User talk:Magnus Manske

Jump to: navigation, search

About this board

Previous discussion was archived at User talk:Magnus Manske/Archive 9 on 2015-08-10.

By clicking "Add topic", you agree to our Terms of Use and agree to irrevocably release your text under the CC BY-SA 3.0 License and GFDL

Proposal for simplification: consistency, gender-neutrality and age-neutrality

Robin van der Vliet (talkcontribs)

I remember you participated in the previous discussion about the gendered properties. I created a new discussion here regarding the property relative (P1038). Could you please take a look? Thank you. Robin van der Vliet (talk) (contributions) 00:31, 20 February 2017 (UTC)

Reply to "Proposal for simplification: consistency, gender-neutrality and age-neutrality"
Parikan (talkcontribs)

Hi Magnus, I have been using mix'n'match for a few days and this is a fantastic tool. I just wanted to let you know of a bug I encountered a couple times in both Firefox and Chrome.

In a nutshell, the page freezes after clicking 'not on wikidata'. It doesn't create the new item, yet the action 'not on wikidata' gets recorded in 'recent changes'. It could be deceptive to users who assume the item was successfully created by the program.

This bug is not super common (for me it occurred maybe once for every hundred uses of the tool) and the only reason I noticed it is because sometimes I want to add more statements manually and the item isn't listed on my 'contributions' page. One more thing I'd like to mention is that I don't have any extensions in Chrome that could interfere with the program. Anyways, thanks a lot for creating the tool, and I hope my feedback is useful.

Reply to "Mix'n'match freezing"
GZWDer (talkcontribs)

please fix it.

Magnus Manske (talkcontribs)

Works for me. Example?

GZWDer (talkcontribs)

Now fixed.

Vladimir Alexiev (talkcontribs)

Rename the current one in MnM from "BMT" to "BM person-institution".

Then could you import 2 more that were just accepted as properties? RDF or CSV is available:

Reply to "2 more British Museum thesauri"

historisches Territorium > Ortsteil

Holger1959 (talkcontribs)

Hallo Magnus, mit Reinheitsgebot erstellst du auch Ortsteil-Items passend zu neuen dewiki-Artikeln. Manchmal steht dann Historical territory (Q1621034) (Herrschaftsgebilde des Heiligen Römischen Reiches) drin, sonst oft "Ortschaft". Das falsche "Historische Territorium" hab ich bislang nur bei Items gesehen, in deren Artikel eine Kategorie "ehemalige Gemeinde" drin ist. Richtig wär aber ortsteil (Q253019). Beispiel Kannst du Reinheitsgebot das beibringen?

Magnus Manske (talkcontribs)

Das wird zu schnell zu komplex. Habe die "Historisches Territorium in Deutschland" Regel erst mal abgeschaltet.

Ham II (talkcontribs)

Hi Magnus. Last year Reinheitsgebot added Grade II listed buildings in England to Wikidata; are there plans to do the same for ones in Wales? They would use Cadw Building ID (P1459) rather than National Heritage List for England number (P1216); the higher grades were added long ago.

Magnus Manske (talkcontribs)

Hi Ham II, I was given that list (by WMUK, I believe), but currently I have no further data to add.

Reply to "Grade II listed buildings in Wales"
Jheald (talkcontribs)

Finally to let you know that all of the Picto-DACS identifiers seem to have changed -- the old identifiers no longer resolve to the names they used to. Presumably it may be possible to update to a new scrape of the database, and carry forward existing matched names; but I don't know enough about the database to know how stable (or stable for how long) these new identifiers are likely to be, or whether this is something that may have to be monitored on a regular basis.

Magnus Manske (talkcontribs)

I can try, but then I will have to remove all matchings that were done, as they are no more reliable.

If the catalog keeps doing this, it will be too unstable to warrant a place in mix'n'match at all.

Jheald (talkcontribs)

The name strings may be the same (I think they are), even if identifier number has changed.

So it may be possible to update the matchings that were done to their new identifiers.

But I agree, if these changes go on, the catalogue will be of limited use.

Reply to "Picto-DACS"
Multichill (talkcontribs)

Hi Magnus, in Mix'n'Match, Art UK artist ID (P1367) is still called YourPaintings, can you rename that? It seems to have 36,784 entries, but their website now has 38,402 entries. Maybe a good plan to sync it?

Magnus Manske (talkcontribs)

Renamed. Tried scraping, but the incremental list only covers ~3800 artists.

Multichill (talkcontribs)

Thanks for the rename. What exactly went wrong with scraping? I made wikidata/, maybe it contains some useful info for you.

Magnus Manske (talkcontribs)

All these searches terminate after ~30 pages. That's enough for your sets apparently, but "all authors" is too gigantic.

Multichill (talkcontribs)

The trick is that you need to grab the json which contains the html. See . That's just normal paging

For example:

import requests
searchurl = ''
searchpage = requests.get(searchurl, headers={'X-Requested-With' : 'XMLHttpRequest',} )
searchjson = searchpage.json()
searchtext = searchjson.get('html')
print searchtext
Magnus Manske (talkcontribs)

Actually, your example link just gives me a 504 gateway timeout.

Reply to "YourPaintings -> Art UK artist ID"
Multichill (talkcontribs)

Looks like Tate artist identifier (P2741) ran out of suggestions in mix'n'match. Can you rerun the auto matching job? I think it will pick up some new suggestions.

Magnus Manske (talkcontribs)

Ran quick'n'dirty auto-matching. Running thorough one now.

Jheald (talkcontribs)

Hi Magnus. I just wanted to thank you for all the work you've put in creating, maintaining and updating Mix'n'match. It really is hugely appreciated.

As regards the Art UK painters, I wanted to let you know that I think all the identifiers currently on Wikidata are correct -- at least insofar as I think they do all now link to currently valid Art UK pages. (Apart from a handful marked as deprecated, for which it seems there is no new updated page).

In about 200 cases this means I have changed an old identifier value to a new one. In a few cases this has also led to two or more Wikidata items being merged. (It seems the Art UK identifiers can sometimes be a bit unstable -- eg they change if Art UK become aware of a date of death for an artist).

I was wondering: will Mix'n'match pick up these changes automatically? Also, am I right to presume there's no risk of Mix'n'match writing an old, defunct previously-matched identifier back to Wikidata?

Thank you again for all the efforts you have been putting in on this -- it is so, so valuable.

Magnus Manske (talkcontribs)

Hi Jheald,

no, mix'n'match will not auto-update from the primary source in general. I am working on a mechanism to do this for some cases, but it's not ready to roll out yet.

An identifier that had been added once and since removed on Wikidata should not be auto-added there.