Wikipedia:Bots/Requests for approval/DipankanBot

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

Denied.

DipankanBot

Operator: Dipankan001 (talk · contribs · SUL · edit count · logs · page moves · block log · rights log · ANI search)

Time filed: 06:09, Wednesday May 30, 2012 (UTC)

Automatic, Supervised, or Manual: Automatic

Programming language(s): Python

Source code available: Pywikipedia

Function overview: ~~Welcoming new users, adding missing <references /> tags to articles, cleaning of sandbox, and~~ tagging newly uploaded files with no information on copyright status.

Links to relevant discussions (where appropriate):

Edit period(s): Daily

Estimated number of pages affected: 2000-3000 Exclusion compliant (Yes/No): Yes.

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

Function details: Hello all, I present you my first bot, DipankanBot. It runs on Pywikipedia and tags newly uploaded files without any licensing tags on them for deletion. It checks the whole page for available license templates from all the licensing categories, i.e Non free categories, Public Domain categories, Copyright from countries and all the copyright categories present. It will further report images that seem to have a fake license at User:Filnik/Report. Dipankan ^{(Have a chat?)} 06:09, 30 May 2012 (UTC)[reply]

Discussion

Way too many tasks for one BRFA, please split this task into multiple BRFAs. Also:

Welcoming new users - No
Cleaning of sandbox - This should be fine, however you need to be careful to ensure that the bot doesn't get into an edit war with one of the current sandbox bots
Adding missing <references /> tags to articles - Is there consensus for a bot to do this? (Does a bot currently do this?)
Tagging newly uploaded files with no information on copyright status. - Needs a lot more detail. What will it use to determine which images should be tagged? Will it notify users? What template will it tag images with? If for some reason an image is tagged incorrectly, what happens?

--Chris 08:04, 30 May 2012 (UTC)[reply]

OK, it might do just one task for now. I'd suggest cleaning of sandbox or tagging of newly uploaded files with no information on their copyright status. As to your queries, this bot will not be edit warring with the current sandbox , I'm sure of that. And for the images, it scans through all newly uploaded files on Wikipedia, and sees if there is any available licenses. If there is no license templates found on the image, it tags them as having no information on their copyright status. It will also notify users. BTW I don't think any bot fixes the reference thing; and since you don't agree; I will not run it. So should you get a trial for this bot? Dipankan ^{(Have a chat?)} 10:09, 30 May 2012 (UTC)[reply]

I would say no to Sandbox task. We don't need a million bots to clean up the sandbox. One bot is sufficient.—cyberpower ^Chat_Online 20:46, 30 May 2012 (UTC)[reply]

That's fine then. We might be observing Memorial Day when a thousand bots clash with each other LOL... The tagging task is pretty fine, though. Dipankan ^{(Have a chat?)} 06:22, 31 May 2012 (UTC)[reply]

Do we not already have a bot that tags images that do not have proper licencing? (talk→ BWilkins ←track) 16:33, 1 June 2012 (UTC)[reply]

ImageTaggingBot handles new uploads with no source/no license. --Carnildo (talk) 21:28, 8 June 2012 (UTC)[reply]

{{BAGAssistanceNeeded}} Get a trial for 50 edits? Dipankan ^{(Have a chat?)} 10:07, 8 June 2012 (UTC)[reply]

So this is just the image tagging bot right now? You still haven't answered some of Chris's questions above. What templates would you use to tag images and notify users, and how would you handle incorrect taggings? Also, I'm wondering if a short delay is a good idea after the image is uploaded and before we tag it, in case the uploader needs a few minutes to add a license. — The Earwig ^(talk) 23:18, 8 June 2012 (UTC)[reply]

It uses {{di-no license}} along with the date for tagging images. It notifies users with {{subst:di-no license-notice|1=Filename.ext}} ~~~~ . It sometimes might clean the sandbox (rarely). Yes, it checks the older files first in the log and then the new files. If this bot goes wrong there will be a false positive page created for the bot. It also reports images which seem to have a fake license at User:Filnik/Report. I don't the technology behind this, but it really works and is great. Dipankan ^{(Have a chat?)} 05:59, 9 June 2012 (UTC)[reply]

What is your criteria for deciding that a file doesn't have a license? Is it simply the absence of a known license tag, or do you also do things like check the "permission =" field of the {{Information}} template? --Carnildo (talk) 21:59, 10 June 2012 (UTC)[reply]

It checks the whole file page for every known licensing tags from all the categories, and if there isn't one, it will tag them. Dipankan ^{(Have a chat?)} 06:17, 11 June 2012 (UTC)[reply]

Before this does anything, please fix the function overview and function details to reflect what this request is actually requesting in light of the above discussion. You may use <s>...</s> to strike out the parts that are no longer being requested, or rewrite them entirely. Thanks. Anomie ⚔ 14:35, 11 June 2012 (UTC)[reply]

Done, updated. Dipankan ^{(Have a chat?)} 06:23, 12 June 2012 (UTC)[reply]

What do you mean by "reports images which seem to have a fake license"? How does it determine this? What do you mean when by "I don't the technology behind this [sic]"? Did you write the code that you are going to be using? If not, who did and will you make this code available for review? I'm seeing a lot of eagerness to run a bot, and not a lot of evidence of the technical skill and the understanding of policy to do so. Thank you, — madman 14:09, 15 June 2012 (UTC)[reply]

It's running on Pywikipedia, source code is there, you can see from there. Dipankan ^{(Have a chat?)} 14:48, 15 June 2012 (UTC)[reply]

Some points:

I think that madman's concern (along with my own) is that pywikipedia has no inherent, built-in way of magically determining if an image is a copyvio or incorrectly licenced (i.e. "seem to have a fake license"). Without further explanation of just how, exactly, your bot will provide this functionality, we'll have difficulty approving your bot.
Although someone may not use one of our licensing tags, it doesn't necessarily mean that they haven't otherwise correctly filled out licensing information. Because of this, it may be inappropriately premature to tag a file for CSD without a little more logic. For example, while the uploader's license might not have a template, it could still be compatible with our licensing (e.g., someone could literally license it as "do whatever you want with it" or under the Do What The Fuck You Want To Public License).

--slakr^\ talk / 23:13, 22 June 2012 (UTC)[reply]

I agree with what you said, but I had tested it. The majority of new users nowadays use the File Upload Wizard which is the default. Using it requires the image to be marked by a license template. Surely the bot would not catch the wrong ones. Dipankan ^{(Have a chat?)} 10:02, 23 June 2012 (UTC)[reply]

Any particular reason you're making edits to someone else's user space? Are you using someone else's script? --slakr^\ talk / 01:32, 26 June 2012 (UTC)[reply]

The fact is that I'm not a admin, so putting up a list in my userspace would be pretty difficult and hard work for me, since I've to report those everytime it checks. Dipankan ^{(Have a chat?)} 11:05, 26 June 2012 (UTC)[reply]

So why not just change it yourself, you don't need to be an admin. Go to the folder where all your pywikipedia stuff is, right click on checkimages.py and select 'Edit with IDLE' and then go to line 331. Swapout Filnik for Dipankan001. There you are. It will report in your userspace now, instead of the guy's this script was originally written for. Rcsprinter (yak) 18:24, 26 June 2012 (UTC)[reply]

Done. Thanks for the help! Dipankan ^{(Have a chat?)} 14:53, 29 June 2012 (UTC)[reply]

All issues seem to have been adressed. As a nearly-but-not-quite BAG member I can't give a trial myself, but am recommending a seven day trial to commence as soon as a BAG puts the proper template up. Neat idea, Dipankan. {{BAGAssistanceNeeded}} Rcsprinter (yak) @ 11:28, 7 July 2012 (UTC)[reply]

I'm not sure that we need a bot task that will duplicate ImageTaggingBot's functionality; as there's no harm inherent in such redundancy, however, you are

Approved for trial (25 edits). Please provide a link to the relevant contributions and/or diffs when the trial is complete. as long as you change this line 'en':[u'{{nld', u'{{no license'], to this 'en':[u'{{nld', u'{{no license', u'{{untagged'], so your bot will recognize files that have been tagged by ImageTaggingBot. — madman 19:13, 10 July 2012 (UTC)[reply]

{{OperatorAssistanceNeeded|D}} Almost done yet? Rcsprinter (converse) @ 16:30, 27 July 2012 (UTC)[reply]

Ya it is. Dipankan ^{(Have a chat?)} 06:08, 28 July 2012 (UTC)[reply]

Trial complete. then, so it can be reviewed. — madman 20:39, 28 July 2012 (UTC)[reply]

Looks like it tagged one file correctly, but all the rest of the edits were making useless reports. What is its criteria for "fake licenses"? FURs certainly shouldn't be considered "fake licenses", or every new file uploaded with an FUR noted. The bot is noting templates like Do not move to Commons and AWM-image as fake licenses, when there are perfectly valid licenses below them. I'm inclined to think this aspect of the Pywikipedia script is hopelessly outdated and as such, not useful at all. — madman 15:37, 1 August 2012 (UTC)[reply]

Denied. – This task is not sufficiently useful, given its problematic implementation, to justify approval of a task already adequately performed by another bot. — madman 03:58, 6 August 2012 (UTC)[reply]

The above discussion is preserved as an archive of the debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA.