Wikidata:Requests for permissions/Bot/Botcrux 10

The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.

Approved --Lymantria (talk) 18:03, 1 December 2022 (UTC)[reply]

Botcrux edit

Botcrux (talk • contribs • new items • new lexemes • SUL • Block log • User rights log • User rights • xtools)
Operator: Horcrux (talk • contribs • logs)

Task/s: Change publication date (P577) of scientific articles from "1 January YYYY" to just "YYYY".

Problem description: Currently we have a lot of wrong statements for publication date (P577) of scientific articles. The error lies in the precision used for stating the date. The correct precision should be 9 (year), while the current precision is 11 (day).

For understanding the magnitude of the problem, please compare the items having in description "01 January" (2200K+) with "02 January" (7K+).

Function details: Once selected all the instances of scholarly article (Q13442814) with some Jan 1st as publication date (P577), the bot would simply:

make edits like this one, changing all the datetimes from "+YYYY-01-01T00:00:00Z/11" to "+YYYY-00-00T00:00:00Z/9" (or "+YYYY-01-01T00:00:00Z/9", for remaining consistent with the source).
in a second moment, fix all the descriptions based on such wrong statements (example).

If possible, the bot would use QuickStatements for solving the issue in a reasonable time. --Horcrux (talk) 09:51, 5 April 2022 (UTC)[reply]

@Horcrux: This is definitely a problem that should be fixed, but what about those articles that legitimately were published on January 1? Can you filter based on how these articles were originally imported? ArthurPSmith (talk) 15:29, 5 April 2022 (UTC)[reply]
@ArthurPSmith: This chould be possible inspecting all the HTML pages looking for a match for "YYYY 1 Jan" in the header's citation. For instance, here the bot would change the date because there is no match in the citation of [1], while here the bot would keep the date because there is a match in the citation of [2]. --Horcrux (talk) 17:08, 5 April 2022 (UTC)[reply]
I noticed that we have to face a similar issue concerning precision 10 (months). For instance, please the numer of items having in description "01 August" (600K+) with "02 August" (10K+). Clearly, for January we have higher numbers because the two problems are mixed. Anyway, the solution for both is the same. --Horcrux (talk) 17:15, 5 April 2022 (UTC)[reply]
Many articles are (or were) published in paper journals with issues that came out nominally on the first of each month, or some months. So many of those could well be correct. I know where I work there are some journals that have/had 1st of month and 15th of month issues, so both the 1st and 15th would give you high counts and other dates not so much. If it really was only a monthly issue then reverting to month precision may be fine, but if the journal has multiple issues per month then losing that 1 is not a good idea. ArthurPSmith (talk) 18:12, 5 April 2022 (UTC)[reply]
@ArthurPSmith: This is indeed a good point, but if the source doesn't report the "1" either, why should we? --Horcrux (talk) 18:24, 5 April 2022 (UTC)[reply]
As long as there's some check like that I guess it's reasonable. Anyway, the January 1 ones are clearly mostly wrong. ArthurPSmith (talk) 18:31, 5 April 2022 (UTC)[reply]
@Lymantria: Sorry for mentioning you. What do we miss for the approval? --Horcrux (talk) 09:02, 28 November 2022 (UTC)[reply]
Hmmm, I previously overlooked this one, as there was no explicit support vote. But I will approved the request in a couple of days, given that no objections will be raised in the meantime. Lymantria (talk) 17:13, 28 November 2022 (UTC)[reply]

The above discussion is preserved as an archive. Please do not modify it. Subsequent comments should be made in a new section.