Wikipedia:User scripts/Requests/Archive 4

Archive 1

Archive 2

Archive 3

Archive 4

Archive 5

Archive 6

→

Archive 9

Script to help automate G5 tagging

Frequently I check special:newpagesfeed and go to 'by blocked users.' A lot of the time the blocked users are sockpuppet accounts that have made 25+ articles that would have to be manually tagged. Tagging all these articles is tedious and I frequently find myself asking sysops on IRC to run a special:nuke. I am requesting a script that could semi-automatically go through and tag G5 eligible articles from a user's contributions page or page creation logs if it has not been edited by another editor and was created after a date that the user would input. Thanks to anyone that might be able to help! Kb03 (talk) 19:28, 8 March 2019 (UTC)

@Kb03: I don't think this is a good idea for a script. For example, Wikipedia:Bots/Requests for approval/PkbwcgsBot 6 was denied because nominating pages for deletion should always be reviewed (as you say you will do) - but it also ties in with communicating with the editors, etc. I'm uncomfortable making this script unless there is a consensus that it would be permissible. Sorry, --DannyS712 (talk) 19:46, 8 March 2019 (UTC)

DannyS712, would an RfC be appropriate for this? Kb03 (talk) 19:47, 8 March 2019 (UTC)

@Kb03: Probably. Maybe even a discussion at WT:CSD or WP:AN. The actual script should be relatively straightforward to create (its a small matter of programming) but the potential consequences make me hesitant. --DannyS712 (talk) 19:48, 8 March 2019 (UTC)

DannyS712, one question. when you quote "Communicating with the editors, etc" are you saying the creator of the article or an editor that may dispute the G5? Kb03 (talk) 20:33, 8 March 2019 (UTC)

@Kb03: I believe that @Xaosflux was referring to the creator. And yes, I understand that this doesn't usually apply with G5, but you should still review pages individually imo. --DannyS712 (talk) 20:35, 8 March 2019 (UTC)

Yes, as such a script could be used for any reason. Perhaps getting "csd-batch" added to Twinkle may be useful, and also automate appropriate settings. Leaving a note at AN/ANI should normally be fine for a NUKE request though - in nuke situations page tagging isn't needed. — xaosflux ^Talk 20:45, 8 March 2019 (UTC)

Xaosflux, would you have an issue with me requesting the feature be added on WT:Twinkle? Kb03 (talk) 20:58, 8 March 2019 (UTC)

Kb03, I'm not sure it's a good fit for Twinkle. Mass-G5 tagging isn't particularly common to most users, and isn't something I think we should suggest is readily encouraged. Mass-tagging is just mass-editing, I think if it's really dramatic people have used AWB, but really as Xaosflux said a post to AN/ANI would be more than appropriate; tagging is just beside the point, especially if they're already blocked. — Preceding unsigned comment added by Amorymeltzer (talk • contribs) 21:53, 8 March 2019 (UTC)

Kb03 guess it is kind of niche, there really isn't a good reason for anyone to spend time mass-tagging, since you can just ask at AN. — xaosflux ^Talk 22:48, 8 March 2019 (UTC)

Xaosflux, Could TW then have a function that automated posting to WP:AN for that reason? RhinosF1^(chat)_{(status)(contribs)} 07:51, 9 March 2019 (UTC)

@RhinosF1: That, on the other hand, is a simple script to create - its just new section: please new pages created by XYZ. I don't think it should be included in twinkle, because of how rare it is; ping @Amorymeltzer to chime in about the need for twinkle vs standalone script. --DannyS712 (talk) 07:54, 9 March 2019 (UTC)

I can have a script in beta within a few days --DannyS712 (talk) 07:55, 9 March 2019 (UTC)

DannyS712, What about mass protection requests? RhinosF1^(chat)_{(status)(contribs)} 08:04, 9 March 2019 (UTC)

@RhinosF1: That would be harder, because RFPP works differently, and has a specific format. AN is just &section=new and explaining the issue --DannyS712 (talk) 08:09, 9 March 2019 (UTC)

DannyS712, Amorymeltzer might be able to advise as Twinkle can do one at a time. Can I also suggest criteria like U5/G11 as a potential for your script? That gets quite quite a few being done at a time then the user never editing again. RhinosF1^(chat)_{(status)(contribs)} 08:12, 9 March 2019 (UTC)

@RhinosF1: let me start with just requesting a nuke, a build from there. I have to go, but I'll get back to you with a draft version soon --DannyS712 (talk) 08:23, 9 March 2019 (UTC)

@RhinosF1: Current script at User:DannyS712 test/nuke.js makes edits like this (I just need to change the target page to AN, but until then it targets my sandbox so feel free to test it). Thoughts? --DannyS712 (talk) 04:24, 10 March 2019 (UTC)

Extended discussion moved to User talk:DannyS712 test/nuke.js --DannyS712 (talk) 08:04, 10 March 2019 (UTC)

ToBeFree, weren't you redesigning WP:RfPP RhinosF1^(chat)_{(status)(contribs)} 18:19, 9 March 2019 (UTC)

DannyS712, that would be a good alternative. I see no problems with it. Kb03 (talk) 21:08, 9 March 2019 (UTC)

I have tried to do the same now. Because G5 only applies to pages "that have no substantial edits by others", I had to manually check the substantiality of all edits not made by the blocked user. I also had to verify that the sockpuppeteer had been blocked before the article has been created by a now-confirmed puppet. I'm not sure if a script can safely automate what I just did. If may be useful for easier cases, though.

@RhinosF1: yes; we're currently waiting for the bot to be updated. ~ ToBeFree (talk) 03:25, 10 March 2019 (UTC)

Redirect page tab behavior

Would really appreciate a script for commons.js, so that when I'm sitting on the Talk page of a Redirect page, and click the "Article" tab at the top of the page, it just takes me back to the Redirect article, and not to the target of it. Full details on this page at WP:VPP. Thanks! Mathglot (talk) 06:22, 9 May 2019 (UTC)

@Mathglot: I'm out of town currently, but I want to work on this. I should have something for you this weekend --DannyS712 (talk) 06:26, 9 May 2019 (UTC)

Wow, thanks! Do you need any additional information from me? Let me know. Mathglot (talk) 06:38, 9 May 2019 (UTC)

Paging User:Brion VIBBER, who may be interested in following this topic as well. Mathglot (talk) 06:49, 9 May 2019 (UTC)

@Mathglot: Can be done using a single line of JS. Just add the following to your common.js file: $("#p-namespaces a")[0].href += '?redirect=no'; This makes the subject page button not follow the redirect. Works for the vector skin, not sure about others. SD0001 (talk) 17:50, 9 May 2019 (UTC)

@SD0001: That seems to work fine, thanks! Also works for non-existent Project Talk pages (e.g., the TP belonging to Wikipedia:WPWO.) I wonder if there's a repository with links to helpful scripts like this one somewhere. Probably would make a lot of editors happy to customize their experience. This one immediately makes my life easier. Mathglot (talk) 19:49, 9 May 2019 (UTC)

@Mathglot and SD0001: Additional line so that the Read tab links to the redirect itself: $("#p-views a")[0].href += '?redirect=no';. It's possible that these both could be merged into one line, but I don't know JQuery or JS in general well enough to tell.

I've also note I've reopened T5324 • Tthe underlying Phabricator issue; there's more discussion here. e^πi (talk | contribs) 01:51, 10 May 2019 (UTC)

Hi, @SD0001, DannyS712, and E to the Pi times i:, Just a heads up that about some unforeseen knock-on effects that this fix may have provoked; see WP:VPT#Invalid url from Create tab. I think this is a minor enough issue that it doesn't really need to be fixed, now that I'm aware of it; but I wanted to draw your attention to it, in case you're interested. In a nutshell: if the url already has a query string, then either the positioning of the added redirect= param or its delimiter might need to be altered. Thanks, Mathglot (talk) 23:52, 10 May 2019 (UTC)

@Mathglot: Yeah, that one's probably my fault. I copied SD0001's script, but I didn't consider the effects it would have on the right tab. But it's not an unfixable problem; it just needs to be more discerning about how it modifies the URL. e^πi (talk | contribs) 01:52, 11 May 2019 (UTC)

oh sorry, didn't foresee that bug. if ($("#p-namespaces a")[0].className.includes('mw-redirect')) $("#p-namespaces a")[0].href += '?redirect=no'; Here, now the ?redirect=no will be added only if the subject page is a redirect, removing all issues. Haven't tested this since I'm on phone. Tell me if it doesn't work. SD0001 (talk) 03:38, 11 May 2019 (UTC)

@Mathglot: Just tested this and it works. If you also want the "Read" tab to not follow redirect, you need if ($("#ca-view a")[0].className.includes('mw-redirect')) $("#ca-view a")[0].href += '?redirect=no'; SD0001 (talk) 06:42, 12 May 2019 (UTC)

@SD0001: thanks. I wasn't quite clear if I need only the last one (@06:42) for Talk & Read, or both (03:38 & 06:42). Mathglot (talk) 08:01, 12 May 2019 (UTC)

Automatic American English

I want the reverse of Ohconfucious's American → British script so that it could convert British → American. IDK why it hasn't been done already. — Mr. Guye (talk) (contribs)  08:32, 26 July 2018 (UTC)

Do you have a link to the original script? --Danski454 (talk) 16:39, 3 August 2018 (UTC)

Is this the page in question? --Kailash29792 (talk) 16:49, 3 August 2018 (UTC)

I think this might be it: User:Ohconfucius/script/EngvarB.js --Danski454 (talk) 17:38, 3 August 2018 (UTC)

Danski454, Ohconfucious does not appear to have responded to a request to make a script that can convert British English to American. Are you working on it? Or does Canadian English follow the same style? --Kailash29792 (talk) 11:44, 24 August 2018 (UTC)

Kailash29792 I am not working on this script. Also, Canadian English is somewhere between British and American. Danski454 (talk) 18:47, 25 August 2018 (UTC)

I must thank Ohconfucius for making this possible, so this request may be archived. --Kailash29792 (talk) 08:54, 13 September 2018 (UTC)

No, this request should not be archived. There still is not an automatic edit script that converts British English to American English, nor does it seem like anyone is trying to create one. My request is far from fulfilled. — Mr. Guye (talk) (contribs)  01:25, 15 September 2018 (UTC)

Is anyone working on this? If not, I can take a stab at it. Warpaltarpers (talk) 16:05, 30 May 2019 (UTC)

Another editor in support of this script - ^Atsme _Talk 📧 21:09, 20 July 2019 (UTC)

Watchlist notifier: number of (automated) updates in browser page title

Automated updates of live blogs are often indicated by a number between brackets shown in the browser's page title, signifying the number of recent changes. The inboxes of most e-mail services display the same, as do Twitter pages. Would it be possible to add a gadget that allows for automated updates of Watchlists (for all projects) with such a notifier in the browser (tab) page titles as well? For me, and I guess for many other users, it would be a major improvement, not having to manually check watchlists anymore and instead being able to view live updates while working in other browser windows. I am not sure if this has been proposed before – if so, my apologies. Jürgen Eissink (talk) 14:24, 17 July 2019 (UTC).

@Jürgen Eissink: that would require the javascript to run when the window isn't in focus, and I'm not sure that is supported. But, you may want to check out User:DannyS712/Global watchlist, which shows you your watchlist from multiple sites. Hope that helps --DannyS712 (talk) 20:01, 17 July 2019 (UTC)

I didn't actually mean a Watchlist for all projects in one, but the possibility to have the proposed feature activated on the different projects (in my case mainly nl., en., de. en Commons). And it doesn't seem like DannyS712 uses automatic updates and a page title signifier, which is specifically what I am inquiring about. Thousands of pages of pages run without the window in focus, I don't see why this would be a problem for specific Wikipedia pages. Oh, I now see you are DannyS712. Thanks for your answer, anyway. Jürgen Eissink (talk) 20:25, 17 July 2019 (UTC).

@Jürgen Eissink: clarification: I'm not sure wikipedia's support for user-defined javascript (eg at your /common.js) will run when the window isn't in focus. I could be wrong though --DannyS712 (talk) 21:50, 17 July 2019 (UTC)

Someone else, who I consider an expert, gave me the impression that it would technically be easy to implement. Let's wait for other reactions. Thanks so far. Jürgen Eissink (talk) 21:58, 17 July 2019 (UTC).

@Jürgen Eissink: I ran some tests. You're right, sorry --DannyS712 (talk) 22:13, 17 July 2019 (UTC)

@DannyS712: That sounds hopeful, thanks! I myself am not enough focused on programming to grasp every possibility, so I have to rely on others in these matters. Jürgen Eissink (talk) 22:19, 17 July 2019 (UTC).

@Jürgen Eissink: User:SD0001/watchlist-update-title.js. SD0001 (talk) 18:52, 18 July 2019 (UTC)

@SD0001: Well, thank you, SD0001, but it doesn't seem to be working for me yet. I copied your script to a new .js-page in my user space, but it doesn't seem to update automatically and neither are the numbers of changed pages displayed in the page title. Am I getting it wrong? Pinging @DannyS712: also. Jürgen Eissink (talk) 20:20, 18 July 2019 (UTC).

Oh, I now see that I have to copy it to my common.js, which I just did. Will try and see again. Jürgen Eissink (talk) 20:30, 18 July 2019 (UTC).

Is this supposed to automatically update the Watchlist when working in another window? If so, how often will it update? Doesn't do anything so far. Jürgen Eissink (talk) 21:07, 18 July 2019 (UTC).

@Jürgen Eissink: It will work only if "Live updates" button on watchlist is selected. (there seems to be no preferences option to have it selected automatically - it has to be selected every time you open your watchlist, but if you wish I can tweak the script to have it selected every time the watchlist is opened). When the watchlist undergoes a live update (done by the mediawiki software, not by this script), the script just updates the page' HTML title prefixing it with a "(1)" or "(2)" or so on. This number is removed when the window comes into focus (i.e, when you switch over to the tab from another, or if you click anywhere within the window if you were already on it).

To install it across all wikis, add

mw.loader.load('//en.wikipedia.org/w/index.php?title=User:SD0001/watchlist-update-title.js&action=raw&ctype=text/javascript');

on your global.js page on meta. Or to your common.js locally for use on a specific wiki. I'd suggest doing this rather than copy-pasting the code as you've done so that any updates/improvements I may make will become available to you automatically. SD0001 (talk) 03:24, 19 July 2019 (UTC)

Okay, that makes sense and now it seems to be working, although so far I've seen updates marked (1) when there were more pages added. So the count seems to be the update and not the number of pages updated, which I'd prefer. A small issue is that when pressing 'Mark all changes as seen' that the counter again appears and will only go away after refreshing the page, which is counterproductive to the initial purpose. I'll keep testing a bit, and while improvements would be welcomed, I thank you for the code! Jürgen Eissink (talk) 11:40, 19 July 2019 (UTC).

It takes some time to get used to this new feature, but the first experiences are satisfying. Thank you once again. I will also keep an eye on DannyS712's developments. Thank you both. Jürgen Eissink (talk) 20:27, 19 July 2019 (UTC).

@Jürgen Eissink: now "Live updates" will automatically be enabled. Counter will count number of new edits rather than watchlist updates. Also fixed the issue with Mark all changes as seen button. SD0001 (talk) 19:31, 20 July 2019 (UTC)

@SD0001: that is pretty amazing and quite literally exactly what I had in mind, so I couldn't be more grateful. I hope more users will find it useful, and I am actually somewhat surprised that this is a first. Thank you again. Jürgen Eissink (talk) 19:40, 20 July 2019 (UTC). P.S. Maybe it's good to list this on Wikipedia:User scripts/List#Watchlist?

One last improvement could be, @SD0001:, that one's own edits are not counted, but I don't know if that is possible. Jürgen Eissink (talk) 21:18, 20 July 2019 (UTC).

It's possible, but isn't it better to just hide one's own edits from the watchlist? There's an option in Preferences for that. SD0001 (talk) 07:19, 21 July 2019 (UTC)

Well I do see the benefit nevertheless for people who don't have that option on. Will add it. SD0001 (talk) 08:27, 21 July 2019 (UTC)

I use the script, because it's convenient to me, but maybe on minor detail could be improved: when hovering over the page histories, those page histories are not updated – would it be possible to add this? Thanks. Jürgen Eissink (talk) 13:32, 26 July 2019 (UTC).

Convert sfn to CS1

What the title says. Leave the sfn(s) if they are botched i.e. the corresponding reference can't be tracked. ∯WBG^converse 14:43, 2 August 2019 (UTC)

WikiProject status

I just want a lightweight script that will allow me to manually fill in the parameters for {{WikiProject status}}. Pretty much exactly like User:Evad37/rater, but simpler and only for one banner template. –MJL ‐Talk‐^☖ 20:16, 9 July 2019 (UTC)

Might do, if nobody takes this before end-September. ∯WBG^converse 14:44, 2 August 2019 (UTC)

Wanted cross-wiki reporting tool

• I am writing here in the spirit of knowing if there is a possibility of having twinkle or any other tool to report users/ips on crosswiki noticeboards? on meta/local.
• Basically twinkle with local/global choice or new script to report users would be great help. thanks QueerEcofeminist ^{"cite! even if you fight"!!!} [they/them/their] 02:31, 6 August 2019 (UTC)

Template:Infobox journal filler/inserter

It would be super, super useful if we could have a one-click script that could insert a partially pre-filled {{infobox journal}} in articles, or rebuild existing ones.

Basically it would insert the default

{{Infobox journal
| title         = 
| image         = <!-- or |cover= -->
| image_size    = 
| alt           = 
| caption       = 
| former_name   = <!-- or |former_names= -->
| abbreviation  = <!-- ISO 4 abbreviation -->
| bluebook      = <!-- For law journals only -->
| mathscinet    = <!-- For the MathSciNet abbreviation IF different from ISO 4 abbreviation-->
| nlm           = <!-- For the NLM abbreviation IF different from ISO 4 abbreviation-->
| discipline    = <!-- or |subject= -->
| peer-reviewed = 
| language      = 
| editor        = <!-- or |editors= -->
| publisher     = 
| country       = 
| history       = 
| frequency     = 
| openaccess    = 
| license       = 
| impact        = 
| impact-year   = 
| ISSNlabel     = 
| ISSN          = 
| eISSN         = 
| CODEN         = 
| JSTOR         = 
| LCCN          = 
| OCLC          = 
| website       = 
| link1         = 
| link1-name    = 
| link2         = <!-- up to |link5= -->
| link2-name    = <!-- up to |link5-name= -->
}}

(and remove {{italic title}} or its redirect if present) in the lead, like so, and would take existing {{infobox journal}} found in articles and standardize them to the above layout, importing whatever parameters it found. Extra non-empty parameters could just be appended, one per line, below |link2-name=, in alphabetical order. Headbomb {t · c · p · b} 00:35, 29 July 2019 (UTC)

Additionally, it could automatically populate certain fields, like

• |frequency=, based on Category:Academic journals by publication frequency and its subcategories
• |language=, based on Category:Academic journals by language and its subcategories
• |publisher=, based on Category:Academic journals by publisher and its subcategories
• |discipline=, based on Category:Academic journals by subject area and its subcategories
• |history=, based on Category:Publications by decade of establishment / Category:Publications by decade of disestablishment and their subcategories

• |title=, could be {{PAGENAME}}, with a warning to verify the name, in case we have an infobox journal on a non-journal article (e.g. like a society that publishes the journal).
• |abbreviation= could be determined by putting {{PAGENAME}} in [1], with a warning to verify the abbreviation, especially for non-English titles, or for non-journal articles
  • [2] could be used to compute the abbreviation for non-English titles (treat multilingual as English)

Headbomb {t · c · p · b} 00:54, 29 July 2019 (UTC)

Going to @Tokenzero: who may have some insight here as well. Headbomb {t · c · p · b} 01:10, 29 July 2019 (UTC)

Still looking into it... my thoughts for now: it would probably be best to use a solid parser like mwparserfromhell: using it to do the basic reformatting of a chosen infobox would be much easier and probably always more robust than any amount of hand-crafted regexes. Looking at mw:Alternative parsers gives essentially two options: Parsoid (the parser behind the visual editor and a few other tools; seems harder to use but can run directly in browser) or python (mwparserfromhell or wikitextparser; easier for me but requires setting up some server to communicate with your browser), I'm still looking into these. Meanwhile, I also learned about TemplateData: ideally the formatting, suggested fields and their ordering could be taken from there, right?

About the UI: you're thinking of e.g. an additional menu entry "add/normalize infobox" on the article page, which would take you to the editing screen with a proposed diff? Tokenzero (talk) 22:58, 2 August 2019 (UTC)

Yeah, taken to diff window would be good yes. Headbomb {t · c · p · b} 04:13, 3 August 2019 (UTC)

I've got a template parser at User:Evad37/extra.js (parseTemplates function), if anyone wants to try using that to make a userscript. - Evad37 [talk] 03:46, 3 August 2019 (UTC)

Thanks! That should work indeed (Parsoid turns out to be overkill and as far as I see it does not allow modifying templates/finding their position with a stable or documented API).

 Partly done Here's my first attempt: importScript('User:Tokenzero/infoboxJournal.js'); // Backlink: [[User:Tokenzero/infoboxJournal.js]]. Does not pre-populate fields yet, only adds defaults, reformats, reorders, and then removes {{italic title}}. It should display an 'Infobox journal' portlet under the 'More' menu. On an edit page it just modifies the textarea, on an article view page it redirects you to the diff page. Tokenzero (talk) 20:03, 9 August 2019 (UTC)

I'll try it later tonight! Very much looking forward to it! Headbomb {t · c · p · b} 21:19, 9 August 2019 (UTC)

@Tokenzero: first thing is it should give an edit summary. Headbomb {t · c · p · b} 01:18, 10 August 2019 (UTC)

[3] add |formernames= as an alias of |former_name= Headbomb {t · c · p · b} 01:26, 10 August 2019 (UTC)

Ok, edit summary is now Add infobox journal (IJ.js) or Reformat infobox journal (IJ.js) (or do you prefer just ce? Also I don't care about the link to the script, I can remove it if preferred). Added aliases |formernames= and |formername=. New infoboxes are now inserted after all initial templates. Tokenzero (talk) 09:03, 10 August 2019 (UTC)

"Adding/standardizing infobox journal with User:Tokenzero/infoboxJournal.js" depending on if it's added, or if one already exists. The link could be an 'instruction page' on how to add/use the script, but it might be worth waiting until it's a bit more polished to have that link. Headbomb {t · c · p · b} 12:07, 10 August 2019 (UTC)

Summary changed, links to User:Tokenzero/infoboxJournal. Pre-filling should now work for |language= and |frequency=. Tokenzero (talk) 13:01, 10 August 2019 (UTC)

Looks good. Only thing that's a bit weird is that saving using the premade summary gives a warning that there is no edit summary. A few scripts seem to be doing that. I wonder why. Headbomb {t · c · p · b} 16:29, 10 August 2019 (UTC)

Fixed: 'wpAutoSummary' should have been set to md5 of empty string. If a script wants users to change the premade summary (e.g. because it's just tagging the script name), the intended behaviour is to set 'wpAutoSummary' to md5 of it. By the way, do you know any user-script that redirect to the edit or diff page? Tokenzero (talk) 18:55, 10 August 2019 (UTC)

not sure what "do you know any user-script that redirect to the edit or diff page" means, but you could take a look at WP:Citation expander. Headbomb {t · c · p · b} 19:15, 10 August 2019 (UTC)

Any script that takes you from the view page to the edit or diff page with proposed changes. I just found that tricky to do and I wonder if there's a better way :) Citation expander, as far as I see, either works directly on the edit page (so it won't show you a diff) or from a web page that submits the edit before you can see the diff (which is much easier to do). Tokenzero (talk) 19:26, 10 August 2019 (UTC)

If you use the Citations button, you're taken to a diff. Headbomb {t · c · p · b} 19:38, 10 August 2019 (UTC)

Empty parameters

Is there a script that removes empty parameters in the cite web template like "df=" that is pretty much always empty? --Mazewaxie (talk • contribs) 13:28, 26 August 2019 (UTC)

t = t.replace(/\|\s*(?:df|mf|dmy|mdy)\s*\=\s*(?=[\|\}])/g, "");

where t is the wikitext, which you've hopefully obtained in some previous step.

Note that this whacks any parameter with one of the four names above (without regard to template name or amount of spacing), just as long as it's blank.

Removal of any blank parameter (but only from specific templates) (if that's what you actually need) would be a bit more complicated to pattern-match because preceding parameter values could contain any number of nested invocations of other templates.

―cobaltcigs 06:56, 11 October 2019 (UTC)

An sidebar on the right would be nice!

Having a floating sidebar on the right of the screen would allow for quicker use of Wikipedia, and I would like this to be implemented as a script (if possible). Cheers! creativeRaja^t@lk 16:20, 14 October 2019 (UTC)

Check out the Timeless skin (in your preferences) if you're not afraid of everything else looking different as well. Can't find anything in WP:US/L about a right-hand sidebar, but I feel like I remember one existing. Enterprisey (talk!) 04:44, 16 October 2019 (UTC)

unreliable.js, a predatory journals/vanity press highlighter

Resolved

Similar to User:Anomie/linkclassifier.js/User:Anomie/linkclassifier.css, I'd like a user script to highlight citations to predatory journals and vanity presses. The regex can be stolen/adapted from Special:AbuseFilter/891.

Basically, if a URL with a domain that matches the urls := part, highlight those links by making them... let's say pink→pink/some annoying color. Likewise, if a DOI matches the dois : part, highlight those in annoying pink as well (doi:10.4172/omicsjournalofnonsense→doi:10.4172/omicsjournalofnonsense). Or whatever other colour. If the script cannot dynamically be kept in sync with Special:AbuseFilter/891, I can take over and put it at User:Headbomb/unreliable.js and maintain it after that.

The idea is that, if you encounter a citation like

• Xu, Ting (2017-01-06). "The Relationship between Interest Rates, Income, GDP Growth and House Prices". Research in Economics and Management. 2 (1): 30. doi:10.22158/rem.v2n1p30. ISSN 2470-4393.

You get presented with something like

• Xu, Ting (2017-01-06). "The Relationship between Interest Rates, Income, GDP Growth and House Prices". Research in Economics and Management. 2 (1): 30. doi:10.22158/rem.v2n1p30. ISSN 2470-4393.

And know something's off. Multiple 'classes' and colors should be supported (each customisable as enabled/disabled), so that additional things like blacklisted links and possibly unreliable sources (like Forbes.com, which will often be fine, but sometimes not) can also be highlighted. Headbomb {t · c · p · b} 07:16, 6 February 2020 (UTC)

The basic version would look like this:

var PREDATORY_HREFS_RGX = /academicjournals\.com|academicjournals\.net|academicjournals\.org|academicpub\.org|academicresearchjournals\.org|aiac\.org\.au|aicit\.org|alliedacademies\.org|ashdin\.com|aspbs\.com|avensonline\.org|biomedres\.info|biopublisher\.ca|bowenpublishing\.com|ccsenet\.org|cennser\.org|clinmedjournals\.org|cluteinstitute\.com|cpinet\.info|cscanada\.net|davidpublisher\.org|etpub\.com|eujournal\.org|growingscience\.com|hanspub\.org|hoajonline\.com|hrmars\.com|iacsit\.org|iamure\.com|\.idosi\.org|igi-global\.com|iises\.net|imedpub\.com|informaticsjournals\.com|innspub\.net|intechopen\.com|intechweb\.org|interesjournals\.org|ispacs\.com|ispub\.com|julypress\.com|juniperpublishers\.com|kowsarpub\.com|kspjournals\.org|m-hikari\.com|macrothink\.org|mathewsopenaccess\.com|mecs-press\.org|oapublishinglondon\.com|oatext\.com|omicsonline\.org|ospcindia\.org|researchleap\.com|sapub\.org|scholink\.org|scialert\.net|scidoc\.org|sciencedomain\.org|sciencedomains\.org|sciedu\.ca|sciencepg\.com|sciencepub\.net|sciencepubco\.com|sciencepublication\.org|sciencepublishinggroup\.com|scipg\.net|scipress\.com|scirp\.org|scopemed\.com|sersc\.org|sphinxsai\.com|\.ssjournals\.com|thesai\.org|waset\.org|witpress\.com|worldwidejournals\.com|zantworldpress\.com/i;

var PREDATORY_DOIS_RGX = /doi.org\/10\.(11648|1166|1234|12677|12692|12720|12988|13005|13172|13188|14257|14303|14419|1453|14569|14662|15415|15373|15761|17265|18005|18052|18311|18775|19030|19044|19070|19080|1999|20472|20849|20902|21102|21767|22158|23937|2495|30654|30845|35841|36648|3844|3923|3968|4018|4156|4172|4236|4303|5267|5296|5376|5430|5455|5539|5567|5580|5772|5812|5815|5829|5897|5899|5923|5963|6007|7243|7439|7537|7575|7718|7763|9734)/i;

var cssRules = {
	"margin-right": "0.1em",
	"padding": "3px 4px 2px",
	"background-color": "#ffdddd",
	"border": "1px",
	"border-radius": "3px",
	"box-shadow": "0.1em 0.1em 0.25em rgba(0,0,0,0.75)",
	"-moz-box-shadow": "2px 2px 4px #A0A080",
	"-webkit-box-shadow": "2px 2px 4px #A0A080",
	"box-shadow": "2px 2px 4px #A0A080"
};

$('.mw-parser-output a.external').each(function() {
	if (PREDATORY_HREFS_RGX.test(this.href)) {
		$(this).css(cssRules);
	} else if (PREDATORY_DOIS_RGX.test(this.href)) {
		$(this).css(cssRules);
	}
});

Hopefully someone else can do the rest and provide customisability you ask for.

You should definitely simplify the massive regex (for instance, make the .com part common for all .com domains) for better performance. SD0001 (talk) 12:43, 6 February 2020 (UTC)

I can regex, what I can't do is .js from scratch, but the above should give me enough to get started. I'll give that piece of code a shot this afternoon. Performance probably isn't much of an issue. Headbomb {t · c · p · b} 13:54, 6 February 2020 (UTC)

@SD0001: User:Headbomb/unreliable.js doesn't seem to work, including your original code. This can be tested on Science Publishing Group, which has both a url and doi that should match. Headbomb {t · c · p · b} 14:45, 6 February 2020 (UTC)

Nevermind, I had a brainfart. Namely, I forgot to do this. Headbomb {t · c · p · b} 16:04, 6 February 2020 (UTC)

I should have mentioned that for it to load reliably, it needs to be ensured that the page has fully loaded before the script code is run. You can do this by wrapping the whole code between (function() { ... script code comes here ... })();

I see that you removed doi.org from the DOIs regex - this would mean that anyinnocentdomain.com/3424310.11648 would be flagged, as the link's url contains "10.11648". Markup like (doi)*?[ ]*?[=\|\:][ ]*? won't work here as we're checking against the HTML of the page (specifically, against external link URLs), not against the wikitext. [4] the i makes the regex test case-insensitive. SD0001 (talk) 17:16, 6 February 2020 (UTC)

Good to know. Didn't know about /i, but that will be useful. As for removing doi.org, that's desired to pickup "Journal of crap 10.4172/asdf10354" that might not make use of templates. false positives are far and few between, although normally I might add a final ...\/ at the end to ensure it matches a 10.4172/ and not just 10.4172. Clearly "(PRED_DOIS_RGX.test(this.href))" needs tweaking to look outside hrefs, but to what, I don't yet know. Headbomb {t · c · p · b} 17:21, 6 February 2020 (UTC)

@Headbomb: You can check whole reference texts (anything within <ref> tags) by using code like:

$('span.reference-text').each(function() {
    if (DOI_RGX.test(this.textContent)) {
        $(this).css(CSSRULE);
    }
});

SD0001 (talk) 17:47, 6 February 2020 (UTC)

@SD0001: Well, I don't want this to be restricted to ref tags either. I want to find those wherever on the page they may be. Headbomb {t · c · p · b} 17:56, 6 February 2020 (UTC)

 Done Also, if something could only be detected/highlighted in specific namespaces, that would be good. For instance the current var WIKI_HREFS_RGX should really only be highlighted in main/draft spaces. Headbomb {t · c · p · b} 23:27, 6 February 2020 (UTC)

@SD0001:, any idea for how to look outside of references/links? Headbomb {t · c · p · b} 01:33, 11 February 2020 (UTC)

@Headbomb: Thanks for all your work on this so far! When finished, I do think this will be a very useful script, particularly for NPP/AFC folks.

Regarding looking outside references, it's a bit complicated. We'll need to iterate over all the text nodes on the page, extract the index at which the search hit begins, and wrap that in a span element, upon which we can then apply custom stylings. I can code this for you when I have time. But where the does the desire for doing this outside ref tags come from? SD0001 (talk) 18:02, 12 February 2020 (UTC)

@SD0001: A typical example would be in further reading section or external link sections where you could have something like

==Further reading
* Smith, J. (2019) "Article of things" ''Journal of Nonsense'' doi: 10.4172/123456798

It doesn't need to highlight the whole doi (although that would be neat if it could), but just the prefix/whatever it matches would be enough. Headbomb {t · c · p · b} 18:08, 12 February 2020 (UTC)

• https://markjs.io/ may be useful for something here. Headbomb {t · c · p · b} 19:00, 12 February 2020 (UTC)

TFD holding cell: button to mark template as ready for deletion

Moved from User talk:Evad37 § Thank you!

So I actually do have a suggestion/request. One thing that would be great is to be able to use this in the Holding cell. What I mean by that is that when a template that is currently listed in one of the holding sections is ready to be deleted, it would be great to be able to just click a button and have the entry moved to the bottom and to have the template in question updated as ready for deletion. Food for thought... --Zackmann (^{Talk to me}/_{What I been doing}) 00:24, 11 February 2019 (UTC)

Does that make any sense? Not sure if I did a great job explaining it... --Zackmann (^{Talk to me}/_{What I been doing}) 19:32, 15 February 2019 (UTC)

Yeah, something like this should be possible. I'll move this thread to WT:XFDC so this suggestion doesn't get lost in my talk page archives. - Evad37 [talk] 02:20, 16 February 2019 (UTC)

Thinking about this further, it's probably better off as a separate script – I don't think being part of XFDcloser would save a significant amount of code - Evad37 [talk] 01:06, 20 November 2019 (UTC)

Moved from WT:XFDcloser § Holding cell suggestion

Ping @Zackmann08: - Evad37 [talk] 01:12, 20 November 2019 (UTC)

Alternatively we could just stop bothering with the delete=1 business since it's completely useless anyway. ‑‑Trialpears (talk) 13:07, 3 December 2019 (UTC)

Extract data from a frozen VisualEditor

Often VisualEditor ends up frozen and won't allow me to save or switch to markup-mode, giving instead a "docserver-http-error", "500", or "404" error (when the article has been renamed while I was editing). This happens about once a day.

It seems VisualEditor's API doesn't allow for extracting the markup data without connecting to the server.

I have two ugly ideas for how to solve this:

1. Regularly log wikimarkup when using VisualEditor to prevent loss of data when it freezes

• Every 30 seconds, generate wiki-markup and console.log it.
• If the connection to the server is broken, alert() me so that I may retrieve my markup from the console and reload the editor.

2. Extract the data model

• Extract the data model, copy it to my clipboard
• Reload the window
• Paste the data model, and the user script applies it to the new instance of VisualEditor

Or if anyone knows a better way to retrieve data from a frozen VisualEditor, that'd be good to know.

With thanks, – Thjarkur (talk) 22:40, 3 August 2019 (UTC)

@Whatamidoing (WMF): Any ideas/comments? SD0001 (talk) 04:58, 21 October 2019 (UTC)

Þjarkur, the data model is actually already saved to localsession storage in your browser. If you randomly refresh the page during an editor session, you will note a small notification top right: "Changes recovered; Your unsaved changes have been automatically recovered". No idea however what happens when the page is moved during that time. It probably wasn't designed with that in mind. You might consider opening a phabricator ticket on that feature request, as it is not an uncommon situation I'd think. —TheDJ (talk • contribs) 09:22, 21 October 2019 (UTC)

SD0001, thanks for the ping. I really appreciate it when editors let me know about conversations. Thjarkur, thanks for posting about the problems you were having. I apologize for the delay in finding it. Inside the visual editor, there's a ? menu. At the end of that menu is an item for leaving feedback on the software. That posts your note over on MediaWiki.org, where the devs are likely to find it faster.

Yes, the ability to (usually) recover unsaved changes already exists. If the visual editor crashes, but your web browser doesn't, then you should be able to select-all and copy everything (or just the changes you were making) and then paste them back in. It won't work in 100% of the cases, but it usually works for me.

There was a spate of 404 problems recently. That's a separate thing (all about mw:RESTBase, and supposedly it's finally fixed now (as of last WP:THURSDAY?). Hopefully you're already seeing fewer of these problems. Whatamidoing (WMF) (talk) 17:10, 23 October 2019 (UTC)

The Restbase improvements did lessen the problem but I still run in to this way too often. The cookie-autosave is cleared if you accidentally clicked on Escape or clicked out of the VisualEditor, I think the cookie has only saved me in about 5% of cases. Does anybody know a way to interact with the VisualEditor instance in Javascript? – Thjarkur (talk) 15:51, 9 December 2019 (UTC)

PROD handler

A script to make it a bit less time consuming for patrolling admins to assess proposed deletions. Single click to open a dialog which would check if:

• the page has ever been PRODed before
• the page has ever been nominated for deletion before
• the page has ever been undeleted
• the talk page has any content other than WikiProject templates

and buttons to:

• open what links here (articles only) in a new tab or list them in the dialog
• open the talk page in a new tab or show it in the dialog
• accept the PROD and delete the page, talk page, and redirects
• decline the PROD because ineligible or for a custom reason

--kingboyk (talk) 18:18, 13 February 2020 (UTC) (ping appreciated if there is a reply)

I now recall that the PROD template itself is supposed to indicate if the article has already been at AfD.

In recent days I have had to decline PRODs due to the article having been PRODded before (this is often difficult or impossible to spot in the history) and one where the article had previously been undeleted. --kingboyk (talk) 13:19, 14 February 2020 (UTC)

Unreviewed featured articles

See Wikipedia:Unreviewed featured articles, Template:Article history, and Wikipedia:Featured article review (FAR)

It would be grand if someone were able to generate a list of current FAs ordered (oldest to newest) by the date of their last review (FAC or FAR), and including their mainpage date if they have run WP:TFA.

This will help in making sure that dated FAs are reviewed at FAR, and that FAs that have not yet run on the mainpage could be prioritized.

• Current FAs are found at Category:Featured articles
• Date of last review is found on the article talk page, in the {{Article history}} template, searching on the actionNdate for the highest N where actionN containes either FAC or FAR.
  • Return for the highest N containing FAC or FAR (whichever is latest) the date and link found in |actionNdate= and |actionNlink=
• Mainpage date is found in the article history template on the article talk page as |maindate=
• The list, tabulated from oldest date of last review to newest, would be placed at WP:URFA, looking like:

Date last FAC or FAR	TFA date	Article	Most recent FAC or FAR	Notes
2006-11-02	2020-03-03	Tourette syndrome	Wikipedia:Featured article candidates/Tourette syndrome
2020-09-26	2008-04-26	Manzanar	Wikipedia:Featured article review/Manzanar/archive1

The notes field would be left blank for reviewers to fill in: for example, we would put on Tourette syndrome that it was completely overhauled for 2020 TFA.

If it makes for less work, it's not really necessary to have FAs that passed FAC after 2017-01-01 ... the idea is to look at the very old FAs.

Thanks! SandyGeorgia (Talk) 21:53, 9 November 2020 (UTC)

@SandyGeorgia:  Done I put it at User:SDZeroBot/Featured articles. Hopefully met all your requirements :D – SD0001 (talk) 21:20, 20 November 2020 (UTC)

Totally awesome ... I am checking through it now and will get back to you. SandyGeorgia (Talk) 21:42, 20 November 2020 (UTC)

User:SD0001 One query (I can adjust this manually unless you tell me it indicates an underlying problem in the script):

1. Cretan War (205–200 BC) returned no last FAC/FAR date

And that is the only issue I uncover. I checked a couple of FAs that have convoluted histories of being featured, de-featured, and re-featured, and this looks perfect.

Would you mind moving the list to Wikipedia:Unreviewed featured articles/2020 ??

Thank you for the amazing work; I must seek out the perfect barnstar! SandyGeorgia (Talk) 22:06, 20 November 2020 (UTC)

SandyGeorgia, Indeed that indicated an underlying problem: the script wasn't configured to deal with "actionN" where N was a 2 digit number. Fixed it and re-run. Also moved it to project space. – SD0001 (talk) 22:51, 20 November 2020 (UTC)

Awesome, we are all set if anyone wants to archive this. SandyGeorgia (Talk) 23:35, 20 November 2020 (UTC)

Good article reassessment script

The process to nominate an article for good article reassessment is very time consuming and rather complicated (see the procedure laid out at Wikipedia:Good article reassessment), and I'd imagine that a script could automate this. Might be better as a twinkle feature— not sure—but here's my (with no coding experience) idea: an option to nominate it for individual or community reassessment, based upon what is selected the script will create the correct subpage. Then there could be a spot to type in the rationale and ideally the script would notify the original ga nominator and the original reviewer and any wikiprojects tagged on the talk. In my mind this would be like a combination of how twinkle handles XfD's and merge discussions, not sure how hard it would be to do, but as I anticipate nominating a lot of poor quality articles for reassessment soon, it could be very helpful. Best wishes, Eddie891 ^Talk _Work 16:31, 19 June 2020 (UTC)

Noting that this was  Done (User:SD0001/GAR-helper). See WT:TW#Good article reassessments. SD0001 (talk) 09:38, 16 July 2020 (UTC)

Userscript help

Resolved

I have imported a userscript from Meta-Wiki and have it at User:FlightTime/confirm-logout.js. I've tried to contact the author, but they seem to be inactive. The script asks for a conformation for logout, as seen here File:Logout screenshot.png. The script uses RGB color calls, What would be perfect would to change all RGB calls to Hex triplet, if that is not possible then the outcome I'm looking for is as follows:

background: #EEEEEE;

border: 2px solid #0000FF; (also the button borders)

text: #000

I am trying to match the boxes on my talk page

Any and all help will be greatly appreciated. Please ping me with comments/results. Thank you for your time, - FlightTime (open channel) 20:32, 23 February 2020 (UTC)

@FlightTime: Make the following changes:

• warning: {text: "rgba(255, 255, 255, 1)", background: "rgba(221, 51, 51, 1)"}, →
• warning: {text: "#000", background: "#EEE"},

• "en_us": "<div style='margin-bottom: 0.5em; '>You clicked on a log-out link. Do you want to continue?</div><div><button>No</button><button onclick='window.location=\"/wiki/special:logout\"'>Log out</button></div>", →
• "en_us": "<div style='margin-bottom: 0.5em; border: 2px solid #00F'>You clicked on a log-out link. Do you want to continue?</div><div><button style='border: 2px solid #00F; color: #000'>No</button><button style='border: 2px solid #00F; color: #000' onclick='window.location=\"/wiki/special:logout\"'>Log out</button></div>",

I think these will work. That whole script could do with a bit of cleanup, since you probably aren't using all of it's features. Regards, —guywan (talk • contribs) 00:23, 21 March 2020 (UTC)

@Guywan: Wow :) Thank you, I was thinking I wasn't going to get a reply. Let's see if I can do this right (if you'd like I can put the script on a subpage, not sure if the author is maintaining it). Thanx again. Cheers, — Preceding unsigned comment added by FlightTime (talk • contribs) 00:25, 21 March 2020 (UTC)

A script to help build an "open tasks" list of things you'd like to monitor

Sometimes, when I throw out an idea on a more obscure talk page, there's no one around to see it and it gets lost (or, at a super busy talk page, there are too many competing discussions and it gets buried). When I make a post I fear this might happen to, I often add the page to my watchlist so I can follow up on it or send out invites if needed later, but this isn't fully optimal, since for obscure pages I might never get reminded to follow up if there's no other activity, and for busy pages it clogs up my watchlist. I'd love to have a userscript that I could use to easily build up a list of such discussions on a subpage of my user page. I envision it working by allowing you to go into a mode on any page where you can click on a section, and it'll add a link to that discussion on your subpage. On the subpage itself, it'd allow you to easily move tasks between the open/active category and a "done" category for completed matters and a "abandoned" category for failed/not-worth-the-effort matters. Does anything like that exist, and if not would anyone be interested in taking this on? (I'm hoping this proposal doesn't itself become one of those things I forget about.) Sdkb (talk) 00:22, 29 March 2020 (UTC)

@Sdkb: Check out the scripts in WP:User scripts/List#Todo lists. I think ToDoLister does most of those things. SD0001 (talk) 08:51, 31 March 2020 (UTC)

A script to handle file renaming requests

A script to handle file renaming requests, which will do following this:

• Rename the file
• Remove the {{Rename media}} template.
• Update file links

Similar to commons:MediaWiki:Gadget-AjaxQuickDelete.js. -- CptViraj (📧) 13:58, 2 March 2020 (UTC)

@CptViraj: Very exciting! I might be able to do this. I'll start working on it some time. Regards, —guywan (talk • contribs) 23:59, 20 March 2020 (UTC)

Ye thankyou. Regards! -- CptViraj (📧) 03:33, 21 March 2020 (UTC)

I do lots of file moving, this script would be terrific. - FlightTime Phone (open channel) 16:09, 24 March 2020 (UTC)

@CptViraj and FlightTime: Unveiling FileMoverHelper. While on File: pages, it creates a link in the More portlet (if you use the Vector skin) called FMH. Click on that link to get started. I recommend checking the results the first few times and letting me know if something goes horribly terribly wrong. Best of luck. —guywan (talk • contribs) 23:05, 2 April 2020 (UTC)

@Guywan: I moved File:Trails of Cold Steel IV Japanese Cover.png to File:Trails of Cold Steel IV Cover Art.png but the script didn't update the file link at The Legend of Heroes: Trails of Cold Steel IV. And if possible can we have custom reason option too? Thankyou! -- CptViraj (📧) 03:48, 3 April 2020 (UTC)

@CptViraj: Should be fixed. You will now be prompted for a move reason. —guywan (talk • contribs) 15:26, 3 April 2020 (UTC)

@Guywan: File:Flash Comics (Feb 1948) cover art.png didn't update file link this time too. Custom reason is working fine. -- CptViraj (📧) 16:05, 3 April 2020 (UTC)

@CptViraj: Now it should be fixed, hopefully. —guywan (talk • contribs) 18:42, 3 April 2020 (UTC)

HarvErrors.js

There is a open discussion at Help_talk:Citation_Style_1/Archive 69#Cite book Harv warning which involves the HarvErrors.js script. The original script is at User:Ucucha/HarvErrors.js; patched versions of the script are available at User:Svick/HarvErrors.js and User:David Eppstein/HarvErrors.js. Nonetheless, its usefulness (both the original and the patched versions) is deemed less than it used to be before the "|ref=harv " default was applied to cs1 cite templates a few days ago. This is not an actual request yet - please discuss at Help_talk:Citation_Style_1/Archive 69#Cite book Harv warning and not here (to keep discussions in one place), until it becomes clear whether a coherently formulated script request would actually make sense in order to address the situation. --Francis Schonken (talk) 07:12, 21 April 2020 (UTC)

@Francis Schonken: The discussion about how to make a new script based on HarvErrors.js should be held here, not at Help talk:CS1 btw, because here is where you can explain what you would like the script to behave like. Headbomb {t · c · p · b} 10:28, 21 April 2020 (UTC)

Is an actual script request now (see below), so struck the recommendation to have preliminary talks elsewhere, which no longer applies. --Francis Schonken (talk) 04:50, 22 April 2020 (UTC)

Some discussion relating to this occurred at the technical Village Pump; the portion of the discussion relating to the development of a user script is being redirected here. Kees08 (Talk) 01:58, 22 April 2020 (UTC)

Kees08, we need a script that did what User:Ucucha/HarvErrors.js did before the recent change that made ref=harv the default for {{cite book}} etc. We need three things:

(a) a warning when we use a short cite (sfn/harvnb) without a corresponding long one;

(b) a warning when we use a long one as a source (i.e. with ref=harv) without a corresponding short one; and

(c) no warning when the template is in Selected works, Bibliography, or Further reading without ref=harv.

We had that functionality before the change. We need to be able to move these citation templates smoothly from sourcing sections (with ref=harv) to non-sourcing sections (ref=none). Is it possible to write a script that does this? The scripts suggested by Trappist and others don't achieve this. SarahSV ^(talk) 03:53, 22 April 2020 (UTC)

Currently, the Harverrors script does (a) with a red error message (not a warning; the two are different, and we should use words carefully when asking a programmer to create a script); does (b) with a brown warning message; and does (c) if the full citation includes |ref=none as a parameter. Those three things have always been true for Harverrors in combination with {{Citation}}, which is a Citation Style 2 template, and for {{cite book}} and the other Citation Style 1 templates when they included |mode=cs2. The recent change is that the brown warning described in (b) has been activated for {{cite book}} and the other Citation Style 1 templates, even if they do not explicitly use |ref=harv, due to a change in the CS1 templates. To be clear, the Harverrors script has not changed.

I don't know of a way that (c) could be achieved without using |ref=none, but there are a lot of clever script writers out there. Depending on the section name will not work, because section names are used inconsistently. At the GA James Francis Dwyer, for example, the section for full citations that are linked from short citations is called "Bibliography". – Jonesey95 (talk) 04:11, 22 April 2020 (UTC)

Or, a script that adds "|ref=none" to {{citation}} and/or cite templates that don't have a "|ref=harv", and/or in selected sections to all such such templates, after which clean-up of all the lists (i.e. "Cited works", "Further reading", "External links" etc) can proceed with assistance of the original HarvErrors.js script (that is the one showing "errors" as well as "warnings"). --Francis Schonken (talk) 04:50, 22 April 2020 (UTC)

c) cannot be achieved as requested because |ref=harv is equivalent to |ref=, which is how {{citation}} worked. The solution to that is to add |ref=none to those {{cite book}} and other templates, just like is done with {{citation}}. That said

• If the warning in the Selected works, Bibliography, Further reading, External links sections are not helpful, those sections could suppress the warnings messages entirely (blacklist approach)
• Alternatively, the warnings could be enabled only in specific sections (e.g. Works cited) instead (whitelist approach)
• Even section names vary, it would be easy to add variations

This would cover both CS1 and CS2 templates. Help talk:CS1#Mode marker? would allow to differentiate CS1 from CS2 templates if that was desired however. Headbomb {t · c · p · b} 05:35, 22 April 2020 (UTC)

We need to be able to move these citation templates smoothly from sourcing sections (with ref=harv) to non-sourcing sections (ref=none). Is it possible to write a script that does this?

Moving text from one section of an article to another can be done in edit-mode but that is a whole different script and operating environment from the post-rendered environment of User:Ucucha/HarvErrors.js and similar scripts. If this functionality is desired (moving text from place to place in an article) that functionality deserves its own request on this page so that this request isn't muddled with unrelated discussion.

At User talk:Trappist the monk § Ref=harv, I suggested a fourth requirement:

(d) no warnings at long-cite templates when the article has no short-cite templates

My fork of User:Ucucha/HarvErrors.js accomplishes this by muting the warning messages when there are no short-cites in the article.

—Trappist the monk (talk) 13:50, 22 April 2020 (UTC)

First, I'm a little confused about (c) "we had that functionality before the change"; is that an assertion that the Ucucha script suppresses (or suppressed before the recent change) warnings in those sections? I never noticed that, and I don't see those exclusions in the JS. I suppose it could be added, but I don't have the skilz. In any case, there are many inventive section names in footers containing citations that I've seen over the years. Not everyone has always hewn to the recommended section structure.

Second, I think the proposal to move text between sections is fraught with danger. There are many patterns of which section or sometimes multiple sections a CS1/2 reference finds itself in, especially in very old articles, which aren't particularly disciplined in their observance of the recommended section purpose and order. I think the bot would need a very clear understanding of the semantics. But I'll wait for that separate discussion and, I hope, a full specification. David Brooks (talk) 15:34, 22 April 2020 (UTC)

@DavidBrooks: "we had that functionality before the change" is a misconception, that functionality was never there. Also, bots shouldn't be doing anything here as far putting |ref=none at this point (although there could be a case made for such a task when the criteria for when |ref=none should be used is clarified), but doing this by script (again, with human-review) would likely be fine. Bots moving citations and references will also never fly, although it's possible that someone can design a script for this. This would be tricky, but is theoretically possible. Headbomb {t · c · p · b} 15:40, 22 April 2020 (UTC)

DavidBrooks, we certainly did have that functionality before the change. I relied every day on all three of the points I mentioned above. SarahSV ^(talk) 17:52, 27 April 2020 (UTC)

You did not, you have a misconception about what the script what doing. Headbomb {t · c · p · b} 18:08, 27 April 2020 (UTC)

The script told me (a) when I used a short cite (sfn/harvnb) without a corresponding long one; and (b) when I used a long one (with ref=harv) without a corresponding short one. It also (c) gave me no warning/error message when I used a citation template in Selected works, Bibliography, or Further reading without ref=harv. Now there is no script that offers (a), (b), and (c). SarahSV ^(talk) 18:41, 27 April 2020 (UTC)

b) and c) are simply incorrect. User:Ucucha/HarvErrors.js had {{citation}} threw warnings those warnings time since immemorial without having explicit |ref=harv set, including in the Further reading section. Now the CS1 templates (such as {{cite book}}) are treated on the same footing as {{citation}}. Headbomb {t · c · p · b} 19:21, 27 April 2020 (UTC)

@Trappist the monk: anyway your script could ignore specific sections from throwing warnings? Headbomb {t · c · p · b} 14:43, 24 April 2020 (UTC)

No idea. My knowledge of javascript is minimal at best. I suppose a script can be made smart enough to some how look for <span class="mw-headline" id="Further_reading"> and then ignore everything in that section until it encountered the next <hn> tag. I don't know how to do that and, frankly, little desire to learn enough js to make an attempt.

—Trappist the monk (talk) 15:10, 24 April 2020 (UTC)

@SD0001: you're good with JS, any ideas here? Headbomb {t · c · p · b} 15:11, 24 April 2020 (UTC)

I am not sure what's being requested here as I know next to nothing about CS1/2. Didn't read most of the above, but regarding Trappist the monk's last comment, it's easy to ignore stuff in a Further reading section using the jQuery. For example, if you're looking for elements with class citation (.citation), var sectionCites = $('#Further_reading').parent().nextUntil('h2').find('.citation').get(); gets you all cites in the section. If CITES is the array of all cites on the page, CITES.filter(function(x) { return sectionCites.indexOf(x) === -1; }) is the array of all cites not in that section. SD0001 (talk) 21:15, 25 April 2020 (UTC)

@SD0001: Thanks for that. The var sectionCites = ... appears to do what you say it does. But, there's always a but ... But, I don't know how to use the CITES.filter(function(x) { return sectionCites.indexOf(x) === -1; }). I have a sandbox that has seven cs1|2 templates. Four of those are in §Further reading. By concatenating sectionSites.length to the end of the script's normal output I can see that sectionSites contains four elements.

In the code, cites is the the array of all cites on the page so this should make filtercites an array of the three elements in cites that are not in sectionSites:

var filtercites = cites.filter(function(x) { return sectionCites.indexOf(x) === -1; });

Alas, it doesn't; filtercites.length returns 7. What am I doing wrong?

The code is at User:Trappist the monk/common.js and the sandbox is at User:Trappist the monk/sandbox2.

—Trappist the monk (talk) 11:14, 26 April 2020 (UTC)

@Trappist the monk: cites is not an array, it's a jQuery collection. To convert it to an array, use the .get() method. So

var filtercites = cites.get().filter(function(x) { return sectionCites.indexOf(x) === -1; });

will work. SD0001 (talk) 12:40, 27 April 2020 (UTC)

@SD0001: Ding ding ding! That works, thank you.

—Trappist the monk (talk) 16:32, 27 April 2020 (UTC)

Given the promiscuous usage of footer headings that we can see in articles old and new, I'm really uncomfortable about assuming semantics for any particular section heading. It's very common to see Further Reading contain external supplementary text, citations used in the article, or both. External Links is often used as a synonym. There are creative heading names like "Notes and References" or "Footnotes". And, of course, it's very common to have citations in the same section as the reflist. If anything, this feels like a call to arms to standardize the semantics of the footer sections of all 6 million articles and enforce it going forwards (with appropriate changes to various guidelines/policies). Sounds like a good idea, but a big one. David Brooks (talk) 23:23, 25 April 2020 (UTC)

They vary a lot, yes, but here the idea would be to identify sections that where the warnings are more likely than not to be pointless for editors that would like focus on more pertinent warnings, even if they miss the occasional one. For example, Further reading, External links, Selected works, Publications, List of publications etc... are very likely to not be the intended targets of SFNs, so the script could ignore those which are deemed safe/safe-ish to ignore. This could be done two ways, conceptually

• a) Centrally, by importing the script with default sections that are ignored
• b) Locally, by allowing something like "IgnoreWarningSections = {Further reading, External links, Selected Works}", etc... to be loaded in a .js or .css page, where a user could customize the list to their liking.

Headbomb {t · c · p · b} 23:30, 25 April 2020 (UTC)

I understand that we agree in principle, but I've found that Further Reading is not uncommon as a place where people put sfn'ed citations. As I think I implied, it's my recollection that footer names and semantics are a guideline, not a policy. And even if it were a policy, it's rarely observed, or corrected by gnomes. David Brooks (talk) 23:37, 25 April 2020 (UTC)

Well, also remember that this would be an optional script. So if there's a centralized default, it would be up to you to decide if you're ok with that behaviour. And if it's customizable, then you can have your own judgment of what you'd want to ignore or not. Headbomb {t · c · p · b} 23:56, 25 April 2020 (UTC)

I'm done hacking on my variant of the script. It ignores the content of §Further reading as long as that section is a level 2 section (==Further reading==) but it also looks for linked citations in that section and emits an error message when it finds them. My thanks to Editor SD0001 for the help.

—Trappist the monk (talk) 16:49, 27 April 2020 (UTC)

@SlimVirgin: try User:Trappist the monk/HarvErrors.js now. It should reduce warnings to only those articles that make use of short footnotes, and ignore those from the further reading section. I'm sure @Trappist the monk: would also be willing to extend the behaviour to others sections too, like External links, where it makes sense to assume those warnings are unwanted. Headbomb {t · c · p · b} 19:29, 27 April 2020 (UTC)

I'm still concerned that, by making these distinctions concrete, we are backing into a requirement to apply stricter semantics to these footer sections (so-called Appendices). That, in itself, deserves an open discussion. The content and use of footer sections does not exhibit unanimity, perhaps because of editors making invaid deductions from existing samples, perhaps just because of inadequate thought on the meaning of the terms, perhaps because of pure cargo-cultishness. For example, "External Links" and "Further Reading" seem to be interchangeable in practice (by which I mean two editors could use the two terms to mean the same type of thing), as do "External Links" and "Sources". Not to mention that all of these are often replaced by boxes pointing to Wikisource copies. To be sure, this is all called out in MOS:LAYOUT, but that's just guideline, not policy, and is pretty loose; see for example the wording of Notes and references. I think it would be a fantastic project to normalize and harden all this stuff, and in particular I think that is a pre-requisite of what's being done here. David Brooks (talk) 22:21, 27 April 2020 (UTC)

Hence the optional nature of these scripts. Headbomb {t · c · p · b} 22:26, 27 April 2020 (UTC)

Trappist, Headbomb and David, the thing is: we need what we had. We had the flexibility of adding ref=harv or not adding it. Together with the script, that allowed us to move citations around, with and without the warnings/error messages as needed. Are you now having to decide in advance the names of sections so you can try to mimic that functionality? SarahSV ^(talk) 02:11, 28 April 2020 (UTC)

Repeating myself from above, in different words that may register this time: the Harverrors script does (a) with a red error message; does (b) with a brown warning message; and does (c) if the full citation includes |ref=none as a parameter. The only change in your workflow should be adding |ref=none as appropriate instead of the previous step of excluding |ref=harv. – Jonesey95 (talk) 03:52, 28 April 2020 (UTC)

Please allow me to repeat myself too in the hope that this registers. This isn't about one person or the small group of people who have complained. It's about a significantly larger group (I assume hundreds) who are now seeing large and very ugly brown warnings after every single citation in their Selected works section that might have taken them hours or even days to compose—and in other articles that they didn't edit, making the sections harder to read—and don't know why. SarahSV ^(talk) 03:58, 28 April 2020 (UTC)

I installed Trappist's script. It does (a) and (b) above (warns if there's a short cite without a corresponding long one and vice versa). But it's doing it where there's no need. For example, in Auschwitz this citation is generating a warning:

{{cite news |last1=Astor |first1=Maggie |title=Holocaust Is Fading From Memory, Survey Finds |url=https://www.nytimes.com/2018/04/12/us/holocaust-education.html |work=The New York Times |date=12 April 2018 |archiveurl=https://archive.today/20180418071414/https://www.nytimes.com/2018/04/12/us/holocaust-education.html |archivedate=18 April 2018|url-status = live}}

It is doing some of (c): not warning about citations without ref=harv. It doesn't generate a warning in Further reading or Selected works, but it's generating one in External links. SarahSV ^(talk) 03:14, 28 April 2020 (UTC)

That's no different than if you used

{{citation |last1=Astor |first1=Maggie |title=Holocaust Is Fading From Memory, Survey Finds |url=https://www.nytimes.com/2018/04/12/us/holocaust-education.html |work=The New York Times |date=12 April 2018 |archiveurl=https://archive.today/20180418071414/https://www.nytimes.com/2018/04/12/us/holocaust-education.html |archivedate=18 April 2018|url-status = live}}

And, as with {{citation}}, |ref=none will suppress those warnings if they are not needed. Headbomb {t · c · p · b} 12:29, 28 April 2020 (UTC)

Correction: it does generate a warning when the headings are Selected works, Bibliography, Sources, and Citations. "Further reading" is the only one that prevents it. SarahSV ^(talk) 06:29, 28 April 2020 (UTC)

So identify which sections should not emit those warnings. Bibliography, Sources and Citations should likely emit them, given those are pretty synonymous with References, but the warnings in External links, Further reading (currently the only one ignored), Selected works, List of publications, and Publications are probably not very useful. Are there others? Headbomb {t · c · p · b} 12:29, 28 April 2020 (UTC)

Just a quick note here that Trappist's script now excludes both External links and Further reading sections by default. Headbomb {t · c · p · b} 19:37, 29 April 2020 (UTC)

Problem

Moved from Help talk:CS1#Problem Headbomb {t · c · p · b} 12:55, 4 May 2020 (UTC)

The new script, User:Smith609/citations.js, is showing an error message for the second long citations in bundles. See the References section of FGM, footnotes 170, 260, 262, 263. I can remove them with ref=none, but why is that happening? SarahSV ^(talk) 02:38, 4 May 2020 (UTC)

User:Smith609/citations.js is not a new script; it was created in February 2011 with occasional changes; the last of which was in October 2019.

Looking at User:SlimVirgin/common.js, I think that you are talking about the messages emitted by User:trappist the monk/HarvErrors.js. Were you to switch to User:Ucucha/HarvErrors.js, you would see the same messages except that they would be larger.

Why the warning message? Each cs1|2 template creates an anchor ID. Take for example this bundle of Wilkerson 1988 and Donaldson James 2012. That bundle uses {{pb}} to separate the two cs1|2 templates. At the html level, the basic bundle (with a lot of stuff stripped out for clarity), looks like this:

<li id="cite_note-200">
	<span class="reference-text">
		<cite id="CITEREFWilkerson1988" class="citation news">Wilkerson, Isabel...</cite>
		<div class="paragraphbreak" style="margin-top:0.5em"></div>
		<p>
			<cite id="CITEREFDonaldson_James2012" class="citation news">Donaldson James, Susan...</cite>
		</p>
	</span>
</li>

User:Ucucha/HarvErrors.js and derivative scripts expect to find an anchor ID (CITEREFWilkerson1988 for the first citation in the bundle) inside two parent elements, where the outer parent has an id="cite_note..." attribute. When these two conditions are met, the script does not emit a warning message. For the second citation (anchor ID CITEREFDonaldson_James2012), the two conditions are not met because the outer parent element (<span class="reference-text">...</span>) does not have an id="cite_note..." attribute. When the two conditions are not met, the script emits the nothing-pointing-here warning.

Remember, you did ask why is that happening?

—Trappist the monk (talk) 03:52, 4 May 2020 (UTC)

Sorry, I posted the wrong script. Thanks for correcting it. Can something be done about it? Lots of people bundle citations. SarahSV ^(talk) 05:18, 4 May 2020 (UTC)

You did say that you can remove them with ref=none; is that not doing something about it?

—Trappist the monk (talk) 12:43, 4 May 2020 (UTC)

Script Request for Specific Editors

Hello, a few of us in the spoken and video guild are looking for a script that would have a manual push button to be able to the following for us:

1. Remove references
2. Remove the {}[] and all content within said brackets (exceptions are infoboxes, tables, and photos as they should not be touched)
3. Any Wikipedia markup (i.e., bold, italic) without affecting the contents between the apostrophes
4. Not allowed to save to the main article; however, by clicking submit, it would keep it to the respective user's sandbox with the same title and appending "-Script and full date as MMDDYY" to it. (This is meant to protect the article)
5. Should only be allowed to run by those in the respective guilds (list of participants), and those that need access like admin, stewards, programmers, etc. Regular editors should not have access to run this script as to what it does. (If possible and not conducive to the WP at large)

I look forward to any discussion on this and to answer any questions you may have. I would also like to assist in testing prior to it being released to ensure it does what I am asking for.

I appreciate all the hard work you do.

Thanks, Galendalia _{CVU Member} \ ^{Chat Me Up} 18:41, 6 May 2020 (UTC)

Galendalia, removing all {{ and }} and all content between them would also remove {{convert}}, but don't you still need to see it and other inline templates? – BrandonXLF (talk) 19:01, 6 May 2020 (UTC)

@BrandonXLF: - we basically need to create a text only version with minimal interfaces that would be like reading a book. I think the {{tl|convert}}, could be made an exception if possible. We do still go through all the articles we have and edit them, this is just looking to save time especially on our longer articles. Thanks Galendalia _{CVU Member} \ ^{Chat Me Up} 19:07, 6 May 2020 (UTC)

Section watchlist

Is there a user script to watchlist sections? I remember reading about one months ago, but I can't recall anything else about it. If not, I was thinking of making one. (Likely architecture: Toolforge job to read from the firehose and use an in-memory hashmap lookup to "deliver" edits to specific sections to users' inboxes, and then a small JSON API to fetch new edits and show them in a new panel on the Wikipedia end.) Enterprisey (talk!) 07:06, 14 May 2020 (UTC)

Button help - English refs >> Czech refs

Hey Guys, I need a simple button in the code editor in the "Advanced" tab for replace English Citation Templates to Czech Citation Templates. For example, from english

{{cite web
 |title=
 |url=
 |date=
 |access-date=
 |publisher=
 |language=
}}

To czech language

{{Citace elektronické monografie
 | titul = 
 | url = 
 | datum vydání = 
 | datum přístupu = 
 | vydavatel = 
 | jazyk = 
}}

Can anyone help me with this? Thanks--MrJaroslavik (talk) 17:10, 14 May 2020 (UTC)

MrJaroslavik, do you have a list of Czech parameters and their corresponding English parameter? – BrandonXLF (talk) 19:10, 14 May 2020 (UTC)

BrandonXLF Yeah, i will send it here. --MrJaroslavik (talk) 19:15, 14 May 2020 (UTC)

BrandonXLF Here

{{cite web=Citace elektronické monografie
|url=url
|title=titul
|last=příjmení
|first=jméno
|author-link=odkaz na autora
|series=edice
|publisher=vydavatel
|location=místo
|pages=strany
|language=jazyk
|type=typ vydání
|doi=doi
|isbn=isbn
|id=id
|access-date=datum přístupu
|url-status=
|archive-url=url archivu
|archive-date= datum archivace 
}}

--MrJaroslavik (talk) 19:37, 14 May 2020 (UTC)

I have more at User:Jonesey95/AutoEd/unnamed.js (do a Find for "Czech"). – Jonesey95 (talk) 22:13, 14 May 2020 (UTC)

Link ISBN direct to WorldCat

Hi, is there a script in existence that, when an {{ISBN}} is linked within a CS1 citation, will link directly to the WorldCat search/address instead of Special:BookSources? czar 15:18, 25 May 2020 (UTC)

Navbox user script

I frequently go through the hassle of filling in navboxes, using subgroups etc. I expected there to be a user script to fill them in but I wasn't able to find one. It would certainly be very useful! — Yours, Berrely • ^Talk∕_Contribs 19:46, 28 May 2020 (UTC)

OneClickArchiver for lists

Scripts listed in Wikipedia:One click archiving implement archiving of whole ==sections==. To make archiving of entries at Wikipedia:Community bulletin board easier, it would be nice to have a OneClickArchiver for * Unordered lists. Pinging Sm8900, who might be interested in such a script. —⁠andrybak (talk) 00:33, 3 June 2020 (UTC)