Wikipedia:Bots/Requests for approval/BU RoBOT 10


 * The following discussion is an archived debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA. The result of the discussion was Symbol keep vote.svg Approved

BU RoBOT 10
Operator:

Time filed: 23:29, Saturday, April 2, 2016 (UTC)

Automatic, Supervised, or Manual: Automatic

Programming language(s): AWB

Source code available: AWB

Function overview: Tag articles identified by WikiProject Women as being under their umbrella with WikiProject Women, assuming more specific project templates aren't already on the page.

Links to relevant discussions (where appropriate):
 * Wikipedia_talk:WikiProject_Women/Archive_6
 * Bot_requests

Edit period(s): Multiple runs based on lists generated and approved by the WikiProject

Estimated number of pages affected: First list currently has 20,191 12,174 articles. See User:Edgars2007/Women_tag/Women.

Exclusion compliant (Yes/No): Yes

Already has a bot flag (Yes/No): Yes

Function details: WikiProject Women is looking to automatically tag articles that fall under their umbrella. They're currently considering different criteria that could be automatically tagged, and I'd like this task to be approved for use on whatever lists they come up with in the future as well. The first such criteria they've agreed upon was "all articles which have an identically named article on the German Wikipedia that has been placed in the category 'Frau' (i.e. Women)," and I think the specificity of this criteria shows that the project is exercising sufficient caution in what is automatically tagged. Articles that already have banners of more specific projects related to women will not be tagged (list found at bot request, linked above). The original request and discussion occurred a while back, so the list of articles will be regenerated before the full task is run. A trial should be fine on the existing list.
 * By the way, this runs with genfixes to handle the placement of project templates appropriately. ~ RobTalk 21:07, 3 April 2016 (UTC)

Discussion
Under the assumption that the bot checks for redirects to the banners that it skips: —  Earwig   talk  20:55, 3 April 2016 (UTC)
 * Yes, it checks for redirects. Thanks, will run this later today. ~ RobTalk 21:04, 3 April 2016 (UTC)
 * Please note that I updated the number of pages above to a substantially higher number. I miscounted before due to an apparent error/limit in AWB. If you use "Links on page (only bluelinks)", it will only populate the list with up to 5,000 links at once. ~ RobTalk 21:27, 3 April 2016 (UTC)

Alright, thanks. Let me test some more stuff in my userspace and then I'll do the trial. I'm going to have to separate this out into cases, I think, and I'll include at least a few edits of each case in the trial. ~ RobTalk 08:55, 5 April 2016 (UTC)
 * Stopped the trial after a single edit because the bot choked on something it shouldn't have. This is being handled with the "prepend text" option in AWB with genfixes used to place the template where it really belongs on the page. Unfortunately, for some reason, AWB is not throwing this template within WikiProjectBannerShell when it exists on the page. I double checked that genfixes are applied after "prepend text", and they are. See ., could you comment on why AWB isn't treating this like a WikiProject template? ~ RobTalk 21:42, 3 April 2016 (UTC)
 * Hmm... Could you do the auto-assessment, too? —  Earwig   talk 21:48, 3 April 2016 (UTC)
 * Not sure what "auto-assessment" is? I've done some more testing and the "prepend text" definitely isn't the problem. When I fed the page through AWB a second time with just genfixes enabled, it returned no changes. I appear to have encountered an unfortunate bug. ~ RobTalk 21:52, 3 April 2016 (UTC)
 * Auto-assessment would typically be copying the class specified in other project banners if they're all identical, as most WikiProjects use the same classification system based on WP 1.0's guidelines. In, it could be identified as a stub. —  Earwig   talk 21:58, 3 April 2016 (UTC)
 * I'm 99% sure that AWB genfixes will handle the auto-assessment once we sort out why this template isn't being detected as a WikiProject template. ~ RobTalk 22:07, 3 April 2016 (UTC)
 * According to AutoWikiBrowser/Order of procedures, talk page general fixes come before prepending text, which is consistent with . Also, AWB genfixes do not include auto assessment.  GoingBatty (talk) 03:04, 5 April 2016 (UTC)
 * I've split this into four cases. The first three cases (WikiProjectBannerShell on page, WikiProject template on page without BannerShell, and no WikiProject templates on page) are handled by regex and include auto-assessment if the WikiProject templates are of the form "WikiProject foo". It will not auto-assess if a redirect such as "WP foo" is used. It also will not auto-assess if an unexpected value is found in the class parameter. The last case is when no talk page exists. That will be handled with the prepend text separately. It appears a very small number of pages will fall under that case (i.e. this bot will not mass-create talk pages). ~ RobTalk 09:39, 5 April 2016 (UTC)

Is this request still active? I am confused. I recall that I have started it at some point but then I stopped for some reason. -- Magioladitis (talk) 10:04, 5 April 2016 (UTC)
 * It is still active, yes. If I recall correctly, you mentioned in the request that you never started because there wasn't consensus for the women's history aspect. This request omits the history stuff and just completes the part with consensus. ~ RobTalk 10:51, 5 April 2016 (UTC)

I updated the WikiProjects module and bypassed all redirects in question. -- Magioladitis (talk) 13:48, 5 April 2016 (UTC)


 * Here's the 50 edits. I was under the impression genfixes placed WP banners at the tops of talk pages whether or not a BannerShell was there, but it only does that as part of the BannerShell stuff. I've corrected the regex to take this into account. By default, the banners will now be placed at the top of the page instead of the bottom (which you'll see a couple times in the edits - lazy regex where I thought genfixes would pick up the slack). There's one really weird edit, where multiple WP banners were pulled out of a BannerShell: . This seems like a genfix oddity rather than anything I've done. I think a small extended trial would be good. ~ RobTalk 18:13, 5 April 2016 (UTC)
 * AWB's WikiProjectBannerShell fixes only work with templates that begin with "WikiProject" and not with their redirects that start with "WP". You might want to load User:Magioladitis/WikiProjects as a custom module to replace the redirects and make sure you're using the latest version of the AWB SVN.  GoingBatty (talk) 02:52, 6 April 2016 (UTC)
 * I've added the module. What's the benefit of the latest SVN? I've always gone with the release version as the latest stable version. Stability is extremely important for a bot run, as an error could potentially impact thousands of talk pages. ~ RobTalk 11:03, 6 April 2016 (UTC)
 * There were some changes in talk page general fixes since the latest stable release that could be crucial. -- Magioladitis (talk) 11:22, 6 April 2016 (UTC)
 * You would consider the latest SVN to be stable enough for a bot run? If so, I'll go with that. ~ RobTalk 11:36, 6 April 2016 (UTC)
 * WikiProjectBannerShell was recently changed to WikiProject banner shell - the module and SVN account for that, but there hasn't been a release version for that yet. GoingBatty (talk) 01:40, 7 April 2016 (UTC)


 * BAGAssistanceNeeded Need an extended trial on this one using the latest SVN and module and with some slightly tweaked regex. ~ RobTalk 21:26, 14 April 2016 (UTC)

I'm assuming the mistake was, which it looks like you cleaned up. Though, it does look like another bot might have already done this task (or possibly portions of it). Be sure to take this into account. -- slakr \ talk / 03:53, 16 April 2016 (UTC)


 * As per the above. I'm beyond befuddled that another bot operator said he'd take up the task, dropped it without comment for over a month, then completed the task out of nowhere without communicating this to me after seeing that I had filed this BRFA. This was a colossal waste of everyone's time. ~ RobTalk 04:22, 16 April 2016 (UTC)

Rob, Slakr I did at some point a part of this task because of the WPBS issue that works correctly only in latest SVN of AWB and not in the stable release (A new release is up this week though), that the list of redirects was not correct. I stopped after realising the list of pages was not entire clean and I sensed they will be troubles again. -- Magioladitis (talk) 07:06, 18 April 2016 (UTC)

Magioladitis (talk) 13:47, 18 April 2016 (UTC)


 * If there's still work to be done, this can be kept open. I checked a good 50 pages and found zero without the project template, plus the bot's contributions showed thousands done. Can you clarify what you mean by "not entirely clean" before I proceed with the extended trial on the latest SVN? ~ RobTalk 15:28, 18 April 2016 (UTC)
 * Rob Well, not anymore :( After you said you withdraw I went forward and tagged the 12,000 pages with Yobot. Next time we 'll cooperate better since we need more tagging bots. This is for sure. -- Magioladitis (talk) 20:36, 18 April 2016 (UTC)
 * You think, there isn't anything to tag anymore (for WP:Women)? You're wrong (I assume) ;) The updated list (I assume, Marios didn't used it). And we have enwiki categories also to tag, so ... -- Edgars2007  (talk/contribs) 20:57, 18 April 2016 (UTC)
 * I'll see about running a trial on that list in a bit. The problem here wasn't really our coordination; it was the bot approvals process. If this hadn't sat here for weeks waiting on approval, the task would have been done ages ago. There should absolutely be a gateway for people to operate bots on the project, but as it stands, bot approvals are a net negative for the project. Simple tasks done by competent bot operators are being held up for weeks or even outright declined (see the direction of my 11th BRFA...). Meanwhile, those who don't bother going through the process can get a lot of productive work done while the BAG turns a blind eye. See Bots/Requests for approval/wargo32.exe, where I brought up the fact that the bot op was operating the bot unapproved and was summarily ignored. At this point, I'd rather sit at my computer clicking away at the Save button in AWB than go through the ordeal of the approvals process unless the bot is making north of 5,000 edits, because it's faster than trying to get the fully automated version through a BRFA. I have no idea when the BAG moved away from being a safety check that made sure bots weren't destroying anything or doing things against consensus into being an online incarnation of the DMV. The whole process fails WP:NOTBUREAUCRACY at this point. ~ RobTalk 21:46, 18 April 2016 (UTC)

Is it OK if I go and merge all undone pages in one single page by removing also all pages taht are tagged wiht a more specific banner? -- Magioladitis (talk) 22:29, 21 April 2016 (UTC)
 * After some additional discussion at WP:WikiProject Women, the project has decided to go with a more restricted list to address concerns that the current one is a bit noisy (I think due to redirects, but I didn't compile the original list myself, so I don't know exactly how). I'm holding off on a trial for a few days until they have time to look over the new list, and then I'll post it here and run the trial. The new list already removes ones with more specific templates on it, I believe (although I still will keep the code to skip articles with more specific banners just in case). ~ RobTalk 22:39, 21 April 2016 (UTC)
 * This is the problem with this bot request from the beginning. Bot requests on tagging should be very specific otherwise there will be complains. Thanks. When you are ready please show the final list and the number of pages ot be tagged. -- Magioladitis (talk) 06:23, 22 April 2016 (UTC)
 * The process was followed correctly here. There was a category of pages discussed for tagging at the WikiProject, an editor requested that specific list to be tagged at Bot Requests, and I filed the request with that list as the basis. Since BRFAs are meant to address technical issues (not consensus-gathering) and the WikiProject indicated they had future tagging processes coming down the pipeline, I requested approval for tagging of WikiProject Women pages more generally (similar to how your own BRFA for Yobot allows blanket tagging rather than a specific list, although I constrained my BRFA to a single WikiProject so the code will not change whatsoever from run to run). The issue here has been that the WikiProject's members have changed their mind about the list several times. I was given a list with 20k articles, then one with nearly 50k articles, and now one with 3k articles (although a large part of the reason for the latest drop was filtering out the 20k you did and any that have more specific templates, which would have been skipped anyway). Unfortunately, that's well outside of my control. It's up to the WikiProject to determine what they want tagged, not the bot op, so I'm kind of at their whims there. I will post the list here when it's confirmed, probably tomorrow. ~ RobTalk 06:37, 22 April 2016 (UTC)
 * At User:Yobot you can find the rues I have formulated about WikiProject tagging. -- Magioladitis (talk)
 * Yes, those were seen prior to filing the BRFA and followed. Step 4 took long enough that the project had revisited step 3/5 before the task was completed. The list is User:Edgars2007/Women tag, and I'll be running the trial later today. There are 3,399 pages on the list, although a few will be skipped (redlinks, for instance). The WikiProject is currently discussing categories on en-wiki for tagging with auto-assessment, but this is all they've decided on for now. ~ RobTalk 15:27, 22 April 2016 (UTC)


 * Edits are here. One extremely small spacing issue at Talk:Alix Le Clerc which is fixed. No other errors that I noticed. ~ RobTalk 20:11, 22 April 2016 (UTC)


 * Spacing
 * No tag -- Magioladitis (talk) 22:25, 22 April 2016 (UTC)
 * No spacing is fixed as mentioned above (and not really a problem in that situation; it caused no problem). The no tag was skipped due to one of the skip criteria to ensure nothing is mistagged. As for why the purely cosmetic edit occurred, I double checked that I had cosmetic-only and genfix only changes set to skip, and they are. Seems like an issue with AWB. ~ RobTalk 23:10, 22 April 2016 (UTC)
 * Yes. Known issue. See -- Magioladitis (talk) 23:29, 22 April 2016 (UTC)
 * Ah. Thanks for the information. Is there any known work-around at this time? ~ RobTalk 02:03, 23 April 2016 (UTC)
 * not yet. We have first to determine if it is AWB's fault or Mediawiki. Case is under investigation. -- Magioladitis (talk) 06:35, 27 April 2016 (UTC)

as soon as you find the reason why a page with no tag was not actually tagged and ensure that banners are put in a newline. -- Magioladitis (talk) 06:55, 27 April 2016 (UTC)
 * ~ RobTalk 12:13, 27 April 2016 (UTC)

Magioladitis (talk) 07:34, 29 April 2016 (UTC)
 * The above discussion is preserved as an archive of the debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA.