Commons:Idea Lab
The Idea Lab is used for discussions about ideas to improve Wikimedia Commons, especially (automated) tools to help with Category:Commons maintenance. If you are ready to make a concrete proposal and determine whether it has consensus, use Commons:Village pump/Proposals. This page is for brainstorming ideas but not about voting, "support", "oppose", etc.
Wondering whether someone already had the idea? Search the archives, or look through Commons:Recurring proposals.
The following pages may be more suitable for other discussions:
- Commons:Help desk, general help.
- Commons:Village pump, general forum.
- Commons:Village pump/Technical, technical forum.
- Commons:Bots/Work requests, for bot-assisted editing.
 |
SpBot archives all sections tagged with {{Section resolved|1=~~~~}} after 20 days. For the archive overview, see Special:PrefixIndex/Category:Idea_Lab/Archives. |
Filemover helper[edit]
a gadget that suggests filenames based on existing file descriptions. for example, https://commons.wikimedia.org/w/index.php?title=File:%E9%BB%9E%E4%BA%AE%E5%8F%A4%E8%B9%9F.jpg&oldid=659577663 is already properly described. a filemover could just copy (part of) the description as the new filename.
things to consider:
- a function to easily check who wrote the existing description (the uploader or someone else? to prevent vandalism).
- if existing file description is too long.
- if file descriptions exist in multiple languages.
- if file has no existing description but proper categories.
make this a semi-automatic tool to load files in a category one by one, so that filemovers can clear Category:Files with bad file names more easily.--RZuo (talk) 13:31, 3 December 2022 (UTC)
- I'd say that filenames are only one of the ways we can identify a file, categories and descriptions often work as well. I'm not saying that a good file name isn't helpful, I just don't think that bad filenames are particularly bad unless they are specifically misleading, that is if they misidentify a subject, location, time, Etc. If the uploaded had trouble thinking of a good file name I don't think that a filemover would necessarily come up with something better. Bad filenames aren't as much of an issue as information lacking in general, a good filename but no description and categories is handy, but a bad filename but with everything else checking out is also handy. Yes, filenames should be descriptive, but I don't think that if they're not descriptive and if other fields are descriptive that we should invest resources into also changing them. --Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 10:28, 8 December 2022 (UTC)
Licence review helper[edit]
a programme/web tool to load files in a category one by one on one side of the screen, and automatically load its source (usually a URL) on the other side.--RZuo (talk) 13:31, 3 December 2022 (UTC)
- So, basically a browser within the Wikimedia Commons for license reviewers? That could work, but this would require a lot of technical skills to create. This is exactly why I want to convince the Wikimedia Foundation (WMF) and Wikimedia Germany (WMDE) to invest more of their time and skills into the Wikimedia Commons, this actually seems like a very handy tool that would greatly aid with license reviews, maybe it should be build on the Internet Archive's Wayback Machine too, so it immediately archives the current version loaded and allows the reviewer to search in older versions of the web page in case the license was changed or the link has expired. --Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 08:29, 8 December 2022 (UTC)
Iranian news website images 2 commons[edit]
for Tasnimnews, Mehrnews, Farsnews, Nasimnews, Mojnews... a tool to scrape a web page for images and import them to commons. by using an automated tool, it's ensured that images are imported directly from the website. human reviewers just need to make sure the source web pages are freely licensed. this saves trouble of reviewing photos one by one manually and eases congestion in Category:License review needed.--RZuo (talk) 13:31, 3 December 2022 (UTC)
- This is a part of a larger issue where we have Flickr2Commons but no similar tools for other websites, I suggest that we should have a "Web2Commons" tool that can import based on every URL and tries to isolate the highest quality images, but for the tool to work users would have to add a free license during the upload process, in fact, I think that "Web2Commons" could just become a feature for the MediaWiki Upload Wizard for users with a set number of uploads or "upon request" (maybe create a new user right called "Web uploader" that would be easier to acquire than Autopatrol, but also easy to lose if abused). --Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 08:25, 8 December 2022 (UTC)
Location-based categorisation bot[edit]
Commons:Bots/Work_requests/Archive_16#Idea_of_a_location-based_categorisation_bot.--RZuo (talk) 13:31, 3 December 2022 (UTC)
- That is actually a great idea, but the issue is flooding. Let's say that 15,000 uncategorised images were taken in San Francisco, would it be wise to "dump" all these images in the main cat or would the bot be able to identify neighbourhoods? --Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 08:22, 8 December 2022 (UTC)
Script / supervised bot to help cfd[edit]
sometimes cfd result in massive rename of cats and moving files across cats. a script to simplify the manual procedure "move (with/without redirect), then cat-a-lot" would be helpful.--RZuo (talk) 13:41, 3 December 2022 (UTC)
- Incidently, consider how "cat-a-lot, then move (with/without redirect)" is a much better way to do it, as it avoids the temp "~" subcat.
- As for the proposal at hand, I cannot see how it’s any different than the existing vfc, or one of the multiple uses thereof.
- -- Tuválkin ✉ ✇ 16:15, 5 December 2022 (UTC)
- it might be possible to adapt en:User:JJMC89 bot III for this. RZuo (talk) 19:16, 14 January 2023 (UTC)
Motd helper[edit]
a gadget, when user is on a video/audio file page, lets the user create motd templates using that file, easily choose a date which doesnt have existing motd, and write descriptions in different languages.--RZuo (talk) 22:47, 4 December 2022 (UTC)
Motd description migration bot[edit]
check motd 5 days ago. migrate descriptions in languages that do not exist on the file page yet. migrate the ones edited solely by autopatrolled or above users, and all "stable" ones contributed by any user before that motd went live. (to minimise sabotage.)--RZuo (talk) 22:47, 4 December 2022 (UTC)
Rapid diffusion bot[edit]
diffuse a target cat into "by year" and "by month" subcats, optimally based on date of photograph. (Category:United States Navy Band has 3k photos right now. best solution is diffusion into yearly subcats.) generates an analysis before actual diffusion, so as to see whether diffusion is feasible and good.--RZuo (talk) 09:17, 17 December 2022 (UTC)
- RZuo, I think that this will probably be possible, but this bot should ignore files without metadata (as upload data is less reliable than metadata), and it should work upon request, that is people should suggest which categories get treated and perhaps the requestor should write the standard template for the dated category. But we should pre-prepare as much as possible so anyone can request anything from this bot without much prior knowledge (just enough to work with it), this way the barrier for entry will be low enough that any users (not just powerusers) can use it. -- — Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 19:27, 17 December 2022 (UTC)
Template:CJK category revamp[edit]
fetch radical (P5280), residual stroke count (P5281), Unicode code point (P4213) etc. from wd, instead of manual parameters. RZuo (talk) 13:15, 20 December 2022 (UTC)
Big button / keyboard shortcut to copy pagename to clipboard (add links)[edit]
js to add a bigass button beside pagename and/or a keyboard shortcut. click the button and the fullpagename is copied.
this saves time to select pagenames with the mouse. RZuo (talk) 15:45, 1 January 2023 (UTC)
- MediaWiki:Gadget-CategoryAboveAll.js MediaWiki:Gadget-CategoryAboveBelowImage.js can be adapted for this maybe. RZuo (talk) 13:50, 11 February 2023 (UTC)
Youtube playlist to commons and motd[edit]
i add good and free videos to a youtube playlist occasionally. a tool to automatically import all videos in a playlist and space them out for future motd would be great. RZuo (talk) 16:35, 1 January 2023 (UTC)
- Isn't there currently an issue with importing videos from Google's YouTube streaming service? I remember that Video2Commons wasn't able to import videos due to Google changing the API's or something. --Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 20:48, 18 January 2023 (UTC)
Video with multiple audio tracks[edit]
this is something beyond commons, but i'll jot it down here first.
i was just thinking of making a video tutorial for v2c, or it could be a cooking recipe. in consideration of different languages, it would be best to make a video suitable for audio commentary in all languages. then other users can make their dubs and add them to the videos.
here comes the problem. afaicu, to add additional soundtracks to a video, the whole video has to remuxed. it's a daunting task for many people, and each new edit would mean creating new versions of a video, i.e. lots of redundant big files.
so my idea is, would it be great if additional soundtracks can be hosted separately, and during video playback, users can choose which tracks to play. that would allow playing different languages, or enjoying a movie (with the original sound) while listening to a commentary soundtrack.
youtube is just experimenting multiple tracks https://support.google.com/youtube/thread/129769858/updates-to-captions-and-audio-features-on-youtube . RZuo (talk) 13:21, 9 January 2023 (UTC)
Gallery generator[edit]
(if commons galleries should still exist)
i imagine an extension of Commons:Glamorous to generate a "deepcat:example" list of images. user can set a minimum number of usage in wikis (instead of the default 1) to filter the files. this list could be used to easily prop up a gallery page.
assumption: file usage implies that those files are more likely to be good illustration of the subject. RZuo (talk) 16:00, 9 January 2023 (UTC)
- for example, for special:diff/724095465, i used glam and got its result in xml https://glamtools.toolforge.org/glamorous.php?doit=1&category=Valentine%27s+Day&use_globalusage=1&ns0=1&depth=99&show_details=1&show_details_limited=1 . then i copied the xml from start to ' usage="5" ' (limiting to files that have been used 5 times or more), pasted into a regex filter and got the list of filenames. finally, i removed a few more files while editing the gallery page. an integrated tool to streamline this process (and include file descriptions for the gallery) will be the generator. RZuo (talk) 16:31, 9 January 2023 (UTC)
- https://gallery.toolforge.org/ https://github.com/ferdi2005/gallery RZuo (talk) 12:47, 19 April 2023 (UTC)
I actually agree with this, good galleries are a severely under utilised aspect of the Wikimedia Commons and this is partially because it's often so difficult to create galleries because one would often have to deep dive to get them. More assistance would help. --Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 20:43, 18 January 2023 (UTC)
One-for-all multilingual motd description[edit]
for Template:Motd/2023-01-19, i wanted to write only {{w|Perchtoldsdorf}} as the description. since the name is the same across many languages, it would be good if only 1 template needs to be created but can be displayed in all languages. RZuo (talk) 09:07, 16 January 2023 (UTC)
- Perhaps this would be better if script based, for example all languages using the Latin script use this name, but then in Arabic, Cyrillic, Greek, Traditional Chinese, Etc. it might be displayed differently. If it's likely that all Latinoscribe languages use the same then them it would make sense. --Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 20:46, 18 January 2023 (UTC)
- actually rather easy. i would create the multilingual using langcode mul, e.g. 2023-01-11 (mul).
- initially i didnt care about non-latin languages, because who cares. they wouldnt have too much difficulty finding out about the subject by googling.
- but since it's mentioned, i thought about using {{Q}}.
- but actually {{Label}} is even better.
- the code that needs changes is Main Page/motd. while checking this i realise current design falls back to en if no description is given for other languages. so all it takes is to add one more layer. first try falling back to mul. if mul doesnt exist, then fall back to en.
- but i cant edit that page now, so i'll hold this off for a while. RZuo (talk) 23:31, 18 January 2023 (UTC)
- expanding on this. perhaps a template could be designed to take a few "keywords".
- so in addition to a description, it's optional to set a few topics for that day's motd. a template like {{motdtopic|123|456|7899}} would then show something like
Topics of MOTD: {{Label|123}}; {{Label|456}}; {{Label|7899}}.RZuo (talk) 23:44, 18 January 2023 (UTC)- this idea is a bit moot, since current design falls back to en anyway. i thought i had seen motd without description. i guess that must be when no description was given in en.
- i think, for people who cant understand english so well, seeing a well written english description is still better than a few keywords in their languages. but maybe the "topics of motd" is still a good idea. imagine seeing
- MOTD主題：甲；乙；丙；丁。
Today's MOTD shows blah blah blah... (detailed description in en)- being able to read a few keywords might be beneficial for users who dont have easy access to translation service, e.g. mobile users. RZuo (talk) 23:54, 18 January 2023 (UTC)
An international team of astronomers using the NASA/ESA Hubble Space Telescope has discovered a ghostly ring of dark matter formed long ago during a colossal collision between two galaxy clusters. This is the first time that dark matter has been found with a distribution that differs substantially from the distribution of ordinary matter. Credit:ESA/Hubble (M. Kornmesser & L. L. Christensen)
MOTD topics: dark matter / gravitational lens / Hubble Space Telescope- i imagine this'll be how Template:Motd/2023-01-20 (en) looks like.
- if someone is seeing this in fr, because Motd/2023-01-20 (fr) is missing, s/he will see the english description, together with the "MOTD topics" line all in french, and the labels linking to frwp. RZuo (talk) 10:58, 19 January 2023 (UTC)
Cat diffusion aid[edit]
- given a cat, a tool to go thru all its subcats and list subcats that contain files more than a number like 180.
- a tool to detect AHKP1 and de-overcat all his files.
- a tool to detect AHKP1 and list all his cats that have multilingual titles.
maybe petscan can do these, but i'll try later when i have time. RZuo (talk) 09:52, 16 January 2023 (UTC)
Bot tagging new uncat files with potential languages[edit]
bot to check new files in cats like Category:All media needing categories as of 2023. tag files with a language-specific "uncategorised" cat based on existing file captions or lang templates. for example, https://commons.wikimedia.org/w/index.php?oldid=615400493 can be tagged for yue and zh-hant.
this bot should patrol files not older than a few days, because the longer after upload, the more likely captions/descriptions would be added in languages that might not be so relevant. but the bot should run the files after #Location-based categorisation bot has patrolled them first, because if cats based on coord locations have been added it's not necessary for this bot anymore.
languages like en,fr,de,es,nl... wont be processed by this bot, because there're a large number of active users of these languages. also files containing these languages are too many. tagging them doesnt have much benefits. RZuo (talk) 20:55, 18 January 2023 (UTC)
Countries of continent[edit]
Template, input=country, output=template:countries of continent[edit]
template to integrate separate "countries of continent" into one.
{{Countries group|Egypt|prefix=xx|suffix=yy}}
=
{{Countries of Africa|prefix=xx|suffix=yy}}
{{Countries of Asia|prefix=xx|suffix=yy}}
{{Countries of the Arab world|prefix=xx|suffix=yy}}
this is not difficult, but i dont have time yet. RZuo (talk) 14:00, 24 January 2023 (UTC)
Populator[edit]
on a cat page using "countries of continent", activate this tool. the tool prepends missing template to all existing cats of other countries (like how it is on the given cat page).
further, instead of only prepending, also fix malformed templates (correcting prefix/suffix parameters, etc.).--RZuo (talk) 11:34, 10 February 2023 (UTC)
Topic year country: 1 template for all the layers?[edit]
would it be possible to create 1 template that can be used by all 5 layers (that result from intersection of topic, period and country), or at least combining some layers in 1 template? if possible, is it actually better for management, or worse because template would be too complicated to maintain?
right now many templates are created for different levels and different countries separately, even though they are parallel to each other. an example can be seen at Commons:Deletion requests/Videos from X by year.--RZuo (talk) 14:27, 24 January 2023 (UTC)
Add URL as an option to the MediaWiki Upload Wizard[edit]
Currently if you use the MediaWiki Upload Wizard (MWUW) you can upload media files from your device, Autopatrolled users can also use a specialised tool to import files from the website Flickr, this Flickr import tool works fine for users who are trusted to import from this website and the MediaWiki Upload Wizard can easily detect the license of the file. I suggest that we use this existing technology to import images from any web page.
I'm not sure if we should only limit this to trusted users as a template to let license reviewers check it might be useful enough to get the right people to confirm. This (license review) template could be automatically added when the URL-import option is used, similar to how Flickr files are treated today. By loading a URL all images are scraped from the URL and duplicates will immediately be detected, the uploader can them remove all the images that (s)he doesn't want to import and select the free educational ones that they do want to import. This option will make it much easier for people to import images in bulk rather than first (1st) having to download the images from the website and then uploading them through the MediaWiki Upload Wizard 🪄 we can now cut out half of the process and directly import from websites saving people a lot of time. --Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 22:11, 24 January 2023 (UTC)
- it doesnt scrape websites. it doesnt work for any website. read mw:Extension:UploadWizard mw:User:Drecodeam/GSoC 2012 Application. RZuo (talk) 23:13, 24 January 2023 (UTC)
- It doesn't currently do that. Having an all-purpose import tool would greatly help most importers and having that tool be in the most visible and accessible place possible is even more important to motivate content creation. The software can always be updated as it does allow users to scrape Flickr albums, which means that the possibility is there. Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 23:17, 24 January 2023 (UTC)
Clever tool to guess existing incremental pages (add links)[edit]
suppose you're on Commons:Village pump/Archive/2022/12. this tool will detect yyyy numbers or serial numbers in the title, then test whether titles with neighbouring numbers exist (../2022/11, ../2023/01, ../2023/12, etc.). if they exist, insert a link to each page on the top of the page.
essentially the same function as some archive header templates or category navigational templates, except no template needs to be placed on the page. RZuo (talk) 17:13, 27 January 2023 (UTC)
js from user talk to their personal DR (add links)[edit]
on user talk, detect presence of "Commons:Deletion requests/Files uploaded by PAGENAME", if yes, add link to it.--RZuo (talk) 07:11, 29 January 2023 (UTC)
Document viewing helper, gadget or built-in[edit]
Commons:Village_pump/Archive/2022/11#Document_viewing_helper,_gadget_or_built-in for better pdf/djvu viewing experience. RZuo (talk) 12:07, 29 January 2023 (UTC)
Fix double cat redirect[edit]
i think a lua module can put the pages into a maintenance cat and then have a bot fix them. RZuo (talk) 11:14, 6 February 2023 (UTC)
Gadget to add helper links on top of search result page[edit]
add some links like MediaWiki:Searchmenu-new does. one click to get "sorted by creation date" "sorted by edit date" etc. instead of the stupid selection menu that takes more clicks and scrolls. RZuo (talk) 09:11, 25 February 2023 (UTC)
Module:category tree[edit]
wikt:Module:category tree looks like a good tool to introduce? produces handy lists of newest and oldest edited pages in a cat? RZuo (talk) 11:34, 26 February 2023 (UTC)
Gadget to view all thumbnails of files linked in DR[edit]
makes it easier to check files in DR.
before that happens, i can add a link to DR pages that links to " Category:Deletion requests Month yyyy (filefrom=dd#SUBPAGENAME) ". RZuo (talk) 12:09, 4 March 2023 (UTC)
Bot to remove fraudulent VI, FI, MOTD, etc.[edit]
a bot is needed to remove fraudulent claims of COM:VI, etc. RZuo (talk) 09:52, 15 March 2023 (UTC)
Newspapers from hkpl[edit]
- turn http://manjai86.blogspot.com/2017/06/blog-post.html into a tool.
- (convert to PDF), mass import PDF together with the "headlines" texts.
RZuo (talk) 16:50, 15 March 2023 (UTC)
Bot reports on licence reviewer stats[edit]
might be a good idea to have a bot weekly generate stats for LR. a bit more scrutiny helps boost trust in the system? RZuo (talk) 12:30, 21 March 2023 (UTC)
AI discussion summary on demand[edit]
if a template "summarise this page" is placed on a page, a bot comes to generate a summary of that page (utilising chatgpt or something similar).
this might help summarise long and boring (for people who dont have an interest in the topic) discussions, like cfd. the bot's summary is not to be taken as perfect, but merely an easy and lazy way to quickly summarise long discussions. users are expected to correct the bot's summary if necessary. RZuo (talk) 17:28, 15 April 2023 (UTC)
- RZuo, This seems like a good idea, I've seen robots do this at Reddit and it seems to work really well. This would especially be good for admins who might not have all the time in the world. — Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 18:09, 15 April 2023 (UTC)
OCR to file description tool[edit]
- have a link like croptool does on the sidebar.
- click the link. automatically load the file into an ocr tool, like https://ocr.wmcloud.org/ .
- have a "submit" button. when copyediting is done, users can specify the langcode and click "submit", which will add the preformated transcribed text to the file description. the format would be like {{en|...}}.
RZuo (talk) 16:32, 5 May 2023 (UTC)
Tool to give links to discussion pages based on filename[edit]
File:Example.svg
if Commons:Deletion requests/File:Example.svg or other similar discussion pages constructed based on filename exists, display the links to these pages, somewhere near span class="mw-page-title-main".
saves the trouble of going to file talk for these links. RZuo (talk) 21:47, 15 May 2023 (UTC)
- This can probably be done via MoreMenu; I'll look into this later. —Mdaniels5757 (talk • contribs) 21:49, 15 May 2023 (UTC)
Done via MoreMenu.- @RZuo Enable MoreMenu at [1] and take a look under "Page" on, e.g., File:ANA_Cargo_logo.png. There will be an option called "DR", clicking it will bring you to the DR. —Mdaniels5757 (talk • contribs) 17:40, 3 June 2023 (UTC)
- thx. i dont have time now, but i'll definitely try it out first. having looked at the documentation, one thing i think that can be improved is if the link isnt hidden in a dropdown menu that requires the mouse to move there (not sure if the menu displays completely with a wide screen, but i dont have a wide screen anyway).
- i have a vague idea about making a tool to add all sorts of helper links that go under "From Wikimedia Commons, the free media repository", but no time to actually do it now.
- will come back to this later.--RZuo (talk) 11:30, 12 June 2023 (UTC)
- @Mdaniels5757 i dont have a solution for cfd, but i have the following code for MediaWiki:Gadget-MoreMenu.extension.js to link Files uploaded by xx.
api.get({
titles: "Commons:Deletion requests/Files uploaded by " + config.targetUser.name,
formatversion: 2,
}).done(function (data) {
if (!data.query.pages[0].missing) {
MoreMenu.addLink(
'user',
'Uploads DR',
mw.util.getUrl("Commons:Deletion requests/Files uploaded by " + config.targetUser.name)
);
}
});
- RZuo (talk) 22:28, 5 July 2023 (UTC)
- @Mdaniels5757 the code doesnt seem to work.
- i dont know whether it's related to the whitespace in
by ". :/ RZuo (talk) 00:35, 6 July 2023 (UTC)- @RZuo It works fine for me on your page. Try bypassing your cache and try again. —Mdaniels5757 (talk • contribs) 00:38, 6 July 2023 (UTC)
- RZuo (talk) 22:28, 5 July 2023 (UTC)
Bot to tag uncat files[edit]
User:SteinsplitterBot was the latest bot in charge of this task but now a new bot is needed to tag uncat (or only under hidden/nonexistent cats) files.--RZuo (talk) 11:30, 12 June 2023 (UTC)
- example: Special:Diff/648063211. RZuo (talk) 14:42, 13 November 2023 (UTC)
Videos from YouTube[edit]
gonna add tracking cats to {{From YouTube}}.
my approach is to use Module:String2 to extract the file extension. then use switch to determine which are videos. i'm not sure if there's a more direct way of determining filetype.--RZuo (talk) 14:36, 12 July 2023 (UTC)
Page status for timedtext[edit]
for com:TimedText, something similar to s:Help:Page status to mark subtitle progress and quality. RZuo (talk) 20:50, 2 December 2023 (UTC)
Uploading textured meshes[edit]
I know this topic was raised already some months ago, but I think it would be the greatest addition since the integration of STL as allowed filetype in 2018 to be able to upload textured meshes. See: https://phabricator.wikimedia.org/T246901
--PantheraLeo1359531 😺 (talk) 16:35, 6 December 2023 (UTC)