Metadata from ScienceDirect

Hi,
Firstly, thanks for a VERY good tool: I think I'm at my 10 year anniversary now.

My Question:
I'm trying to see how ScienceDirect (I guess that means Elsevier in general?) allows access to its MetaData.

Specifically in terms of how Zotero can potentially get MetaData from references that are listed by ScienceDirect.

I see that they seem to say they do NOT automaticaly give full access to it:
https://www.elsevier.com/solutions/sciencedirect/librarian-resource-center/api
At least not unless you are a "subscriber": which means a university or other large library.
Right?

Although, somewhat confusingly, they DO say the following:
"Anyone can obtain an APIkey and use our APIs free of charge, provided that our policies are honored."
And a bit more about data-mining.

Therefore, has Zotero requested access to their API's?
In other words, a "free" key?
If so, great.
If not, why not ask?

If it has been granted, then to me I don't think the Zotero Metadata search engine is using it very well (no offence) but when I try to get MetaData for something that is definitely on ScienceDirect, I am frequently told "no Metadata available" by Zotero.

Let's use this as an example:
https://www.sciencedirect.com/referencework/9780122274107/encyclopedia-of-physical-science-and-technology

It has about 1,447 articles on it, and I have them.
But if I try to index on Zotero, and retrieve the MetaData using Zotero by dropping them on Zotero, I find it gets about 2/3 of them.
It misses about 500.

All are unlocked/no-security.
And their properties/metadata etc in Adobe seem identical.

If I use Google Scholar to check, using ones that Zotero DOES retrieve, and ones that it does NOT, then it's about the same.
As in, I can find the article when Zotero can, and cannot when it cannot.

In which case, I suspect it won't help me to write my own custom translator (right?).

I'm asking this from the perspective of of a non-specialist in terms of Information Science: please forgive the no doubt in-exact terminology.
Thanks,
Alan
  • There's a number of misunderstandings about how Zotero works here, I'm afraid, but to answer the main questions:
    - Zotero downloads the metadata directly from Sciencedirect, without using an API, by accessing the RIS available on article pages
    - Zotero doesn't use any publisher APIs to retrieve metadata for PDFs, nor would that likely be helpful. Zotero does use the CrossRef API for anything with a DOI or where a DOI can be determined.
    - If your goal is to write custom functionality to recognize a specific set of PDFs, that wouldn't be a translator. You'd have to hook into the recognizePDF functionality, which is integrated a good bit more deeply into Zotero and thus is significantly more difficult to modify than a custom translator.
  • Thanks for the reply: and I did say I'm a non-specialist user I don't really know the way Zotero works.
    I'm just going by what is on the Zotero site.

    (1) I WOULD like to understand Zotero better;
    Could you please give me a link that might explain things better?
    I searched before I asked my question.
    Perhaps what's there needs a bit of a re-write?

    (2) However, your reply did not answer my example. If Zotero can access metadata directly from Sciencedirect, then why doesn't it access the example I gave?

    The articles in my example all have DOI's.
    In fact the files are LABELLED by the DOI.

    And why I thought it involved a Translator, was that the Zotero documentation says translators are what Zotero uses to extract Metadata from websites.
    That's what it says here:
    https://www.zotero.org/support/adding_items_to_zotero

    So surely, Elsevier/ScienceDirect must have a pretty consistent set-up for Zotero to retrieve MetaData from?

    I can give you another example:
    V/Stol Airplanes by McCormick
    I search on Sciencedirect, and it finds it easily:
    https://www.sciencedirect.com/topics/earth-and-planetary-sciences/takeoff

    And here's its DOI:
    https://www.sciencedirect.com/science/article/pii/B012227410500911X

    And again, maybe my jargon is not right, but surely if the article has a DOI, and it's easy to find, then why can't that be used by Zotero to retrieve MetraData?

    (3) There DOES seem to be a bug in Zotero about how it accepts DOI's:
    Please read ALL my step:
    - I add the article in pdf form into Zotero
    - It says "No matching references found" (which is how my question started)

    So I then try to use a DOI:

    Method A:
    I right click on the pdf file, and select "Create Parent Item"
    It then gives me two choices:
    - Manual Entry
    - "Enter a DOI etc etc"
    BUT if I try to use the DOI here, Zotero gives an error:
    "Lookup Failed: Zotero could not find any identifiers in your input."

    So that stopped me.
    However, I see on reading more on using DOI's in Zotero you can input it another way:

    Method B:
    Toolbar on Zotero with Wand: "Add items by identifier"
    I then enter EXACTLY the same as I did in Method A...and it WORKS.
    It FINDS the article.
    Great.

    But Method A has a bug in it.
  • The DOI isn't in the actual PDF -- not sure what you mean by "labelled with" but if you mean that it's in the filename, then Zotero doesn't look there simply because there's typically no useful information.

    As for Methods A vs. B I can't replicate this.
    If I download the PDF from
    https://www.sciencedirect.com/science/article/pii/B012227410500911X
    as you say, retrieve metadata doesn't work (because we don't have a DOI to use), but right-click --> Create partent item, and pasting 10.1016/B0-12-227410-5/00911-X into the window works for me. It doesn't for you?
  • Hi Adam,
    And firstly, thanks for your usual patience in answering questions.

    But no, Method A doesn't work for me.
    Specifically, when I am at the pdf in the main pane of Zotero, and Right Click.
    I get various options, from:
    View PDF
    Show File
    Show in Library
    ...
    Create Parent Item
    Reindex Item
    Update Citations

    So I click on "Create Parent Item", which gives me an entry sub-window with:
    area to input data
    "Manual Entry", "Cancel", Create Parent Item" (greyed out until you enter some characters in the data entry spot)

    A few issues:
    I CANNOT use Paste from my mouse to enter data there.
    I also cannot use Ctrl-V/Ctrl/C from, keybourd to copy/paste the DOI info.
    It ONLY accepts manual typing in.

    And THEN when I type in the correct DOI (https// etc etc etc)....it THEN says "Trump is the Emperor.....you are forked".
    OK, I joke.
    It says "nothing found".

    I'd send you screen shots but I think this forum too basic, yeah?

    HOWEVER, when I use the option at tool bar, it works (Method B).

    So look, no big deal, I figured it out.
    But there definitely IS a bug in Zotero.

    I'm on a Mac, High Sierra.
    Zootero 5.0.95.3

    If you want to follow it up, tell me what you need.
    But from my end it seems unless I gear up to to use JavaScript better I'll leave it as it is.

    However, I see that as of 2019 there's a plug-in for DOI's: I'll try that.

    Alan
    P.S. I am not actually a COMPLETE idiot....a few parts are missing.

  • I CANNOT use Paste from my mouse to enter data there.
    I also cannot use Ctrl-V/Ctrl/C from, keybourd to copy/paste the DOI info.
    It ONLY accepts manual typing in.
    It looks like right-click isn't working there — we'll fix that.

    But you paste on a Mac using Cmd-V, not Ctrl-V, and that works fine there. You absolutely don't need to type manually.
    And THEN when I type in the correct DOI (https// etc etc etc).... […]
    It says "nothing found".
    I'm not sure what you're typing, but to be clear, it has to contain a DOI. It can't just be a URL, though we might support that in the future. A doi.org URL is fine, because it contains an actual DOI.

    If you think you're pasting in something that should work, we'd want to see a Debug ID for that.
Sign In or Register to comment.