@William Avery: The bot has made some incorrect links today. For example, the en article, Senecio apolobambensis has been linked to Q15595087 Senecio appendiculatus. I have found several Senecio species today and corrected them. I don't know how widespread this issue is. (And I don't understand why my username, Oronsay, doesn't show in the preview of this post.) 02:23, 1 August 2020 (UTC)Reply

Treccani mistakes

edit

Hi @William Avery:! First of all, thanks for the great work the bot has just started! I have just found a little case of error due to Treccani (there are some mistakes like this in Treccani website, unfortunately) and I have removed it; I don't know if we can use some sort of structured whitelist. In any case I will send you other Treccani-mistakes I removed. Thanks again, --Epìdosis 12:14, 5 January 2022 (UTC)Reply

Thank you. I will review and add a check this evening. I think that before adding a new id the bot should check that either the id matches the existing id (perhaps not exactly) or the "named as" value should match. The batch of 25 processed at 12:33 seems to all be OK. William Avery (talk) 12:50, 5 January 2022 (UTC)Reply
I agree, the string check will probably avoid all errors; keep me informed if you need some manual checks, of course. Thank you again, --Epìdosis 16:26, 5 January 2022 (UTC)Reply

Another case: two IDs for Francesco Adorno (Q61470049) wrongly connected by Treccani to Francesco Adorno (Q1440881); fixed now on Wikidata. --Epìdosis 09:54, 7 February 2022 (UTC)Reply

Thank you. I am going to add a couple of features to the bot:
  1. Make addition of references easier to review, either by listing them on a talk page or doing them in separate runs of the bot
  2. Check that a claim being added does not duplicate a claim already present on another item
I feel more and more that there is no substitute for review by an editor, even though the bot saves a great deal of work by finding the apparently missing claims. William Avery (talk) 11:52, 7 February 2022 (UTC)Reply
Very good. Point 2 would have avoided the addition of one of the two IDs, in fact. Let me know the page for reviews when it is ready. --Epìdosis 11:55, 7 February 2022 (UTC)Reply
Another suggestion, maybe: the value of the qualifier publication date (P577) (if present) should always be higher than the value of date of birth (P569) (if present), for obvious reasons. This would avoid another group of mistakes. --Epìdosis 11:56, 7 February 2022 (UTC)Reply

Thanks

edit

Recently I've been creating a lot of plant articles and thinking I would have to manually add them all to Wikidata. When I went to start trudging through them this bot has already done it for me, you've saved me hours :)

Thanks again

John Cummings (talk) 13:31, 16 February 2022 (UTC)Reply

Treccani mistake on homonyms

edit

Hello @William Avery,

I have to notify about a conflation made by the bot.

Here are the two entities wrongly linked (Vincenzo Miceli (Q87647568)):

These 2 Vincenzo Miceli have different birth dates and professions.

In the DBI page of a person, there's often a right column called "Altri risultati per ..." which means "Other results".

The problem is that the match is made lexicographically and not semantically. So, it's not a reliable source of linking.

I split the two entities, and ask you to check if this behaviour in your bot is just a case, or maybe there are other elements to fix.

Thanks and have a nice day! :) Luca.favorido (talk) 07:19, 21 January 2023 (UTC)Reply

Delay

edit

Would it be possible to add a slight delay to the bot when it adds sitelinks? I often find myself in edit conflicts with it, because after I create an article I always come to Wikidata and add the sitelink manually. About 5000 ms should be plenty. Thanks, Abductive (talk) 06:57, 27 January 2023 (UTC)Reply

Thanks for the suggestion, I will look into implementing this. William Avery (talk) 15:11, 27 January 2023 (UTC)Reply
You should get at least three minutes now. William Avery (talk) 15:32, 27 January 2023 (UTC)Reply
Wow, thanks! Abductive (talk) 07:36, 28 January 2023 (UTC)Reply
I think 30, or even 60, minutes would be a more appropriate delay. Editors have a lot on their minds after they have created a new article (well I do, maybe I can't speak for all of us), including checking that the article is displaying correctly, that all the links are working, that references are formatted correctly, etc. It's very frustrating for a live editor to spend time fixing issues but then to be presented with an edit conflict when they try to save because a bot has jumped in ahead of them. Junglenut (talk) 10:44, 17 April 2023 (UTC)Reply
Thank you for the suggestion. I have lengthened the wait time to 30 minutes. William Avery (talk) 18:31, 17 April 2023 (UTC)Reply
Thank you, much appreciated Junglenut (talk) 20:44, 17 April 2023 (UTC)Reply