Library Proxies : basics questions

Could somebody explain to me the value and the recommended practices for proxies settings when consulting i) press articles ; ii) academic articles from your Library/University distant access please?

To my view the documentation page does not explain this with clarity (https://www.zotero.org/support/connector_preferences#proxies_preferences). And the forum has specific questions but nothing I could really understand from a quick search.

I can see from a personal test (press article added from my library proxy with Zotero web plug) that letting default parameters adds the information that an article x has been added from my instition vpn (i.e the institution acronym appears in the url field in Zotero).

...But what are the nuts and bolts of this? What are the others options available and what are the corresponding pros and cons? Any impact on the printed bibliography (let say in APA7 or Chicago style) ?

Thank you!
  • I can see from a personal test (press article added from my library proxy with Zotero web plug) that letting default parameters adds the information that an article x has been added from my instition vpn (i.e the institution acronym appears in the url field in Zotero).
    It's not supposed to do that and that'd be a bug. The proxy should never be in your URL field.

    The recommended settings for any resource for which you get access to your university are the default settings. Not sure what "press" articles are, but if you mean Newspapers, you shouldn't see proxy access their unless, again, access is through your university.

    The settings you see either turn off specific features (the notification pop up or automatically adding new sites) or the the entire automatic redirection feature -- there's not much more to say about this?
  • Thank you for that quick answer,
    The press article means newspaper yes indeed, please forgive my french.
    Ok there is no problem regarding academic article (the url remains the same as without the proxy).

    However, a suffix is added in urls from some newspapers items, see for example the two following:

    https://www-mediapart-fr.bnf.idm.oclc.org/journal/international/241022/en-autriche-des-trains-pas-chers-contre-le-rechauffement-climatique
    or : https://www-arretsurimages-net.bnf.idm.oclc.org

    Actually it can be problematic because if I want to share the item later on with someone or if I canceled my subscription (next year, next month, etc.), I won't be able to access the item clicking on it from zotero interface url field (though it is true that I still have a html snaphot).

    See that bnf.idm.oclc ? Any idea on how to make that suffix disappear?
  • I think @adomasven is in charge of that code. For exactly the reason you said, the proxy should never end up in the URL field, so something isn't working as it should.
    To make sure I understand correctly -- do you need the proxy for full access to the papers?
  • edited October 24, 2022
    I suppose I should edit the Schem field in Zotero connector preferences" " %h.example.com/%p " with the following variables and help:

    You may use the following variables in your proxy scheme:
    %h - The hostname of the proxied site (e.g., www.example.com)
    %p - The path of the proxied page excluding the leading slash (e.g., about/index.html)
    %d - The directory path (e.g., about/)
    %f - The filename (e.g., index.html)
    %a - Any string

    But I won't risk anything right now.

    As for the second question, the usecase is a newspapaer review (press review) that implies articles locked by a premium subscription. So yes the proxie enable me to access the full article through my institution (I need it to reach the full access).

    I understood that as an alternative I could re-open the article without the proxy (so without full view), add it with Zotero connector to get the metadatas, then switch back to the full-access version, download it as pdf (if enabled by the newspaper), and then add it as a child item with zotfile. Not optimal but that should works.
  • No, you're really overthinking this. Something isn't working correctly and someone needs to look at this. You shouldn't change anything or do anything differently.
  • edited October 25, 2022
    @alflamingo: Did you configure this proxy manually or was it auto-detected? If you click on the proxy entry, is "Automatically convert between dots and hyphens in proxied hostnames" enabled? (It should be.)
  • I'm quite sure it has been auto-detected, and the option you mentioned was not enabled.

    I ticked the option you mentioned but I still got the same issue. Expect for the mediapart website. I tried to manually add hostnames (+ button in proxy setting) but still it is not working.
  • Can you provide a Debug ID from the Zotero Connector for loading a proxied page, saving, and ending up with a proxied URL in the URL field?
  • edited October 26, 2022
    Ok, here is the ID: 724853195. I saw that the only error message from there is something related to my rss feeds, which is not related to our problem and was an error I got few weeks ago.

    Anyway let me know if I could do something else.

    Edit: Updated Debud ID instead of Report ID: D450346522
  • Those are both from Zotero, not the Zotero Connector.
  • My bad : D962533228
  • (5)(+0000013): Deproxifying item url https://www-arretsurimages-net.bnf.idm.oclc.org with scheme %h.bnf.idm.oclc.org/%p to https://www-arretsurimages-net.bnf.idm.oclc.org
    This is a bit of an odd example. It looks like you're just trying to save the homepage of a random website, which shouldn't need to be proxied in the first place. Zotero uses the embedded metadata provided by the site, which includes the homepage URL without a trailing slash, and that fails to match the proxy scheme. That's a bug, I guess, and we can try to fix it, but I can't recall this ever coming up — saving the homepage of a website loaded via a proxy that happens to include the root URL without a trailing slash isn't a normal thing.

    If you have other examples, we can look at those. We'd want to see Debug IDs from both Zotero and the Zotero Connector for a save attempt.
  • Hi, seems to be ok for everything else so nothing to add on my side.
    That is, every article saved has now a normalized url without suffix/prefix.
    Thanks!
Sign In or Register to comment.