Commons:Bots/Requests/SchlurcherBotT

From Wikimedia Commons, the free media repository
Jump to navigation Jump to search

SchlurcherBotT (talk · contribs)

Operator: Schlurcher (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information)

Bot's tasks for which permission is being sought: Extending Testing new functionality within the scope of the following tasks previously approved for User:SchlurcherBot to auto-protected and otherwise fully-protected images:

  1. Internationalisation
  2. Fix http to https
  3. General fixes
  4. Add structured data based on information provided on file description page according to Commons:Structured data/Modeling

Automatic or manually assisted: Automatic Fully supervised

Edit type (e.g. Continuous, daily, one time run): One time run, then on demand

Maximum edit rate (e.g. edits per minute): 5-10

Bot flag requested: (Y/N): Y

Sysop flag requested (administrator status): (Y/N): Y

Programming language(s): Bash + Python

Statement from the operator: I'm operating User:SchlurcherBot for already 10+ years, performing internationalization, https and general fixes. Over the last two years, notably also adding structured data to most files hosted here. Motivation for this request is to extend this activity to the most prominently placed and viewed images which are therefore auto-protected. Files that we place on the front page of any Wikipedia, e.g. to illustrate recent events, should follow our highest standards and have complete and accurate structured data. I expect reservations that I am not an administrator myself and hope that my dedication shown to the project so far will be taken into consideration instead, hopefully resulting in trust from the community that the sysop flag will only be used to edit permanently protected files as per the task description. From a technical perspective additional precautions were taken as follows: 1) The bot will run on a separate dedicated infrastructure independent from User:SchlurcherBot which is also planned to be offline when not needed, 2) the account uses two-factor authorization for login, and 3) API edits will be performed with oauth-authorization technically restricted to edit-changes only. The bot account will never perform any non-automatic edit or action. Looking forward to the discussion. Best regards, Schlurcher (talk) 08:06, 20 March 2021 (UTC)[reply]

Discussion

If I understand correctly, those auto-protected files are for a short period of time due to Main Page appearances, right? They're usually not protected for a good length of time before their appearance and after, at which point the bot should be able to edit them? ProcrastinatingReader (talk) 13:40, 20 March 2021 (UTC)[reply]

That is correct, but addresses only part of the problem. I noticed this during en:2021 storming of the United States Capitol, where we had a couple of pictures on the English Wikipedia front page on the same day so directly after upload. Structured data was added after they were removed from there. So, in the time they were most visible, there was no chance to add structured data. WMF is building more and more tools that seemingly rely on structured data, so we should make sure that our most visible images have them. For the same reasons we started adding structured data first to our excellent and valued images, prior to moving to all other. --Schlurcher (talk) 19:18, 20 March 2021 (UTC)[reply]
As for the few files protected for a longer duration (like logos), they seem to be static and so small in number that manual edit requests would be more effective. ProcrastinatingReader (talk) 13:44, 20 March 2021 (UTC)[reply]
Last time I checked there were about 500 permanently protected files. Years ago I made these edit requests when needed. In end of 2019 I also wanted to 'finish' my bot tasks and process all protected files. That time, I had checked all permanently protected files with my internationalization and https fixes, logged approximately 100 actions and asked at the administrators noticeboard for help. From the version history of User:SchlurcherBot/Blocked you can see that it took User:Majora almost 4h to complete this task. Two years later, I tried the same for structured data. You will see that 5 structured data related edit requests where not done yet (File talk:Article1 B.svg, File talk:Licence Art Libre.svg, File talk:AnonEditWarning.png, File talk:BlueLine.jpg, File talk:Phone.png). This might suggest that adding structured data manually is a tedious process and that our admins might want to focus on other aspects. I also think they should. --Schlurcher (talk) 19:18, 20 March 2021 (UTC)[reply]

Different opinions welcome, but I think we cannot grant sysop to a bot operated by a non admin. I'd appreciate if you could open an RFA first for your main account. --Krd 08:53, 21 March 2021 (UTC)[reply]

@Krd: Thanks. Apart from what is described in this request, I do not remember an occasion where I would have even wanted to perform any admin action. My involvement at Commons is almost exclusively focused around my bot work (and related activities), commenting on Commons:Bot requests and following Commons:Administrators' noticeboard. This has almost no overlap with expected work of admins. So, I see no use in opening an RFA for my main account. If the community thinks this is an requirement then I am ready to accept this and we can close this request. I would like to note though that api-access from the bot to admin related tools can be customized and will be restricted to edits only. Do you think it would be beneficial to cross-include this request at RFA to reach further groups from the community for feedback? Please let me know. --Schlurcher (talk) 16:49, 21 March 2021 (UTC)[reply]
I don't have a definite opinion, as I think there is no precedent. In any case I think that access to the admin tools, which includes access to deleted content, whether used or not, requires formal community support, i.e. RfA. Also, e.g. there is an activity policy in place for admins, and an admin bot run by a non-admin likely shouldn't be an exemption. I don't know how to resolve that. --Krd 17:33, 21 March 2021 (UTC)[reply]
@Krd: As I understand continuing to use {{Edit protected}} seems to be the way forward; all other alternatives would require me being an admin myself, which then invalidates this request here. User:AntiCompositeNumber has just now addressed all currently outstanding requests. Thank you.
This said, I would like to withdraw the original proposal and fully revise this request. There would still be benefit for me to have a second bot account: to test functionality within the scope of my current approved tasks for User:SchlurcherBot. Due to the constant edit rate there, it's getting complicated for me to see and supervise any new functionality edits. Within this revised scope all edits will be fully supervised. If no obvious flaws are identified the updates would be copied over for routine processing to User:SchlurcherBot. Hope that approval of this revised scope is rather straight forward ;-) --Schlurcher (talk) 18:28, 26 March 2021 (UTC)[reply]

Discussion on revised scope

I think it can easily be approved this way. Before we do so, one thought from my side. The auto-protected files are protected mostly against simple vandalism, which could be achieved as good as with the template protection level, and I think it would be no problem to assign template editor rights to a bot. Perhaps this would resolve it for at least a large part of your initial idea. Is that worth discussion further? --Krd 20:50, 27 March 2021 (UTC)[reply]

Most of the files that would fall under this request have cascade-protection, not file level protection. I have not found much about this on Commons, but according to En-Wikipedia cascade-protection is restricted to full protection. The template editor right seems specific to Commons, so I'm not sure if a cascade-protection with template editor right is technically possible. Do you/Does anyone know? If confirmed, I would start a proposal for further input. --Schlurcher (talk) 07:51, 30 March 2021 (UTC)[reply]
The protection happens at Commons only, so IMHO enwiki isn't much involved, and I don't see why a cascade protection shouldn't work also with different protection levels. I may be mistaken, that's just an idea how to address the initial intention with least effort. --Krd 08:13, 30 March 2021 (UTC)[reply]
I was referring to the enwiki description on how cascade protection works. I now found similar description on Commons:Protection_policy#Cascading_protection, which suggests (together with the linked tickets) that cascade-protection is not a form of inheriting the protection level, but simply puts each linked paged to full protection. --Schlurcher (talk) 20:24, 30 March 2021 (UTC)[reply]
I'm not sure this is true, but at least per Commons:Protection_policy#Cascading_protection "Cascading semiprotection is disabled, because it enabled non-sysops to fully protect pages by transcluding them." is a showstopper. I cannot really follow the logic, but I think discussing and changing that is then also too much effort. If nothing else arises, I think the bot request should be approved. --Krd 20:48, 30 March 2021 (UTC)[reply]

Approved. --Krd 17:40, 5 April 2021 (UTC)[reply]