Wikipedia:Bots/Requests for approval/TestEditBot

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section. The result of the discussion was

Approved.

TestEditBot

Operator: tj9991 (talk | contribs)

Automatic or Manually Assisted: Automatic, unsupervised

Programming Language(s): Python (wikipedia module)

Function Summary: Constantly monitor main namespace for short edits which contain nothing other than a trip-phrase (such as '''Bold Text'''). Will not revert the same article twice in a day in case of false positives.

Edit period(s) (e.g. Continuous, daily, one time run): Continuous

Already has a bot flag (Y/N): No

Function Details:

Periodically check for minor changes in the main namespace using the IRC recent changes channel
Check if it matches a pre-defined list of common test edits and make sure it is the sole changed content (other than newlines) so it doesnt interfere with useful additions
Revert the edit the user made if it seems likely to be a test edit and either warn them on their talk page, or report if final warning has been given

Discussion

How will the diff be loaded? As far as I'm aware the IRC feed doesn't include diffs, only links to diffs. --Nn123645 (talk) 18:28, 26 July 2008 (UTC)[reply]

The diff is loaded using standard methods with the wikipedia module. Specifically, a diff notification shows up in the IRC feed, the script loads the page which was modified and checks if it is interesting (main namespace), fetches the version history for diffing, diffs the last version to the new version, computes if it matches the short-edit threshold, and then checks against all of the trip-phrases. tj9991 (talk | contribs) 18:40, 26 July 2008 (UTC)[reply]

How does this bot improve over ClueBot? BJ^Talk 18:32, 26 July 2008 (UTC)[reply]

Because of the pinpointed task this bot will perform, it should identify and take action against a test edit faster than ClueBot. Through my Huggle usage I have seen ClueBot either completely ignore blatant test edits or delay further than a few seconds, in which a person doing counter-vandalism work gets to it. I've wrote the bot in a very concise way towards diffing and checking the edit, to be as speedy as possible. tj9991 (talk | contribs) 18:40, 26 July 2008 (UTC)[reply]

Are you using Pywikipedia or custom interface code? BJ^Talk 18:48, 26 July 2008 (UTC)[reply]

I am using pywikipedia. tj9991 (talk | contribs) 18:51, 26 July 2008 (UTC)[reply]

Are you aware that I am already running a bot that performs a similar task? User:AmeliorationBot (BRFA). Also, are you aware that there are instances where bold text italic text etc. are added legitimately into articles? ~ Ameliorate _{U T C} @ 02:04, 27 July 2008 (UTC)[reply]

I was not aware that you ran a bot with a similar goal, and I was aware that there will be a small amount of legitimate cases for the trip-phrases. I believe that the small amount cases where the terms are actually being put to good use will not render the usefulness of quickly reverting the many test edits that pass through every day. I believe that when I have the bot set up and functioning it will streamline the process because of the absence of any other processing work. tj9991 (talk | contribs) 02:15, 27 July 2008 (UTC)[reply]

This sounds pretty good to me, and, redundancy is not a problem IMO. You may want to consider adding some sort of on-wiki or off-wiki title whitelist, which the bot would ignore. You also may want to do something similar for editors, etc. SQL ^{Query me!} 06:59, 27 July 2008 (UTC)[reply]

This will actually render my bot useless, however that is because it performs my bot's task faster and more efficiently.

Just some ideas; to (help) prevent mistakes my bot ignores certain word combinations, such as: "bold text denotes", "in bold text", "bold text means", "bold text indicates" (and the same again for italic text), perhaps you could implement that? Also, what does your bot determine to be a test edit? I assume it covers the single button presses such as "bold text", "italic text", "[[Image:Example.jpg]]" etc., does it also find combinations, such as "<blockquote> {| class="wikitable" |Block quote |} </blockquote>"? ~ Ameliorate _{U T C} @ 09:18, 27 July 2008 (UTC)[reply]

Actually, it seems to me that this task forcuses only on edits that only add "'''Bold text'''", whereas yours does more things. I can't speak for the author, obviously, but, that's how it reads to me. SQL ^{Query me!} 09:54, 27 July 2008 (UTC)[reply]

I believe SQL Is correct. From what I've read about AmeIiorate's bot, it also has functions which go further than simple trip words. And to answer AmeIiorate's question: It does not look for combinations. The bot is designed specifically for small test edits in order to keep false positives to an absolute minimum. I will also take your ideas in to consideration, thanks. tj9991 (talk | contribs) 12:04, 27 July 2008 (UTC)[reply]

Will the list of pre-defined test edits be public? If so, where? And will this bot be limited to a certain list or will its approval apply to any and all future test edits it chooses to look for and revert? --MZMcBride (talk) 13:36, 27 July 2008 (UTC)[reply]

You can view the list yourself by clicking on each of the buttons above the editing text box. The bot only searches for the default text they insert. And no, the list will not change without me asking for permission first. tj9991 (talk | contribs) 14:52, 27 July 2008 (UTC)[reply]

Let's see how well this works. Approved for trial (25 edits or 1 days). Please provide a link to the relevant contributions and/or diffs when the trial is complete. Leaving up here for discussion. BJ^Talk 18:20, 27 July 2008 (UTC)[reply]

I'd like to note that the first case of the bot detecting a test edit has taken place. Relevant diffs: [1] (article), [2] (user talk page). The bot is currently operating on my home cable connection, and will pick up speed when I have an account on the toolserver. tj9991 (talk | contribs) 21:20, 27 July 2008 (UTC)[reply]

Done I've disabled the bot until a BAG member informs me of what happens next. tj9991 (talk | contribs) 10:33, 28 July 2008 (UTC)[reply]

Shouldn't it be issuing the warnings using the {{uw-test}} series rather than {{uw-vandalism}}? —Ilmari Karonen (talk) 05:36, 28 July 2008 (UTC)[reply]

You're right. I was using an anti-vandalism bot for reference and never even thought about uw-test. I've changed it over, thanks. tj9991 (talk | contribs) 10:41, 28 July 2008 (UTC)[reply]

Approved. Looks good, bot approved --uǝʌǝs ʎʇɹoɟ ʇs(st47) 21:44, 28 July 2008 (UTC)[reply]

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.