Page MenuHomePhabricator

eBay to Commons: tool needed
Open, LowestPublic

Description

There are a great many images on eBay, of out-of-copyright works such as old postcards, ephemera, or paintings, that can legitimately be uploaded to Commons. A while after an eBay sale ends, its images disappear from view and are lost to us.

The task of uploading such images manually is laborious and time consuming, so much so as to be a deterrent.

It would be beneficial to have a tool or script that can automate as much of this task as possible, akin to Flickr-to-Commons.

Because the majority of images on eBay are still in copyright, use of the tool should be limited to (or rate-limited for everyone except) users who have applied for and been granted the ability to use it (such as is done with AWB, for example).

The tool should account for the fact that some offers on eBay include multiple images; the user should be able to select (or deselect) some or all of the images from a given offer for upload. It should work on any eBay domain (ebay.com, ebay.co.uk, etc.)

All edits made by the tool should use an identifying tag in the edit summary, and all images uploaded by it should be added to a category associated with the tool.

This may make a suitable project for a hackathon or student activity.

Event Timeline

Hi @Pigsonthewing, thanks for taking the time to report this! Do you plan to work on this at the next Hackathon and/or mentor this?

Do you plan to work on this at the next Hackathon and/or mentor this?

I'm happy to do the latter in a non-technical capacity, but I am not a coder.

I also see that you removed the projects 'Google-Summer-of-Code' and 'Wikimania-Hackathon-2021'. How do you suggest that a proposal such as this be tagged for the attention of people interested in such activities?

I often upload from eBay and fully agree with this, it's not rare that highly valuable images from eBay are lost forever because nobody bothered importing them. Such a tool should not be limited to eBay, there are many auction websites and there are many general websites, in my opinion we should have a tool like Flickr2Commons that can import from all websites, but if copyright issues from new users are concerned then limit it to users with a set number of non-deleted uploads or all users with 1000 (one-thousand) edits and upon request for those that fail to meet this threshold.
A general-purpose tool would really be handy, it's such a shame that we don't have a Community Tech Wishlist like the German-language Wikipedia has for Wikimedia Commons.
The tool should require manual review and not bot-review and users with a thousand or so upload could be bot-reviewed, or it could just rely on copyright © tags. Anyhow, I think that a bot should make a gallery per website (and/or per user) like with YouTube imports, and it could suggest pages to be imported to encourage more imports. It should be customisable so users can bulk import from multiple pages.

Odd, when I come with tool and/or feature suggestions they are closed as "not bugs".

we should have a tool like Flickr2Commons that can import from all websites

I doubt this is feasible, due to lack of structure.

Aklapper triaged this task as Lowest priority.Jul 19 2021, 9:52 PM

I also see that you removed the projects 'Google-Summer-of-Code' and 'Wikimania-Hackathon-2021'. How do you suggest that a proposal such as this be tagged for the attention of people interested in such activities?

Random proposals that lack mentoring capacity are out of scope for these tags; see their descriptions. I'm going to add Technical-Tool-Request.

I also see that you removed the projects 'Google-Summer-of-Code' and 'Wikimania-Hackathon-2021'. How do you suggest that a proposal such as this be tagged for the attention of people interested in such activities?

Random proposals that lack mentoring capacity are out of scope for these tags; see their descriptions. I'm going to add Technical-Tool-Request.

So this hasn't been included in the Wikimania-Hackathon-2021 Phabricator board, despite the instructions on [1] saying "To propose a project... Please add or create a task in the Projects column of the Phabricator board."?

I'm also not clear how this long-considered proposal qualifies as "random".

[1] https://wikimania.wikimedia.org/wiki/2021:Hackathon

Hi, I'll also quote the third item: "In the task description, add any information that will help new people join your project: additional links, what kind of help or skills you are looking for, [...]" In my understanding, project proposals imply that you plan to work and hack on these projects yourself during the Hackthon.
Maybe I misunderstood you, I am sorry in that case.

"add any information" clearly implies that such information is optional.

There is a significant disconnect between the guidance for making suggestions, as shown on the Hackathon page on the Wikimania wiki, and which also incudes the statement "everyone is free to suggest projects, tasks to work on, or sessions in the schedule") , and the reception of such proposals here on Phabricator. Whichever is correct, the other should change to match it.

I'm still not clear how people interested in 'Google-Summer-of-Code' and 'Wikimania-Hackathon-2021' are/ were expected to know about this proposal.

Clearly, there is no logical reason to expect that anyone proposing 'Google-Summer-of-Code' projects (for "a global program focused on bringing more student developers into open source software development" [1]) must "plan to work and hack on these projects yourself ".

[1] https://summerofcode.withgoogle.com/

@Pigsonthewing: Hi, anyone is free to share their ideas as Technical-Tool-Request tasks, and anyone is free to decide [not] to work on these tasks.

I'm still not clear how people interested in 'Google-Summer-of-Code' and 'Wikimania-Hackathon-2021' are/ were expected to know about this proposal.

In my understanding, as long as nobody is available to mentor, outreach program students or organizers do not need to know about this proposal, as it cannot be implemented due to lack of technical mentors. Mentors are very welcome to look into Technical-Tool-Request tasks (and other tasks) and tag them with an outreach program tag (e.g. 'Google-Summer-of-Code') if they are available as technical (co-)mentors (reviewing code, etc) for the next round of an outreach program.

As long as nobody plans to work on these projects at the Hackathon, any Hackathon attendees are welcome to pick a project task from Technical-Tool-Request or a project task from somewhere else when they are interested to work on that task at a Hackathon.

Basically: Every task has at least one person (the reporter) who thought it's a good idea if someone (the person or someone else) worked on their task. :) Following that logic, every open project task could be tagged with a Hackathon tag, as an attempt to get attention. And that would make using a Hackathon project tag useless. Simplified, Hackathon project tags are for project tasks, or hosting sessions that have at least one person who plans to actively do that work.
Does that help?
See also https://www.mediawiki.org/wiki/Bug_management/Development_prioritization

Odd, when I come with tool and/or feature suggestions they are closed as "not bugs".

@DonTrung: If you have specific examples in mind you'd like to look into, feel free to bring them up on https://www.mediawiki.org/wiki/Talk:Phabricator/Help - thanks!

Does that help?

No.

As I said (emphasis added for clarity): There is a significant disconnect between the guidance for making suggestions, as shown on the Hackathon page on the Wikimania wiki, and which also incudes the statement "everyone is free to suggest projects, tasks to work on, or sessions in the schedule") , and the reception of such proposals here on Phabricator. Whichever is correct, the other should change to match it.

I see, thanks. Looks like next year's Hackathon docs could be improved.

For the records, I've clarified the docs and hope they'll also serve as a base for next time, to avoid this misunderstanding.

Maybe it could be implemented in UploadWizard?

That would suit me, with suitable caveats about only uploading out-of-copyright material.

BTW, permissions could be aligned with those in T281599.

The benefit of using UploadWizard is that you don't need to implement an upload interface from scratch - UW already contains a gallery interface for selecting images from a list + an interface for adding metadata etc. My recollection (which you should take with a grain of salt - it's been almost a decade since I last worked on that codebase) is that changing the Flickr upload code to support other websites wouldn't be too hard, it just involves reimplementing the parts that call the Flickr API to get an account's image list / the full image URL.

The drawback is that the upload workflow would have to be fairly similar to that of Flickr (tag images with some sort of "review needed" template, and then on-wiki some sort of bot or manual process ensures that these really are compatible with Commons policy). Although if access is restricted, I suppose that's less of a concern.