SilentSpike

talk

Previous discussion was archived at User talk:SilentSpike/Archive 1 on 2019-03-26.

I started trying to match the UKRI Gateway to Research organisations Mix'n'match catalog to Wikidata items. I noticed that every item in the catalog returned an error page from UKRI Gateway to Research, and initially assumed that this was an error on UKRI's server. However, now I have noticed that one item in the catalog, Chatherm Technologies (Q122841603), had an entry in the catalog for https://gtr.ukri.org/organisation/CF962C09-F22F-4837-AB1C-D9747E89A3E0 which is an error page like many other items, but has a valid identifier at https://gtr.ukri.org/organisation/F3DFE2CF-BED9-42CA-9DD2-D9DA15FC8003 . What's going on here? Options I see:

1. The Mix'n'match catalog has been wrong since its creation due to vandalism or a technical error.
2. UKRI identifiers have not been persistent, and have changed at some point between the catalog being created and now.
3. UKRI had duplicate entries and instead of merging the identifiers through redirects, simply deleted the duplicate entries.

In any case, my experience indicates potentially serious problems with the Mix'n'match catalog data, and I wonder if we should:

1. Report the issue
2. Remove the catalog from Mix'n'match.
3. Replace the catalog with updated identifier data.

What is the source for the data in the Mix'n'match catalog?

Hey there, I'm not as active as I once was, but good find!

The identifiers for the catalog were scraped directly from the website and I can say with certainty were correct before (from my contribution history I can see one's I manually added are now broken).

I suspect it's not the catalog that's broken as much, but that the UKRI has re-assigned all of their identifiers. In which case, if you want to chase this up, you should probably start a discussion on the property's talk page or project chat. I'd start by checking if anyone knows more or if we've seen this in the past with other identifiers.

I want ahead and kicked off a section for this on project chat. Suggest further discussion take place there!

I came across the Aberdeen Built Ships database you imported last year, and noticed that you left the last 30 or so, that lacked names, unmatched. I decided it was worth adding them anyway, as they do have at least minimal info in them, enough to distinguish them. Before I go thru the rest (and in case you feel like doing them or have suggestions), I thought I'd check with you.

Hey man, what happened to your announced big overhaul of the WikiProject Mountains project page? I'm looking forward to see somebody who intends to revive this project and I'd be glad to cooperate. All the best!

Hey, that might be partly correct. However, both pages which I edited do have Wikipedia articles connected. The description is shown as short summary of the idea behind the object on various places (e.g., Commons search). “Wikimedia compound page” might be an internal information but isn't useful for other users which likely see that even more often than internals. Since I do not want to start an edit war, I'm asking you to help me find a solution for both requirements. A comma? “Wikimedia compound page about XYZ”?

Hello!

Just to provide context for any future readers: pages in question are floor exercise (Q828144) and still rings (Q655318).

I also have no intention of edit warring here, I'm primarily a Wikidata contributor so was considering the description from a Wikidata centric point of view where is it useful for the description to clearly mark the item as intended to be a compound page.

As for resolution, this is a classic and long-standing problem (described at Help:Modelling/Wikipedia_and_Wikimedia_concepts#Compound_Wikipedia_articles) which I admit I don't fully have the answer to. In an ideal world I suppose the commons category would be linked to the item specifically for the event since it seems to be representing that over the apparatus itself, but I'm guessing that would break the direct link from commons to other wikis?

Perhaps, as you suggest, a description "compound page about the event and the apparatus" would be best. We may also get more experienced suggestions if asked at project chat.

Thanks for the hints. I was pretty sure that it was discussed before. I'll read through that when I find a bit of time and do as suggested. Best!

Scraper added

Hi! Could you please add scraper to mix-n-match catalog you created, so it could automatically update on its own?

https://mix-n-match.toolforge.org/#/scraper/new

Here's parameters:

Catalog ID: 3476

Press "Add keys level"

URL pattern: https://musicbrainz.org/genres

RegEx entry: <li><a href="/genre/(.+?)"><bdi>(.+?)</bdi>

id: $1

name: $2

url: https://musicbrainz.org/genre/$1

Press "Test scrapper"

If it gives 800 entries then press "Save scraper/catalog".

Hey, thought I originally had. Now done.

Proposal reviewed

Hi!

Could you take a look at Wikidata:Property proposal/Tax identification number and create this property when you have a chance?

Hey there, I can take a look yeah! :)

Thank you! Please let me know if something needs to be improved before the property is created

Hello! I think that these auxiliary matches in UKRI catalogue are hilarious ;-) They are connected to cities, villages etc.

Hey there :) Yeah it's pretty silly, seems to just be matching the street name for some, I was trying to see if there's a way to just unlink them all so if you know of one let me know

Hello Vunja6dcd966mmru,

Really sorry for the inconvenience. This is a gentle note to request that you check your email. We sent you a message titled "The Community Insights survey is coming!". If you have questions, email surveys@wikimedia.org.

You can see my explanation here.

What is the cause that you are removing the TripAdvisor ids (P3134)?

Hey, the property format was recently updated to be more correct (only the digits are the true ID).

So my Quickstatements batch should not be removing IDs, but replacing them with the correct format. If you've come across a case where the ID was strictly removed please let me know, thanks.

Hello. Why is bot replacing Q48799541 with Q28378282?

Hey, that's to make the data homogeneous. You can see from the following query that the latter is preferred:

SELECT ?qualifiervalue ?qualifiervalueLabel (COUNT(?item) as ?ct) 
{
   ?item p:P2002 [ pq:P1552 ?qualifiervalue ]
   SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }               
}
GROUP BY ?qualifiervalue ?qualifiervalueLabel 
ORDER BY DESC(?ct)

Try it!

I'd refer you to Property_talk:P2002#Representing_various_account_qualities where I've started a discussion on how to represent various account qualities.