User talk:CJMbot

From Wikidata
Jump to navigation Jump to search

No test edits on actual item pages, please

[edit]

Hey CJMbot, it seems that you are using Jef Wauters (Q2155599) as a test item for your editing. The revision history of that item is meanwhile pretty messy. Can you please not use an actual item for testing purposes? There is a sandbox item for testing at Wikidata Sandbox 1 (Q4115189); otherwise, you can also try things in test.wikidata.org, although this is not very convenient. Thank you, MisterSynergy (talk) 19:59, 14 November 2020 (UTC)[reply]

Got it, thanks for your suggestion! Meemoo BE (talk) 16:59, 10 December 2020 (UTC)[reply]

Care to explain this edit history, and why it happened more than a month after @MisterSynergy: first noted the issue and started this conversation? Roger Hespel (Q21645578) was the most-edited item in December and the instance of (P31) on it has a cool 500+ references, all identical. --Matthias Winkelmann (talk) 23:59, 12 January 2021 (UTC)[reply]

Article <-> Human

[edit]

Hello, please take a look to the edit. It is looked as bug. — Ivan A. Krestinin (talk) 08:07, 3 December 2022 (UTC)[reply]

Indeed, this was a human error that was written to Wikidata by a bot. Thanks for signalling and correcting it! Beireke1 (talk) 19:18, 9 January 2023 (UTC)[reply]

This edit seems like it might be an error of some kind; I cannot find the claim that the subject was born in 1905 (which seems very incorrect) anywhere in the URL cited as a source, including the page source and JSON manifest. Am I missing something? Thanks, Jamie7687 (talk) 14:26, 24 May 2024 (UTC)[reply]

I have similar concerns about recent edits by CJMbot to Bartholomeus van der Helst (Q380704) and Willem van der Borcht (Q3568363); there could be more, but these cause at least one date of birth (P569) claim value to be a date later than at least one date of death (P570) value for these items. Jamie7687 (talk) 14:36, 24 May 2024 (UTC)[reply]
Thanks @Jamie7687. I checked it and there was indeed a mistake in the preparation of the data, not in the bot itself. I will remove these statements and double check the initial dataset on more potential inconsistencies. Beireke1 (talk) 13:24, 13 June 2024 (UTC)[reply]

Linking artists to chemical compounds

[edit]

@Beireke1 See w:Special:GoToComment/c-Fram-20241004101200-Aaron_Liu-20241003201200. For example, the bot added information that matches George J. Zobel to Q82151957, 4-[2-[[4-amino-5-[4-(difluoromethoxy)phenyl]-1,2,4-triazol-3-yl]sulfanyl]acetyl]-1,3-dihydroquinoxalin-2-one. Aaron Liu (talk) 13:58, 4 October 2024 (UTC)[reply]

@Aaron Liu Thanks for notifying. This indeed not the desired result. Luckily I have access to the source data to track down how this could happen. It seems that the correct item to add this information to should have been Victor Jacques Bertauts (Q82151959). Instead of copying and pasting that item number into several rows of data, the person doing the preparation in a spreadsheet made the mistake of filling down these rows of data by dragging up the bottom right corner of the field. I'll check and revise the possible corrupted items. We're also working on an update of the bot. By using OAuth authentication, users will do their edits in their own name instead of having all edits under the bot account. Beireke1 (talk) 13:33, 7 October 2024 (UTC)[reply]