User talk:ChristieBot

Archives

1  · 2

Inaccurate GA's reviewed count

Hello,

I believe I was told it was fine to make a manual update to User:GA bot/Stats before, but that no longer seems the case - the Bot reverted me here. I'm not sure how ChristieBot decides who the reviewer was, but I'm guessing it's whoever started the subpage. However, that isn't quite correct for Talk:Pitfall!/GA1, where the original reviewer just abandoned it. I guess that by the letter of the law, it could have been a procedural fail followed by an immediate renomination, but that was skipped here in the name of NOTBURO and not having a misleading "failed" nomination in the history through no fault of the nominator. (Checking the history... I guess it updates the bot updates the counts when a review is claimed, not completed? It looks like Indrian did get a credit around 11 June. I'd definitely assumed it'd be on completion, myself.) With the 2023 GA changes to the rules, the GA reviewed count is actually kinda important now. Would it be possible to update the numbers to include my review here? SnowFire (talk) 03:56, 30 October 2023 (UTC)

I'm afraid the numbers reflect what's in the database, and as you've guessed that is determined by the creator of the review page. There has been periodic discussion about creating |nominator= and |reviewer= parameters for {{article history}} and/or {{GA}}, which would allow this to be resolved, but there has never been support for that. Perhaps it would be best to fail and renominate in these cases; an editor's failed GAs aren't easily visible as the nominator is not recorded in the database so there's little stain on the nominator's reputation. Mike Christie (talk - contribs - library) 10:37, 30 October 2023 (UTC)

Well, the issue is that it's obviously too late in this particular circumstance to do such a quick-fail. And I'm not sure it's a good idea anyway - the tech should adapt to reviewers, not the other way around. If it is truly a hard requirement, then I think this needs to go in the GA instructions somewhere that reviews must be performed by whoever started the GA review page. Although really, fixing the tech is still preferable.

I looked at the code snippets from the user page. Does the Bot "reconstruct history" from the ground up, or are there reviewer totals in the DB that can just be incremented with deltas of recent changes? If the latter, then can you just increment the count? It seems like the kind of admin function that would be useful to have anyway. This case will almost assuredly happen again.

I'm aware we're all volunteers here so I don't want to be too demanding, and I'm also aware that maybe this wouldn't be a big deal for most GA users who do quick reviews on short topics. I try to provide somewhat in-depth reviews and also tend to nominate topics that are very long, so being short on the new QPQ system is a lot more significant; it's not just a vanity concern anymore (I wouldn't have cared under the old system). SnowFire (talk) 08:42, 1 November 2023 (UTC)

Yes, I could just modify the internal database -- the numbers were reconstructed initially, but I could edit the output. I agree that the tech should adapt to the needs, not the other way round. I would want to see a consensus that this should be done, and one thing I think would have to be settled is what is acceptable as authority to change the attribution of a review. Can I just put in a delta because an editor requests it? Or does someone need to verify it first? And I think there would have to be an audit trail of some kind -- if someone says "why is X credited with N reviews" the database should contain the answer to that question. So I don't think it's as simple as just editing a number. I should also say that though it's clear the new sort order has had an effect, it hasn't had as big an effect as you might think --- old nominations by nominators who rarely review still get reviewed; it's not like a poor ratio leaves you languishing unreviewed for ever. And there are other things about the current system that are less than perfectly accurate already -- for example, anyone who responds to a "second opinion" request gets no credit for it in the sort order. Anyway, I would suggest starting a conversation at WT:GAN about this -- it's come up before and that's probably the best place to have this discussion. Mike Christie (talk - contribs - library) 09:47, 1 November 2023 (UTC)

Personally, my suggestion would be to just accept any requests you get and assume good faith (including mine and the requests above), and if something really wacky is going on then to raise the issue at WT:GAN for whatever the problem seems to be. If someone exploits this to get "extra" GA reviews under their belt, shame on them, but the harm isn't that huge, and it'd be rapidly obvious after you get asked the second or third time that something is up.

I think it's more about the new social expectation. Doug Coldwell was being a huge leech by nom'ing tons of GAs while being too incompetent to do any reviews. If I see someone nominating a big long article and not having enough reviews under their belt, I'd definitely be very uninterested in doing the review. Maybe the articles still get reviewed, sure, but I'd want to attract high-quality reviewers to my noms, whom I suspect would care more about meeting QPQ expectations. (And I realize after writing this that this makes me come across as some sort of GA megalomaniac despite having fairly few nom's and review's, but... so it is.) SnowFire (talk) 23:46, 1 November 2023 (UTC)

Before this gets archived, can I again request the GA count be updated? Nobody really responded at Wikipedia_talk:Good_article_nominations/Archive_30#Troubleshooting_review_count_mismatches, but I'd argue that fixing this should be uncontroversial - the review actually happened. SnowFire (talk) 06:54, 28 December 2023 (UTC)

I'm reluctant to make a change like this without a consensus that it should be done. There's definitely no consensus (and in fact some minor opposition has been expressed in the past) to handling this by recording the nominator(s) with parameters on the GA template. The only other option is ad hoc changes, which would take fewer code changes and would be easier to do, but for which I feel it would be necessary for the community to agree to rules for when the numbers can be changed. I don't feel I can unilaterally make these changes; sorry. Mike Christie (talk - contribs - library) 11:24, 28 December 2023 (UTC)

Drive-by removal notifications

Hi Mike, happy 2024! In Special:diff/1197628592 Christiebot made a talkpage notification of a failed GA when I removed a template as a drive-by nomination. Is this intended behaviour? I wouldn't say a reversion is technically a fail, at if it was I would suggest an exception here as these are often from new editors and slapping them with a big you failed message seems bitey. Best, CMD (talk) 08:47, 21 January 2024 (UTC)

I'm not where I can look at the code at the moment, but as I recall, once the bot sees a review has been opened, it has to figure out what happened when the review template disappears. It's the creation of the sub review page that causes the problem -- if nobody gets around to doing that just removing the nomination takes it off the GAN page. With the review page having been created, as was the case here, the bot won't think it's a pass (it looks for the {{GA}} template to establish that) so it must be failed. The only other approach I can think of is that it could look to see that the review page no longer exists. Do you think that would be a reliable test for "this review shouldn't count"? I do have a "superseded" status I use, but at the moment it only gets used once the second review for that number is created. E.g. if the review page for nomination 1 is created, then deleted, so the nomination template still has a "1", when the review page is recreated the old one gets tagged as superseded. So that half of the process is OK. The hard part is how the bot can tell at the time the nomination template is removed that the nomination is going to be superseded. Is the deletion of the subpage enough to say that for sure? Mike Christie (talk - contribs - library) 09:42, 21 January 2024 (UTC)

Deletion of the subpage wouldn't have helped here as they were deleted after the notification went out. If this is isolated to the specific edge case of page creation that's less urgent than my initial concern. It does happen sometimes, but usually the user will have received separate advice at some point. Best, CMD (talk) 01:21, 22 January 2024 (UTC)

Enhancement request for ChristieBot, sir:

When the bot goes to an article talk page to add the edit "Transcluding GA review", may it also check for older, out of date transclusions on the same page and remove them? Otherwise, two GA reviews might appear on the talk page at the same time, when certainly only the most recent one is relevant. (Even Lowercase sigmabot III will archive everything--except the transcluded, irrelevant GA review.) For an example, see the Revision history of Talk:Chris Rock–Will Smith slapping incident. Thank you for considering this. Prhartcom (talk) 12:02, 5 February 2024 (UTC)

Seems reasonable, though I wonder if some editors might disagree -- e.g. if a failed GA is still transcluded, should that be left for the new reviewer to looks at? This is not something I can do soon, but will add it to the list. I'll probably check at WT:GAN to be sure everyone agrees it's a good idea, if/when I get to it. Mike Christie (talk - contribs - library) 12:40, 5 February 2024 (UTC)

And one more: When ChristieBot adds the GA icon, may it not mark it as a "minor" edit? It is a monumental edit! Prhartcom (talk) 12:31, 5 February 2024 (UTC)

That should be easy; will try to get to it in the next couple of weeks. Mike Christie (talk - contribs - library) 12:40, 5 February 2024 (UTC)

Thank you, Mike; and may I again say it has been good since ChristieBot took over managing the GA pages. Prhartcom (talk) 22:29, 5 February 2024 (UTC)

Bug report: order of items

The bot puts {{Good article}} as a first line of an article; still, according to MOS:LAYOUT, "{{Good article}}" should be after {{Short description}}, DISPLAYTITLE and Hatnotes. Please consider improving the bot the following way: finding an offset from the start of the article, i.e. finding a value GA_OFFSET which is a max value of the four of the following:

1. 0
2. {{Short description}}
3. DISPLAYTITLE
4. Hatnotes

and putting the {{Good article}} not at the beginning but a position of that value GA_OFFSET. You can use the other way to improve the bot to be compliant with the MOS:LAYOUT. Thank you! Maxim Masiutin (talk) 19:12, 3 April 2024 (UTC)

Hi -- thanks; I agree this should be fixed but I don't know when I'll get to it. I've been working on the bot this week so I might be able to fit it in. Mike Christie (talk - contribs - library) 22:28, 3 April 2024 (UTC)

Do you have the bot on Github? Maybe other people could contribute. Maxim Masiutin (talk) 10:21, 4 April 2024 (UTC)

I don't; I've never used it though I should probably figure out how. Until I do I'm afraid it's just me maintaining the code. Mike Christie (talk - contribs - library) 10:26, 4 April 2024 (UTC)

The citation bot is developed this way: https://github.com/ms609/citation-bot

However, if you manage to maintain alone, than it is good and no code sharing is useful.

Probably, at this point, just putting the "good article" after "short description" and "DISPLAYTITLE" should cover most cases. Maxim Masiutin (talk) 11:51, 4 April 2024 (UTC)

Maxim Masiutin: it should now be fixed for those two cases. I haven't come up with a reliable way to identify the hatnotes so the template will still be inserted before those where they exist. Mike Christie (talk - contribs - library) 14:11, 10 April 2024 (UTC)

@Mike Christie Short article and display title probably cover 95 percent cases or more, Thank you! Maxim Masiutin (talk) 16:56, 10 April 2024 (UTC)

After user space renaming

I can find my reviews GA stats after I was renamed. What do I do since the bot has changed the my names at where it do be; the question is why is my stat showing wrongly? Safari Scribe^{Edits! Talk!} 07:56, 4 April 2024 (UTC)

When you change your username the bot will follow a redirect from the old username if there is one; there is not in this case so the bot didn't know that Otuọcha was now SafariScribe. Your nominations have been edited to correct the problem -- e..g BlueMoonset fixed your nomination of Lalita Tademy. The review statistics are wrong because the bot still doesn't know that Otuọcha's statistics should now be attributed to SafariScribe. To fix that you just need to let me know, which you've done. I'll add your usernames to the list the bot keeps of user name changes and it should be corrected the next time the bot runs. Mike Christie (talk - contribs - library) 09:54, 4 April 2024 (UTC)

GAs nominations count

My issue is pretty simple, I have brought 5 articles to GA and the bot is only counting 4. What can I do about it? Thanks. The Blue Rider 14:20, 30 April 2024 (UTC)

See here -- the bot thinks you've made six nominations. For the GAN page it only counts successful nominations, so four seems to be the right number. Is there another nomination the bot is missing? Mike Christie (talk - contribs - library) 15:01, 30 April 2024 (UTC)

Thank you for the reply, yes it is missing Resplendent quetzal. The Blue Rider 18:34, 30 April 2024 (UTC)

Oh, now that I see, I made the nomination with a previous username, maybe it is that. The Blue Rider 18:36, 30 April 2024 (UTC)

Yes, that's under your old username. I assume you'd like to have that credited to you instead? I'll connect the two names in the database this evening and leave you a note when it's done. Mike Christie (talk - contribs - library) 19:27, 30 April 2024 (UTC)