Wikidata talk:Zotero

From Wikidata
Jump to navigation Jump to search

Thank you very much for writing this translator. I tried out the translator and I have found it to be very accurate and easy to use. I noticed that in the publications I created items for using zotero -> quickstatements statements using published in (P1433) do not get created. Is that correct, and is that on purpose? I added them myself, but I wanted to ask as I imagine many people will want to also have published in (P1433) statements get created via this translator. Thank you! YULdigitalpreservation (talk) 16:26, 12 July 2017 (UTC)[reply]

I also noticed that the translator provides full work available at URL (P953) links. The majority of the links added for the ~30 publications I was working with are behind publisher paywalls. Do we want to provide those links if so many of them are barricaded? YULdigitalpreservation (talk) 17:39, 12 July 2017 (UTC)[reply]

published in (P1433) needed to be added manually; the relevant QIDs do not exist in Zotero libraries. There is no prohibition on paywalled links; but that does not stop another editor, or bot, from either adding a second link or replacing the link with an open one Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:35, 12 July 2017 (UTC)[reply]
Thank you, that answers my questions. 130.132.173.232 18:24, 13 July 2017 (UTC)[reply]
@YULdigitalpreservation, Pigsonthewing: There is (now) a new version of the script which allows also to export additional PQ-pairs from the extra field in Zotero. For example you can add "P1433: Q29310506" in the extra field of an Zotero item and it will be exported as the correpsonding statement. This can be already useful if you add several articles from the same journal, where you can look up the journal once in Wikidata and then copy this line manually in many Zotero items. Moreover, it should be possible to write a little Zotero script which could perform an automatic lookup in Wikidata based on the ISSN. Stay tuned for that. --Zuphilip (talk) 20:24, 29 December 2019 (UTC)[reply]

Can you add any Mac Specific instructions for the installation?[edit]

I have loaded the file into the Translator folder, but the option isn't appearing in the Default Format listing on Export tab. Trilotat (talk) 01:04, 12 October 2018 (UTC)[reply]

There is no mac specific installation for mac, because Zotero and the Zotero Translators are platform-independent. Just make sure that you copied the correct file and restarted Zotero once. --Zuphilip (talk) 11:36, 27 October 2018 (UTC)[reply]

Can someone fix the Zotero Quickstatements exporter?[edit]

There are issues that I always have to fix using this tool:

Can someone address this? - PKM (talk) 21:06, 14 June 2020 (UTC)[reply]

@PKM: The labels values for the QuickStatements import are in quotes and dates work for me as well. Can you share your Zotero item which does not work for you as expect it?
Changing the type for books to export in Wikidata is easy to change, but do you have some source that this is the correct way to do? --Zuphilip (talk) 13:07, 24 October 2020 (UTC)[reply]
@zuphilip: I've been trying to pick up some examples for you, but Zotero isn't grabbing the author, publication date, or ISBN for anything I try. Here's an example https://boydellandbrewer.com/medieval-clothing-and-textiles-16.html and here's another https://www.amazon.com/FBI-History-Rhodri-Jeffreys-Jones-ebook/dp/B00I50EXVY/ref=sr_1_9?dchild=1&keywords=Yale+university+press&qid=1603569098&s=digital-text&sr=1-9. In both cases, the exported text file has no quotes on the EN label.
As far as the P31, see Wikidata:WikiProject Books for the distinction between a "work" (notional text) and an "edition" (specific published form of a work). Only editions should have publishers and ISBNs, and be used for citations. - PKM (talk) 20:06, 24 October 2020 (UTC)[reply]
@PKM: Thank you for the examples. The Boydell Press website has no specific metadata for Zotero and there is also no other specialized Zotero translator for that. Therefore, the Zotero symbol in your browser is only showing a website and not the book, which means that you get almost no metadata by using it. However, you can copy the ISBN and use the magic wand in Zotero to add this book by identifier. Amazon is supported by Zotero and the example you mentioned works fine for me. However, I think usually the hardcover version in Amazon has some more metadata for Zotero to save.
In both cases I see then that EN label in the WQS export has quotes around, e.g. LAST Len "Medieval clothing and textiles", which works without problems. You should see exactly the same line in your export.
Okay, thank you for the info about the work vs. edition. As I want to export as much as possible this is only possible in an edition. Thus, I will change that. However, it would be possible to add a hidden preference in Zotero to switch to export as book, but then restrict the exported fields, but this is maybe something for the future to work on... --Zuphilip (talk) 08:46, 25 October 2020 (UTC)[reply]
Thank you for the explanations. - PKM (talk) 21:07, 25 October 2020 (UTC)[reply]
@PKM: The best way to add book in Zotero is to use: Add item(s) by identifier. The identifier is the ISBN. LaMèreVeille (talk) 22:31, 25 October 2020 (UTC)[reply]
@LaMèreVeille, Zuphilip: "Add item(s) by identifier" worked very well - but the EN label still comes over without quotes for me (easy fix in QS) and the title has "english:" rather than "en:" (also an easy fix). Thanks for the hint! - PKM (talk) 22:59, 25 October 2020 (UTC)[reply]
@PKM: Is it possible that you speak about the title (P1476) instead of the label? Anyways, try to download the newest version of the WQS translator from Zotkat, which I updated yesterday and possibly you need to fix the language codes in Zotero before the export, e.g. use "en" for English instead of "english". --Zuphilip (talk) 07:36, 26 October 2020 (UTC)[reply]
@Zuphilip: it’s definitely labels - titles are fine. I probably have an old version. I’ll work on getting everything updated today. Thanks for all your help. - PKM (talk) 19:47, 26 October 2020 (UTC)[reply]
Much better now! Labels have quotes. "Pages" value causes an error in QS (LAST P1104 "201"). Thanks again! - 20:55, 26 October 2020 (UTC)
@PKM: Ah, yes, the values of P1044 are not strings and therefore no quotes should be around them. Fixed in the newest version just pushed now. Thank you for the report. --Zuphilip (talk) 08:32, 27 October 2020 (UTC)[reply]

Google Books[edit]

I've tried this out for the first time and it works great, thank you! I grabbed all the data I needed from Google Books, but the output doesn't include a Google Books ID. Is there any way to add that to the QS output? Thanks. Gamaliel (talk) 13:19, 17 July 2020 (UTC)[reply]

@Gamaliel: Good idea, yes I can add that identifier as well in the next version. Since you are working with books as well, what type should be used for them? book (Q571) or version, edition or translation (Q3331189)? Cf. discussion above. --Zuphilip (talk) 13:20, 24 October 2020 (UTC)[reply]

Example[edit]

I've moved the appended here, from the page, since it seems to duplicate a lot of what already existed, and not to be integrated with it.

'Example' 
==Example==
[[File:ASC_Leiden_-_Library_-_Journals_-_2017.jpg|thumb|Journals at the Library of the African Studies Centre in Leiden]]

'''How to import scholarly articles to Wikidata using Zotero'''

=== Introduction===

The [[w:en: Afrika-Studiecentrum Leiden|ASCL]] started adding journal articles to Wikidata by using Zotero because this method of bibliographic data ingestion is likely the easiest way to import bibliographic data to Wikidata. The upload procedure consists of "prefabricated" parts; special technical skills are not required. 

Once the bibliographic data is copied to [[w:en: Zotero|Zotero’s library]] - via f.e. Zotero’s in-built RSS Feed Reader or the Zotero connector, a browser extension that recognizes bibliographic data on websites - the data can be easily converted into the Wikidata Quick Statements format and uploaded to Wikidata. 

This workflow is especially suitable for journal articles and less for other material types as Zotero’s download file in the Wikidata Quick Statements format lacks essential information for other material types than journal articles. For example, the publisher and place of publication are missing in a book’s download file. This information is missing because there is no direct link between Zotero and Wikidata. Zotero cannot retrieve the publisher’s unique identifier (its Q-number).

===Workflow in a nutshell===
[[File:How to import scholarly articles to Wikidata using Zotero.webm|thumb|How to import scholarly articles to Wikidata using Zotero, screencast video, 2020]]
#Add the scholarly articles to your Zotero Library.
#Download the scholarly articles using the Wikidata Quick Statements export format.
#Check and amend the scholarly articles in a text editor.
##Check for duplicates.
##Add the source (P1433 ''published in'').
##Check and correct - if necessary - the language of the title (P1476 ''title:'' '''und''':“Title“ > '''en''':“Title“).
##Check and correct - if necessary - strange/garbled characters.
#Upload the scholarly articles via [https://quickstatements.toolforge.org/#/ Wikidata Quick Statements 2].

===Documentation===
*[[Media:How to import scholarly articles to Wikidata using Zotero.webm|Workflow ASCL, screencast video, 2020]]

and further

* [https://www.youtube.com/watch?v=lP4ugVkcJSU How to install Zotero]
* [https://www.zotero.org/support/feeds Zotero feeds functionality] 
* [https://github.com/UB-Mannheim/zotkat Zotkat (for downloading Wikidata QuickStatements export format)]
* [https://notepad-plus-plus.org/ Notepad++]
* [[User:Walkuraxx/e-scholarship#How_to_import_books_to_Wikidata_using_a_spreadsheet_and_OpenRefine|How to import books to Wikidata using a spreadsheet  and OpenRefine]]
* [[User:Walkuraxx/e-scholarship#How_to_import_library_materials_to_Wikidata_using_MarcEdit_and_OpenRefine | How to import library materials to Wikidata using MarcEdit and OpenRefine]]

Perhaps it should go on a sub page? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:42, 3 March 2021 (UTC)[reply]

Dashboard[edit]

Please kindly send me the link to the dashboard. Thanks DaSupremo (talk) 21:37, 20 May 2021 (UTC)[reply]

Widen and standardise Zotero support of Wikidata properties and classes[edit]

On a separate topic where I wondered about the criteria to decide whether a source belongs to Wikidata or not, I wondered whether having a unique identifier could be one such criterion.

This led me to think that it would be nice that Zotero's Wikidata translators (both import and QuickStatements export) supported unique identifiers other than just DOI and ISBN. A recent change made to the Wikidata import translator added support for QID, but there are many other unique identifiers that could be supported as well. In absence of custom fields, users usually add this information to the extra field (e.g., PMID).

Hard-coding a map between Wikidata properties and their corresponding Zotero field (as currently done in Wikidata translators) is both cumbersome and fragile. Same thing applies to mapping Wikidata classes to their corresponding Zotero item type. See for example this issue where some encyclopedia articles are no longer supported, or this other one where some trial cases are not supported either.

I think one long-term solution for this could be adding a equivalent property (P1628) property to supported properties, indicating the name of the corresponding Zotero field, according to the Zotero schema. See, for example, how this is done with DOI (P356) and the Bibframe ontology. Analogously, one could add a equivalent class (P1709) property to a Wikidata class (such as scholarly article (Q13442814)), indicating the name of the corresponding Zotero item type, also according to the Zotero schema.

I guess we could already use these equivalent property (P1628) and equivalent class (P1709) properties. We would need to have a URL to point to the corresponding Zotero fieldnames and item types, though. For example, Wikidata property DOI (P356) could have property equivalent property (P1628) = https://.../zotero/DOI, or Wikidata class scholarly article (Q13442814) could have property equivalent class (P1709) = https://.../zotero/type/journalArticle.

In cases where Zotero doesn't have a specific field for a Wikidata property, we could come up with a way to indicate how it should be included in the extra field. This could in turn help make sure all users do it the same way. For example, the Wikidata property PubMed ID (P698) could have equivalent property (P1628) = https://.../zotero/extra/PMID, or something generic such as https://.../zotero/extra/wd.P698

Alternatively, if having a schema URL wouldn't work, we could create two new properties in Wikidata: Zotero fieldname, and Zotero item type.

@Zuphilip: what do you think about this? I'd appreciate it if you could ping anyone who you think could contribute to this discussion.

Thanks! --Diegodlh (talk) 16:19, 31 May 2021 (UTC)[reply]


@Diegodlh: Sorry, for the delayed answer here.
That is an interesting idea to use equivalent property (P1628) and equivalent class (P1709) to make the mapping to Zotero part of the knowledge graph of Wikidata directly. I would however concentrate on the publication types and think that a property list and mapping could be easier to have directly in a translator.
Technically I could become a little more trickier to implement this, because AFAIK this information would not be part of what we get from one individual item. At the moment Zotero takes the RDF-XML data of an item (e.g. [1]), where no further information/data from the class scholarly article (Q13442814) like P1709 is present. Correct? Or is there any other more extensive RDF support?
One could also try to make further calls for all P31 nodes. That can be a little tricky and should be discussed before. But we do something similar in the Amazon translator and call OCLC for the publisher (?) if missing. If you would here also look up each property node, then I fear that these are too many lookups such that the data grabbing by Zotero is becoming too slow.
Besides or instead the Zotero vocabulary one could also use CSL vocabulary, which is very near and maybe of larger interest.
--Zuphilip (talk) 12:06, 18 August 2021 (UTC)[reply]

Intetion[edit]

I am not sure if this still work. Was the intetion to generate citation via QID, or hot it suppose to work? Juandev (talk) 22:46, 11 February 2024 (UTC)[reply]

Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata_talk:Zotero&oldid=2074901339"