Wikipedia:Bots/Requests for approval/EranBot
- 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 Approved.
Operator: ערן (talk · contribs · SUL · edit count · logs · page moves · block log · rights log · ANI search) Was developed based on the request of User:Jmh649 among others.
Time filed: 01:24, Friday August 22, 2014 (UTC)
Automatic edits but not to main space
Programming language(s): Python
Source code available: source in github
Function overview: This bot submits newly added text to medical articles to the iThenticate API which determines if other sources are similar to it. This information is than placed on this page Wikipedia:MED/Copyright for human followup. The bot run on period basis.
- Links to relevant discussions (where appropriate): An entire project surrounds the creation of this bot with many subpages. This can be seen here Wikipedia:Turnitin. A memorandum of understanding has been signed with Turnitin. User:Ocaasi can send it to anyone who wishes it
Edit period(s): A few times per day
Estimated number of pages affected: 1
Exclusion compliant (Yes/No): N/A
Already has a bot flag (Yes/No): No (using autopatrolled
for trial)
Function details: Bot only changes one page and that is Wikipedia:MED/Copyright
Discussion
[edit]As this bot does not edit mainspace and their is no plan for it to ever edit mainspace the discussion at Wikimania was that it does not need official approval to run but should get it eventually. Happy to provide more details and to be corrected if this was a misunderstanding. Would appreciate the help of those who know the bot approval process better than I in correcting any malformed answers. Doc James (talk · contribs · email) (if I write on your page reply on mine) 01:29, 22 August 2014 (UTC)[reply]
- All bot edits are ultimately the responsibility of the operator, but the updates are being done by others.— xaosflux Talk 03:00, 22 August 2014 (UTC)[reply]
- What is your criteria to determine if an article is a "medical article"? — xaosflux Talk 03:07, 22 August 2014 (UTC)[reply]
- It is tagged with the WP:MED tag on the talk page. Doc James (talk · contribs · email) (if I write on your page reply on mine) 04:42, 22 August 2014 (UTC)[reply]
- Can you provide the current, approximate, count of these ? — xaosflux Talk 15:44, 22 August 2014 (UTC)[reply]
- I assume that means talk pages which transclude {{WikiProject Medicine}}. There are over 25 thousand transclusions of that template. – Wbm1058 (talk) 13:24, 11 September 2014 (UTC)[reply]
- Can you provide the current, approximate, count of these ? — xaosflux Talk 15:44, 22 August 2014 (UTC)[reply]
- It is tagged with the WP:MED tag on the talk page. Doc James (talk · contribs · email) (if I write on your page reply on mine) 04:42, 22 August 2014 (UTC)[reply]
- Do you plan on only look at current articles in whole, certain diffs, or every diff (even ones already reverted?) — xaosflux Talk 03:23, 22 August 2014 (UTC)[reply]
- Do you plan to notify editors that you are creating reports about their edits? — xaosflux Talk 03:26, 22 August 2014 (UTC)[reply]
- Only if their are concerns. Doc James (talk · contribs · email) (if I write on your page reply on mine) 04:42, 22 August 2014 (UTC)[reply]
- Currently done manually by Jmh649 or other patrollers of the report page. However, the bot may add [in future] signature (~~~~) to its edits and then editors will be notified using the notification system ("someone mentioned you..."). Eran (talk) 07:50, 22 August 2014 (UTC)[reply]
- Looking for additional community feedback on this, conducting automated investigations of editors without involving them may be problematic. — xaosflux Talk 12:40, 22 August 2014 (UTC)[reply]
- (If this aspect has already had a community discussion, please provide a link to it here). — xaosflux Talk 15:44, 22 August 2014 (UTC)[reply]
- Looking for additional community feedback on this, conducting automated investigations of editors without involving them may be problematic. — xaosflux Talk 12:40, 22 August 2014 (UTC)[reply]
- Currently done manually by Jmh649 or other patrollers of the report page. However, the bot may add [in future] signature (~~~~) to its edits and then editors will be notified using the notification system ("someone mentioned you..."). Eran (talk) 07:50, 22 August 2014 (UTC)[reply]
- Only if their are concerns. Doc James (talk · contribs · email) (if I write on your page reply on mine) 04:42, 22 August 2014 (UTC)[reply]
Resolved discussion over initial trial pages
|
---|
Bot blocked pending the operator's response here. — xaosflux Talk 03:58, 22 August 2014 (UTC)[reply]
|
((BotTrial|edits=100|days=10|userspace=yes)) Initial approval for user-space technical trial, there are some community-interaction components that need further exploration, but a trial is a good way to gather the type of output that is expected. — xaosflux Talk 12:36, 22 August 2014 (UTC)[reply]
Consensus from WP:MED for this bot is [1]. Doc James (talk · contribs · email) (if I write on your page reply on mine) 02:01, 27 August 2014 (UTC)[reply]
Jmh649 did you finish bot trial? Is anything to be done to reduce false positives? Du you want one more trial? -- Magioladitis (talk) 16:25, 8 September 2014 (UTC)[reply]
- Efforts to reduce false positives are coming along nicely. We now have a blacklist that is machine readable. As this list is improved false positives will decrease. Other improvements are listed here [2] A 50/50 rate of false positives to true positives I deem to be a huge success.
- Plans are not to have the bot do anything other than create a list of possible concerning edits. There is no discussion of having automated warnings or editing. That would be years out if at all.
- At this point I hope the bot could be approved to operate on medical articles going forwards ( and expanded to pharmacology and anatomy articles ). I would want proof of a community willing to follow up on concerns if it was to expand to other topic areas. Doc James (talk · contribs · email) (if I write on your page reply on mine) 10:38, 10 September 2014 (UTC)[reply]
The trial has been underway at two locations:
- Wikipedia:MED/Copyright, where trials began on August 10, and continued until August 22
- User:EranBot/Copyright, where the trial has been ongoing since August 22
I suggest that the next step should be to approve for an extended trial at Wikipedia:MED/Copyright, and merge User:EranBot/Copyright into that page. I learned of this trial from reading The Signpost here, so that high-profile notice should draw more attention to this trial. – Wbm1058 (talk) 14:13, 11 September 2014 (UTC)[reply]
((BotExtendedTrial|edits=100|days=10|userspace=yes)) Approved for extended trial. Please next time post a BotTrialCompleted after bot trial is done. -- Magioladitis (talk) 14:38, 14 September 2014 (UTC)[reply]
Doc James when bot trial ends please notify using {{BotTrialComplete}} and provide diffs. -- Magioladitis (talk) 08:21, 15 September 2014 (UTC)[reply]
- Provide diffs of what exactly? Doc James (talk · contribs · email) (if I write on your page reply on mine) 06:28, 16 September 2014 (UTC)[reply]
- Whichever edits (or other log entries) the bot made as part of its trial. In this case, you could probably just link to Special:Contributions/EranBot with appropriate offset and limit URL parameters since the bot hasn't made any edits other than for this trial, or maybe to the history of User:EranBot/Copyright with appropriate offset and limit. Anomie⚔ 11:26, 16 September 2014 (UTC)[reply]
- Provide diffs of what exactly? Doc James (talk · contribs · email) (if I write on your page reply on mine) 06:28, 16 September 2014 (UTC)[reply]
Doc James is bot trial done? It's been 20 days. -- Magioladitis (talk) 22:14, 8 October 2014 (UTC)[reply]
- Trial complete. Bots working wellish. Of the last 36 edits 18 have been true positive copyright issues. 18 have been false positive copyright issues. Has allowed us to keep a fair number of copy and paste issues out of Wikipedia and given us the ability to educate new editors. More tweaking is needed but that is not unsurprising.
- We have two long term editors who are doing the follow up of all edits that are flagged. The number of concerns brought to our attention is not unreasonable for us to handle. We are getting 2 to 8 concerning edits per day from ~30,000 articles.
- Before this bot were to expand to other topic areas there would first need to be dedicated editors who wish to do the review. The list of sites that the bot ignores as they are known mirrors of Wikipedia has grown nicely [User:EranBot/Copyright/Blacklist]
- Unclear what you are wishing before bot is approved to function in an ongoing manner? Doc James (talk · contribs · email) (if I write on your page reply on mine) 00:47, 9 October 2014 (UTC)[reply]
- Doc James If you are satisfied with the result and the minor issues can be handled I will aproove the bot. -- Magioladitis (talk) 19:45, 23 October 2014 (UTC)[reply]
Approved. -- Magioladitis (talk) 09:02, 24 October 2014 (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.