Wikipedia:Bots/Requests for approval/ArmadilloProcessBot

From Wikipedia, the free encyclopedia

Contents

[edit] ArmadilloProcessBot

taskscontribscountlogspage movesblock userblock logflag logflag bot

Operator: ArmadilloFromHell

Manually Assisted:Yes

Programming Language(s): Pywikipedia framework

Function Summary:To scan through User pages listed in userbox/template categories - to make a list of user pages that should be edited to remove the template category/ies. See http://en.wikipedia.org/wiki/Wikipedia:Village_pump_%28policy%29#User_pages_in_template_categories

Edit period(s) : As required, probably a few times a day until the problems as cleared up, then once or twice a month.

Edit rate requested: 6 edits per minute

Already has a bot flag (Y/N):N

Function Details: The initial version of the bot will not do any user page (or article) editing, just create a list to see what the possible workload is.

[edit] Discussion

So it crawls through a list of template categories and finds user pages? That seems like a lot of page requesting. Also, where does it get the list of template categories from?Voice-of-All 19:23, 12 November 2006 (UTC)

Please Make a suerpage for your bot. — xaosflux Talk 20:36, 12 November 2006 (UTC)
Sorry - the link was wrong --ArmadilloFromHell 20:58, 12 November 2006 (UTC)
Starting at Category:User templates it would find all subcategories e.g. Category:Food user templates. In each of the categories and subcategories it would find entries beginning with User: (I'm only interested in top level User pages, as opposed to subpages) as opposed to Template: and create a unique list of those, together with a count of many times each is found. --ArmadilloFromHell 21:11, 12 November 2006 (UTC)
Im not sure you want to use python. Python has a bug with userpages. (I know I have pywiki) I would sugest AWB. Betacommand (talkcontribsBot) 13:58, 13 November 2006 (UTC)
Thanks for the suggestion, I just applied. However, I'm sure I tried it in the past and it would not work, but I'm on a new computer with a freshly installed WinXP so maybe it will be ok now. --ArmadilloFromHell 14:40, 13 November 2006 (UTC)
Is this task to "make a list", as it says above, or to actually remove the categories? If the former, I assume no bot-flag is require, and nor would it be making 6 edits per minutes (surely more like, one edit per run, if any). If the latter, please clarify. Alai 14:28, 14 November 2006 (UTC)
Initially it was only to make a list, so for that maybe it does not need bot mode, but I assumed it would have to be throttled, so it doesn't go flying through all the template categories enumerating them. Once I see how much is involved, I may decide to make it a bot to do the user page changes. --ArmadilloFromHell 20:31, 14 November 2006 (UTC)
Some throttling is probably a good idea, but "reads" are much cheaper than "writes", and I don't think it would have to be restricted to six page-fetches per minute, if that's what you're proposing. OTOH, getting the list of candidates via the database dump is also an option. Alai 01:05, 15 November 2006 (UTC)
robots.txt used to throttle the page-read speed to sixty fetches per minute, but the devs removed the throttle recently. It's still a good guideline. --ais523 11:11, 15 November 2006 (UTC)
Sounds like a good idea (and a plausible ballpark number) to me. Alai 16:58, 15 November 2006 (UTC)

I've been playing with AWB, and have found there are in fact two distinct problems, the original one (as above), where users have added templates using subst: and failed to remove the categories, and a second one - some (unknown) number of templates have been created incorrectly and do not have noinclude around the category see User:ArmadilloProcess, I suspect the number of changes required to be signicantly less than originally estimated, e.g. fixing one template, fixes all users who have used the template. Can I use User:ArmadilloProcess (and not User:ArmadilloProcessBot) to run AWB to fix the templates without getting bot approval? I've done a couple of tests and changed one or two - see for example [1] --ArmadilloFromHell 04:19, 16 November 2006 (UTC)

Based on some sampling - only about 0.3% of templates will need changing. --ArmadilloFromHell 07:14, 16 November 2006 (UTC)
It's my understanding that yes, you can do that without prior approval, so long as it's running unflagged, and you're manually checking each edit. Alai 20:39, 20 November 2006 (UTC)
OK, I need to do some tests first, and see how much is is involved, then I'll decide what to do with this request, it may turn out not to be needed and be a manual process. --ArmadilloFromHell 14:56, 21 November 2006 (UTC)