Sorry, the person who was usually merging these requests hasn't been working on the project for some time lately. I just merged the pending pull request.

Gracias @Sophivorus! That's all very useful information.

Also @Sophivorus recommended maybe using mw.track() as used in ProveIt gadget.

I see today a" Powered by Wayback Machine" message in the citation dialog (with Web2Cit disabled). Consider T408872.

What about {{#invoke:wd|title|linked|Q2}} or {{#invoke:wd|label|linked|Q2}}?

Make sure we resolve T408860 before.

Apparently not yet supported for web services? T362077

The undefined fields of the resulting citation (...) are automatically filled in with the corresponding Citoid response.

Alternatively, in the meantime, just remove the column to avoid confusions.

In wikitext, pipes can used as {{!}}.

Consider this comment here: T308328#7927379

Also note what would happen if a path and field, for which a procedure has been defined but no goal has been set, is opened by the Web2Cit-JSON-Editor form? Will it fail? Will it prevent users from saving until the error has been fixed?

The idea is to post this spreadsheet to Meta

Nidia posted this to Meta and linked to it on 2023-05-11: https://meta.wikimedia.org/w/index.php?title=Web2Cit/Research&diff=prev&oldid=25006825

Then use range transformation "0" to keep the first item selected.

When this is fixed, consider notifying interested users that they can now watch result pages to monitor score changes:

it gets the list of configuration files, and asks for the latest revision timestamp, one by one.

Consider potentially related tasks T324933 and T387039.

Analogous to T317168.

a collaborative repository of community-curated transformation functions.

This seems to be working much better with Citoid/Zotero now. Only the first name is retrieved slightly wrong, with a "Por" (By) prefix.

I just checked these three URLs again and they seem to be working with Citoid now. If you can confirm this maybe @Mvolz would like to remove the Citoid project tag from this?

may reduce the cases where a date field would be missing, but may result in wrong force converted dates.

Note this would already be available from the monitor's result pages if T317977 is addressed.

Alternatively, consider adding a tag to the edit summary if a Web2Cit-generated citation was inserted. This seems to be what Proveit does.

Whatever wasn't working seems to be working now.

This seems to be fixed already.

Hi, @AlexisJazz! Just checked this again for https://dictionary.cambridge.org/grammar/british-grammar/could-may-and-might and it seems to be working at least for me now. Could you check it too? Shall we close this?

There is a problem with this domain https://www.bcn.cl/, when we use it in automatic citations it only retrieves the main domain but not anything after.

Hi, @AlexisJazz! Sorry, I just saw this.

In the meantime, a Split transformation step may be used first to split a string into its characters. However, this would only work as expected if the original input is a single string (i.e., not an array of strings).

the idea would be that requests originating from user scripts or gadgets set their own api-user-agent. However, because of the way how Web2Cit-Gadget patches the Citoid extension, doing this may not be as straightforward.

Fixed with https://gitlab.wikimedia.org/diegodlh/w2c-server/-/commit/1d92b7b6a17cca4dae646c13410a29440ddd8d5a

Thanks for reporting this.

As described in T305166, there may be cases where the fallback template (which currently uses Citoid selection for all fields) does not apply.

Interestingly, this does not happen when the template uses fixed-value selection steps only:

This selection step may be useful in cases where we can't get the webpage because we have been blocked. As suggested in T305168, we could use such a URL selection step to guess a title.

Tool is cool, but it seems not entirely compliant with https://www.mediawiki.org/wiki/Recommendations_for_gadget_developers_on_Wikimedia_wikis

Client-side metadata extraction sounds like an interesting idea to both circumvent sites blocking Citoid (as mentioned in T362379) and to support sites which rely on the browser running some JavaScript code to display content.

Client-side metadata extraction and citation generation, as proposed in T362379 to circumvent the problem of sites blocking Citoid, may also result in supporting JavaScript-generated content.

As mentioned in the original thread linked to above, we have decided not to make these message key id and variable renames for the moment. We may do them some time in the future, but maybe Fluent file format will be supported by then, who knows.

Prototype available at https://meta.wikimedia.org/wiki/User:Diegodlh/search-wpl.js

One concern here (apparently not mentioned before) is that not all constraints can be satisfied by a single edit, for example “symmetric” and “inverse"

Hi, @Ivi104! Thank you for your interest in Web2Cit and for posting your question here.

Merged. Sorry for the inconveniences. We will try to merge this more regularly. Thanks for your understanding!

no easy way to edit this configuration (in cases where doing so may be useful).

Note that the error persists when the template is changed to be non-applicable.

@DrThneed reported in T313236 another URL for which Citoid (and therefore Web2Cit fallback template) fail: https://teara.govt.nz/en/biographies/4g2/galbraith-alexander.

Thank you @DrThneed for reporting those sites.

@CristianCantoro would this be related to what you mentioned at the pitching session? I'd love to collaborate!

Also potentially related: https://wikicitations.wikibase.cloud/wiki/Main_Page

@Isaac, have you heard of something like this before? What do you think about it?

Do you (or could you) support Cite Q through this tool? Also see T289287.

I've just asked the community around the Cite Q template whether they know if such an analysis has been done already: https://en.wikipedia.org/wiki/Template_talk:Cite_Q#Template_usage_across_Wikipedias

@Pablo mentioned the tool https://github.com/internetarchive/iare. A couple months ago there was https://internetarchive.github.io/ware/, which seemed to be a frontend using IARE on the background. However, the site is down now, and the repository (https://github.com/internetarchive/ware) redirects to the IARE repository. Maybe @Harej can provide further information?