Wikipedia

Wikipedia:Bots/Requests for approval/VIAFbot

< Wikipedia:Bots | Requests for approval
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.

VIAFbot

edit

Operator: Maximilianklein (talk · contribs · SUL · edit count · logs · page moves · block log · rights log · ANI search)

Time filed: 17:19, Friday August 17, 2012 (UTC)

Automatic, Supervised, or Manual: The bot is technically automatic, but logs cases for humans to check where it found conflicts between dewp and enwp.

Programming language(s): Python

Source code available: https://github.com/notconfusing/VIAFbot

Function overview: This bot takes a list of links from viaf.org to Wikipedia and reciprocates those links by way of the {{Authority control}} template after checking to see if the reciprocal link would agree with dewp's Normdaten template. Full details: Wikipedia:Authority control integration proposal


Links to relevant discussions (where appropriate): Passed RfC: Wikipedia:Authority control integration proposal/RFC and before that at the village pump

Edit period(s): It will be run one-time. A maintenance script could be developed later to refresh content when VIAF is refreshed.

Estimated number of pages affected: ~250,000

Exclusion compliant (Yes/No): No, this bot takes the list of articles from a file that's generated by viaf.org they are virtually all WP:BLPs.

Already has a bot flag (Yes/No): No

Function details: Takes this list Run's this switch: 

 Key 
  nd - normdaten; ac - authority control; vl - viaf->wikipedia link number
  (xx) - template does not exist; |xx| - has template but no viaf parameter; |(xx)| -> (xx) or |xx|
  
  1 - |(nd)|, (ac)
      AC template added VIAF parameter written.
  2 - |(nd)|, |ac| 
      VIAF parameter written.
  3 - |(nd)|, ac == vl 
      No writing necessary.
  4 - |(nd)|, ac != vl 
      Requires human attention. Nothing written.
  5 - nd , (ac), nd == vl.
      AC template added VIAF parameter written.
  6 - nd, (ac), nd != vl.
      Requires human attention. Nothing Written.
  7 - nd, |ac|, nd == vl.
      VIAF parameter written.
  8 - nd, |ac|, nd != vl.
      Requires human attention.
  9 - nd, ac , nd == ac == vl.
      No writing necessary.
  10- nd, ac, nd == ac != vl.
      Requires human attention. Nothing written.
  11- nd, ac, nd != ac, nd == vl.
      Requires human attention. Nothing written.
  12- nd, ac, nd != ac, ac == vl.
      Requires human attention. Nothing written.
  13- nd, ac, nd != ac != vl.
      Requires human attention. Nothing written.

Discussion

edit

Please review Wikipedia:Bot policy. In particular, note:

Code comments:

  • Why does it try to write Template:Db-meta? Just to test the error?
    • Right, that was just to test the error handling for locked pages. It was cruft, and is removed now.
  • Why does pageValidate follow 10 levels of redirects, when only 1 is valid?
    • I know sometimes that double redirects exist before they are cleaned, and some of these pages jumped around a bit, so I just set the maximum number of redirects to follow to 10, to make sure we're not in a loop. I can reduce it to 5 or even 4.
  • The bot does not need to be editing its status pages (via calls to logOnWiki and from exception handling all over the place) for every edit it tries to make. Log them locally if at all possible. If it is really necessary to write them to the wiki, batch them to make one edit every few hours.
    • The reason I am logging onWiki is that this information is useful to those other editors, also on German Wikipedia who are attempting to improve VIAF representation. Each time a dewp/enwp conflict is detected I believe that information needs to be reviewed - likewise where dewp disagreed with enwp & viaf.org. I refactored the code to batch the edits and write once there are 100 errors or conflicts of every variety. The plan is for humans to be able to follow along and look at conflicts as the bot is running, so I feel that every 100 is a good compromise.
  • Similarly, the bot writes its stats far too often. Rather than posting to the wiki every 100 edits (which probably works out to once every few minutes), write the stats after a reasonable period of time has elapsed.

I'll leave any comments on whether there are more efficient ways to use pywikipedia to others, as I'm not very familiar with that framework. Anomie⚔ 14:39, 18 August 2012 (UTC)[reply]

Some more code review:
  • I don't think it's a safe idea to do from wikipedia import *, though I don't see any conflicts.
    • Done
  • Why are you using the replace.py class? Isn't it simpler (and faster) to just call wikipedia.replaceExcept() directly?
    • Done.
  • Same with add_text.py (using add_text allows for text to be put before interwiki links as Max explained on IRC)
    • Correct add_text.py is nice because it puts the text after the article and before the interwiki and category links, which is very nice.
  • I'm not sure what the point of pageValidate is. Couldn't you just use:
if page.isRedirectPage():
    return page.getRedirectTarget()
else:
    return page
LegoKontribsTalkM 20:11, 18 August 2012 (UTC)[reply]

There's a clear consensus for this task, and it appears all questions regarding the code have been answered. Let's see a trial to evaluate results and iron out any bugs. Approved for trial (35 edits). Please provide a link to the relevant contributions and/or diffs when the trial is complete. — madman 04:10, 2 September 2012 (UTC)[reply]

Thanks for the trial. I am busy but I hope to have the trial completed by Oct 1st. Maximilianklein (talk) 10:18, 6 September 2012 (UTC)[reply]

Trial complete. See [[1]]. Here's the log from the run:

Articles with VIAF numbers already on DEWP

edit
  • 15:53, 13 September 2012 . A.C. Reed added VIAF number 76504615
  • 15:56, 13 September 2012 . Abel Decaux added VIAF number 34650276
  • 15:56, 13 September 2012 . Abel Ferrara added VIAF number 12486003
  • 15:57, 13 September 2012 . Abram Ioffe added VIAF number 10024694
  • 15:58, 13 September 2012 . Adalbert Krüger added VIAF number 30290947

Articles with VIAF numbers not already on DEWP

edit

Conflicts of Type 6

edit
  • 15:56, 13 September 2012 . Abdoulaye Mamani requires human attention: nd != vl . VIAF number: 56613360
  • 15:56, 13 September 2012 . Abdülaziz requires human attention: nd != vl . VIAF number: 122114473

This will be reviewed when possible. You really need to make sure you're never editing under the bot's account when you're not performing an automated task, though. Making human edits under a bot account is an excellent way to get the account blocked. — madman 00:14, 14 September 2012 (UTC)[reply]

Understood. I'm very sorry for doing that. There may want to be better guidlines about how to prototype in the pre-trial period. Or a note on not protoyping in the pre-trail period. Maximilianklein (talk) 11:51, 14 September 2012 (UTC)[reply]
I have no idea what you mean by prototyping, but the Wikipedia:Bot policy clearly states "Bot accounts should not be used for contributions that do not fall within the scope of the bot's designated tasks. In particular, bot operators should not use a bot account to respond to messages related to the bot. Bot operators may wish to redirect a bot account's discussion page to their own." — madman 15:07, 14 September 2012 (UTC)[reply]
Hmmm. Everything looks good except the VIAF record for Abraham Kaplan. This may be a case of "garbage in, garbage out" as the match does make sense; some of the works are correct on the VIAF record. It still may be worth filtering records like this out, however, as the birth years don't match. — madman 23:21, 17 September 2012 (UTC)[reply]
Yes, there maybe be some errors. As it says on WP:VIAF the matching was done with this code that does attempt to match names, birth and death dates, works, and some other factors. Checking by hand it's about 98% accurate. In the RfC it was acknowledged that even with these imperfections there was consensus to go ahead. Maximilianklein (talk) 16:20, 18 September 2012 (UTC)[reply]
{{BAG assistance needed}} It's been one week since the last comment on this bot discussion. Maximilianklein (talk) 00:14, 27 September 2012 (UTC)[reply]
 Approved. — madman 16:09, 27 September 2012 (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.
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia:Bots/Requests_for_approval/VIAFbot&oldid=1131473481"