User:HersfoldArbClerkBot

Current status: . The bot is currently running normally. (Last run: ) Hers fold  (t/a/c) 18:31, 9 January 2013 (UTC)

Non-administrators may also shut off the bot by posting a new comment to the bot's talk page.

Function
HersfoldArbClerkBot is a automated Arbitration Committee clerk that will periodically check evidence pages in open Arbitration cases for the lengths of evidence sections submitted to the cases. The Arbitration Committee (or rather, its clerks) enforce a length limit on these sections, which is often exceeded by editors participating in the case. This bot will keep running tallies on the word length of each section, as well as the number of diffs and other links posted in each section. These counts will be posted at the top of each section for the benefit of the submitted editor, the clerks, and Committee. If any section goes significantly over any of these limits, the bot will leave a notice on the submitting editor's talk page asking that they shorten their evidence. This bot will not modify submitted evidence in any way.

Future updates
To suggest an update to the bot, please contact Hersfold. Suggestions for updates should relate to the bot's current task; if you would like a bot to perform a totally different task, please make your request at WP:BOTREQ.


 * None currently

Completed updates

 * ✅: Basic word, diff, and link counter algorithms
 * ✅: Retrieve list of open cases and obtain data from evidence pages
 * ✅: Provide on-wiki list of all evidence section lengths in userspace
 * ✅: Set up the bot to read configuration data regarding the limits and tolerance level from a user subpage
 * ✅: Set up the bot to notify users exceeding limits outside of tolerance levels
 * ✅: Find a way to get the bot to run continuously (current plan is a scheduled task to run on my computer every two hours, with the computer rebooting itself every 24. :-/)
 * ✅: Set up the bot to make edits to each section to provide length counts using a user template
 * ✅: Set up bot to check talk page for messages between each case
 * ✅: Add an override configuration option for specific users in specific cases (suggested by NuclearWarfare)
 * ✅: Rig diff and link counts to discount content within templates and HTML code
 * ✅: Fix null edit issue
 * ✅: Bot doesn't seem to be handling hat/hab sections correctly, it didn't correct word or diff count for Chester Markel's evidence on the last run
 * ✅: Add (uninvolved) to the possibilities for marking one self as uninvolved
 * ✅: Find a better way to get the bot to run continuously
 * ✅: Bot does not correctly remove layered list formatting, i.e. multiple :'s, #'s, or *'s, or some combination thereof. (mid priority)
 * ✅: diff not recognized as a diff (gets stripped out with all other templates) (high priority)
 * ✅: Malformed section headers (i.e. "Evidence presented by User:foo") cause bot to crash
 * ✅: Bot returns a word count of 1 for blank evidence sections regardless of whitespace present.
 * : Set up the bot to maintain on-wiki logging pages in userspace - After further thought, this could cause problems, as the bot would be constantly editing the log page. Contributions and the on-wiki length reports should suffice, and logs will be kept on the local machine in case of a significant problem.
 * : While not breaking the bot, tables seem to throw off the word count and need to be handled better - While this does significantly throw off the word count, tables can be used to convey a great deal of information, on a level analogous to that provided by the amount of words the bot thinks the table contains. It will be very difficult to strip out table formatting while leaving the contents of the table intact, due to the way Mediawiki displays tables. As these are infrequently used at arbitration, this is not really worth putting effort into fixing just now. Should it become an issue, this can be revisited later.

Configuration and logging
While running, the bot will maintain logs of its actions on the WMF Labs server. The logs from trial runs will be posted here. In the event of a problem, these logs will be used to determine the cause of the problem and possible solutions. The bot may also be reconfigured using these pages by currently-serving Arbitrators and Committee Clerks.

/Configuration (Easier forms) - /Trial logs - /Length reports - /User Notice - /Length header - /IRC Help - Source Code (at WMF Labs)