nothing we can do about that - I believe there are plans to integrate the proxy feature in the standalone connectors, but I don't think that will happen soon.

Simon

Only Zotero for Firefox can understand URLs that look like that. The Chrome and Safari connectors can only understand URLs like http://www.jstor.org.proxy.university.edu, which are more common these days.

I think that Chrome 17 added the features we need to support proxies like yours, but it will be a little while before we port this functionality from Zotero for Firefox to the Chrome connector.

EDIT: You might also see want if you can get VPN access to your university, which would allow you access JSTOR without a proxy so that the Chrome connector works.

shel

Thanks for the responses. I'll speak to our librarians about it, but since the university has just signed a contract with refworks, I doubt they will care to help much. Sounds like I will either go back to Firefox--assuming it will work, given AdamSmith's response, I'm now worried about even that--or capitulate to our administration and use RefWorks instead of Zotero.

Edit: I have used the beta and release candidates of standalone with Chrome and did not have this problem, which is why I thought it might have something to do with the 3.03. Guess not. Thanks anyway.

Edit 2: Re-installed Zotero for firefox. No problem using it on my university library website. Jstor works again, thank goodness, though I don't like having to rely on Firefox.

Barbara Belland

I too would like to use Zotero with Chrome but given the response above will revert to Firefox until this can be fixed.

adamsmith

@Barbara - it really depends on the specifics - what site & what URL are you looking at?

qenghis

Hi Adam, with JSTOR, pdfs are automatically downloading for me, but with ProQuest, they are not. I haven't tested other sites.

I'm using standalone 3.0.3 with Chrome 18.0. The same process works fine with the FF plugin.

Thanks.

adamsmith

we need to see actual URLs to say more. I believe proquest should typically work with Chrome.

qenghis

This url for example: http://search.proquest.com.ezproxy.cul.columbia.edu/docview/144672462

Thanks.

Simon

qenghis: that ProQuest page won't work through a proxy at the moment, because the PDF is on a different domain from the page being saved. However, a workaround for this is next on my list.