Commons:Bots/Requests/DieBucheBot

DieBucheBot (talk · contribs)

Operator: DieBuche (talk)

Bot's tasks for which permission is being sought: This is a bot for various smaller maintenance tasks. It has a flag on de.wiki & it.wiki and about 13k edits on commons. I used it before to tag suspected thumbnail uploads, right now it is tagging images having only "no original description" as description with {{Description missing}}. Afterwords it will try to fix those images whose description was lost by commonshelper, but which is still in the upload log.

It also applies the fixes to allow localization (own -> {{Own}} etc.) at each edit.

Automatic or manually assisted: Automatic

Edit type (e.g. Continuous, daily, one time run): one time runs

Maximum edit rate (eg edits per minute): 8/min

Bot flag requested: (Y/N): Yes

Programming language(s): pywikipedia (python)

Source Can be found here

Discussion

  • I think will be good idea to create more descriptive edit summaries. Minor fixes looks too general for me. --EugeneZelenko (talk) 15:27, 20 May 2010 (UTC)[reply]
  • Looks fine for me. Even the Minor fixes is ok, as the main main task is another one. --Schlurcher (talk) 16:23, 20 May 2010 (UTC)[reply]
  • @Eugene: minor fixes links to this. It's a collection of replacement, similar to those which Slobot and other use, to allow for the internationalisation of pages. For example == Licensing|License information|{{int:license-header}} == is changed to == {{int:license}} ==.

@all I just ran a batch of 50 pages to show how the bot restores lost descriptions due to an CommonsHelper bug--DieBuche (talk) 17:20, 20 May 2010 (UTC)[reply]

Looks good, I checked a couple of edits. No objections --Schlurcher (talk) 18:41, 20 May 2010 (UTC)[reply]
Here's another thing I could do: Go through missing desc. cat. & run magnus "what is that" script on them. This would be a local clone to reduce the load on the toolserver. If the result is non-empty, place it onto the image in addition to a short template similar to {{BotMoveToCommons}} (maybe something like {{BotFetchedDescription}}). If there is already a {{BotMoveToCommons}} present, we don't add our template, assuming that when a user checks the bot move, he will also look at the descriptions. This is not yet coded, but shouldn't be to hard. --DieBuche (talk) 12:23, 21 May 2010 (UTC)[reply]
One more; I got list of description-less roadsigns, to which I could add a bot generated description like:"Sign of the A1213 running from Town, UK to City, UK", or is that just busy-work?--DieBuche (talk) 12:26, 21 May 2010 (UTC)[reply]
Here's a possible template:
Moved
This image description was added by a Bot. It may be inaccurate and should not be considered reliable until it has been reviewed and any needed corrections have been made. Once the review has been completed, this template should be removed. Check now!

English | +/−

--DieBuche (talk) 15:19, 21 May 2010 (UTC)[reply]

Have a look at my last 20 edits to see how the bot adds Descriptions from "What is that"--DieBuche (talk) 19:47, 21 May 2010 (UTC)[reply]
You don't have to add {{Description missing}}, if you just make the field empty {{Information}} will show this.
You don't have to add a template to indicate a bot added the info, it will probably stay on there forever. Just add the source of the info as a comment (like the this is not a vote part at the top of this section) so the source is clear. Multichill (talk) 20:18, 22 May 2010 (UTC)[reply]
I agree with Multichill. Kameraad Pjotr 10:29, 23 May 2010 (UTC)[reply]
Ok, I'll just add smtl. like <!-- Bot generated description -->--DieBuche (talk) 13:58, 23 May 2010 (UTC)[reply]
With a source where the bot got the info from so people can traceback errors or debug it. Multichill (talk) 14:52, 23 May 2010 (UTC)[reply]
How about <!-- Description fetched by DieBucheBot using whatisthat.php-->? But i feel that's a bit long--DieBuche (talk) 16:34, 23 May 2010 (UTC)[reply]
Personally I think comments in file description pages aren't of much use. They are hard to find and read. I'd rather see a compact template. -- User:Docu at 15:16, 24 May 2010 (UTC)[reply]
I actually tend to agree with multichill. A lot of pages are already "spammed" with templates & reviewing the added descriptions, i find that they are quite accurate.--DieBuche (talk) 15:47, 24 May 2010 (UTC)[reply]
Well, your sample wasn't exactly "compact". Most other templates get added by Multichill's bot. I think he even converted some comments to templates. -- User:Docu at 15:57, 24 May 2010 (UTC)[reply]
  • You might want log the imported description to the bot's edit summary. -- User:Docu at 15:16, 24 May 2010 (UTC)[reply]
  • Is it correct that the language it picks is pretty much random? -- User:Docu at 15:16, 24 May 2010 (UTC)[reply]
    • The language it picks is the language of the local wiki, assuming that most text in a specific wiki is in that language. Were there any edits where the language was obviously false? Is there a maximum length of the edit summary?--DieBuche (talk) 15:37, 24 May 2010 (UTC)[reply]
      • I was under the impression that from several possibilities it picked just one language, but apparently, the image I looked at was described properly in just one language, e.g. this has three. -- User:Docu at 15:57, 24 May 2010 (UTC)[reply]
        • I rewrote the script to find more cases; BARI.jpg would now get this, 6 instead of three:
          Eesti: Bari Ülikooli peahoone.
          Íslenska: Háskólinn í Bari.
          Italiano: Il palazzo dell'Ateneo dell'Università, in piazza Umberto I. Universita.
          Norsk bokmål: Universitetet i Bari.
          Srpskohrvatski / српскохрватски: Univerzitet u Bariju - zgrada Ateneo.
          Српски / srpski: Универзитет у Барију - зграда Атенео.
          --DieBuche (talk) 17:26, 24 May 2010 (UTC)[reply]
✓ Approved as there seem to be no outstanding major concerns. –Juliancolton | Talk 12:04, 3 June 2010 (UTC)[reply]

In case anyone is still watching this page; Here are the stats of the first run:

First run

  • Images analyzed: 70306
  • Description(s) found for: 17065 images
  • Average number of languages per Image: 1,68

  • Total descriptions: 28225 ± 900

--DieBuche (talk) 12:31, 22 June 2010 (UTC)[reply]

Please don't remove __NOTOC__, it's there for a reason.
These are added on default by Commons Helper, even if there are less than four headings. Other bots seem to remove them as well--DieBuche (talk) 13:46, 23 June 2010 (UTC)[reply]
The edit summary isn't very descriptive, could you improve this? Multichill (talk) 13:32, 22 June 2010 (UTC)[reply]
"Those were just some final cleanups (images which had a desc but the string "no original desc")--DieBuche (talk) 13:46, 23 June 2010 (UTC)[reply]