User talk:Magnus Manske

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

Could you add Isabelle de C harrière (Q123386) with zulu wikipedia, where she in the wrong Isabelle wikidta.

https://zu.wikipedia.org/wiki/Isabelle_de_Charri%C3%A8re

Not sure what needs to be entered in Petscan to query categories of https://wikisource.org .

"sourceswiki" works for the "manual list" feature.

Catalogs needing auxdata:

• catalogs for PLWABN (4050; 4051; 4052; 4055; 4056) need dates, VIAF, ISNI  Done
• PIAC (4066) needs dates  Done
• UCAL people (3965) needs dates  Done
• KANTO people (4094) needs dates  Done
• AAROME (4152) needs dates and Library of Congress authority ID (P244)  Done
• Patrinum people (4182) needs dates  Done
• ISVROMA authors (4189) needs the formatter URL https://biblio.isvroma.it/cgi-bin/koha/opac-authoritiesdetail.pl?marc=1&authid=$1 and dates  Done using 4197
• INE authors (4207) needs dates
• BFS authors (4208) needs dates

Catalogs needing a new import:

• MutualArt (1140): it contains both good entries (e.g. this) and wrong entries not respecting the regex (e.g. this), so it would be necessary a new import
• OLID author (98): it contains both good entries and wrong entries not respecting the regex (e.g. this), so it would be necessary a new import
• Bénézit: the old catalog (143) has slightly fewer entries and shorter auxdata, whilst the new catalog (768) has IDs not fitting the regex; it would be necessary a new import in order to have a catalog possessing the best qualities of both the above
• FAST (150): the sex or gender (P21) contained in the auxdata is often wrong (e.g. this and that contain "sex: female" which is absent in the relative entries, and wrong); since unfortunately the error seems widespread (e.g. also this and this other), I ask the removal of all sex or gender (P21) from auxdata, preferably soon. Thanks!  Done
• Artnet (523): the encoding of non-ASCII letters violates the regex (e.g. in this); it would be necessary a new import in order to have a catalog fully respecting the regex

Catalogs needing a change in formatter URL:

• Saamelaiskulttuurin ensyklopedia (3115): new formatter URL https://saamelaisensyklopedia.fi/wiki/$id

The "Google-search" links for unmatched entries on Mix'n'match are broken. Could you fix them? Thanks!

Hi! I and @Bargioni: have just uploaded https://mix-n-match.toolforge.org/#/catalog/4189, representing https://www.wikidata.org/wiki/Wikidata:Property_proposal/Svenska_Institutet_i_Rom_ID (will be created in a few days). However, there's a problem: the new import interface, while reading the formatter URL from the P if present, doesn't allow to define a formatter URL if the P doesn't exist yet. Could you add https://biblio.isvroma.it/cgi-bin/koha/opac-authoritiesdetail.pl?marc=1&authid=$1 as formatter URL to this catalog and fix the problem in the import tool? Thanks!

I've fixed the problem of catalog 4189 deactivating it and reimporting it as 4197 after the property had been created. However, the problem is still to be solved: it should be possible to indicate a formatter URL if the P doesn't exist, as it was possible before, with the old import interface. Thanks!

Hello Magnus Manske, the name is spelled Johannes Henderikus Morrie͏̈n, and not Johannes Henderikus Morrie ¨n. can you fix this? Thank you for your time.

In pagepile tool, there are several places referring the domain toolforge.com instead of toolforge.org. Please fix it.

Dear Magnus! I recently uploaded a csv file to Mix'n'match to create a new catalog. The data was in this format:

40,ANDRÁSSY Karolina csíkszentkirályi és krasznahorkai gr.,1865. IX. 18.,1937. IX. 10.
41,KEGLEVICH Imre gr.,1897. X. 27.,1964. VIII. 19.
42,PIRET DE BIHAIN Lajos Jenő br.,1862. I. 27.,1910. XII. 2.

However, after uploading, I noticed that birth-death data can not be found. Will you help me see what the error was and whether it can be fixed by re-uploading the repaired dump?

Thank you!

m:Mix'n'match/Import explains the date format.

Also, names (read labels) would ideally have the same format as on Wikidata. For Hungarian, there is the additional problem that the label is different from English.

Not sure if there is a way to upload both formats in one catalogue. Maybe you could upload the data twice: once in English (<given name> <surname>), once in Hungarian (<surname>, <given name>). If someone matches a name in either catalogue, eventually resync will match the second entry.

Hello Jura1! Thanks for the reply! The header originally looked like this:

id, Name, BirthDate, DeathDate

I need to fix this:

id, name, born, died

and the date format

1786. XI. 9 -> 1786-11-9? So it will be good? Shall I upload as a new catalog? (Don't sync old 4183)

Unfortunately, I can't correct the name, not only because of the East-West order of names, but because the noble first names are included, I can't take them out

I suppose you could update the catalogue per explanation at m:Mix'n'match/Import or create one (or two) new ones. If new ones are created, 4183 can be de-activated.

I don't quite understand why you don't want to change the names to Wikidata label format. This would simplify matching and avoid that items are created with problematic labels (requiring correction later). The second name format could still be used as alias or with a P1932 qualifier.

Hello! I uploaded again, but not good. What's wrong? I modified the format and header based on the description.

id,P97,name,desc,born,died

26591,,ABAFFY Ágost Sándor abafalvi és felsőlehotai ,,1830-09-12,1907-04-14

4261,Q165503,ABELE Albert Lajos József Mária lilienbergi br,,1857-03-1,

Thank you! Palotabarát (talk) 00:16, 19 February 2021 (UTC)

type is missing (Q5)

Given the format of the names, it seems unlikely that much will automatically in MxM, e.g. second record with Albert Abele von und zu Lilienberg (Q12784118).

You might want to go ahead and create items (with normally cased names: Abaffy, Abele) and then merge applicable ones that appear on Wikidata:Database_reports/identical_birth_and_death_dates/1

Dear Jura1! Unfortunately, I can't correct the names. There are several different data types in a cell, and I can’t fix them in 39,000 cells individually. That's how I got it ready. In the example:

ABELE Albert Lajos József Mária lilienbergi br

Abele = family name (P734) (really bad spelling, not ABELE, but Abele is correct)

Albert Lajos József Mária = given name (P735)

lilienbergi = noble forename (no Wikidata property)

br = noble title (P97) (abbreviated in Hungarian = báró baron (Q165503))

In this, for example, the German and Hungarian spellings are different, in the Hungarian the noble first name is written with lowercase letter, in German it's with capital letter.

You say if I put another column with type = Q5 (for each item), it recognizes the dates and P97?

Maybe Magnus has advice that suits you better

Hi (sorry for my english),

I don't know how to use github, so I write you here.

Can you check why we can't refresh Wikidata:WikiProject Association football/reports/P641 only?

Thanks you!

"SPARQL=" needed to be changed to "sparql="

Thanks!

https://wikidata-todo.toolforge.org/quick_statements.php is down for a couple of days.

Would you kindly re-start its server?

(BTW its breakdown helped get another problem fixed)