Commons:Bots/Requests/Frettiebot
Frettiebot (talk · contribs)
Operator: Frettie (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information)
This bot adds Wikidata Infobox to czech community pages (by Mike Peel script). Sometime there will be other tasks by czech communnity. I am requesting for bot flag, because my bot (Frettiebot) was blocked after few edits, its a mistake, because at Czech language Wikipedia or Wikidata i have bot flag and it is ok for thousands of edits and there is not at commons, unfortunately, sorry.
Bot's tasks for which permission is being sought: Wikidata Infobox for czech articles
Automatic or manually assisted: Half automatic, i manual start my bot and looks, which its do.
Edit type (e.g. Continuous, daily, one time run): One time / continous
Maximum edit rate (e.g. edits per minute): 30 edits per minute
Bot flag requested: (Y/N): Y
Programming language(s): Python
Frettie (talk) 16:16, 9 May 2018 (UTC)
Discussion
Comment This isn't a problem from my perspective, but I should note that User:Pi bot will run through these categories eventually. It's currently going through Category:Art as a test of a big category, and I'll start it from Category:CommonsRoot soon (maybe next week), but I'm focusing on adding ~600,000 commons sitelinks through d:Wikidata:Requests for permissions/Bot/Pi bot 2 first. Thanks. Mike Peel (talk) 16:54, 9 May 2018 (UTC)
- Perfect, good work!--Frettie (talk) 18:57, 9 May 2018 (UTC)
- Just to note: Pi bot's run to add the Commons sitelinks on Wikidata is now complete, and it's now focusing on infoboxes here, starting from CommonsRoot. Thanks. Mike Peel (talk) 23:17, 11 May 2018 (UTC)
- Perfect, good work!--Frettie (talk) 18:57, 9 May 2018 (UTC)
- (Edit conflict) Does this run the exact same script as Commons:Bots/Requests/Pi_bot_1. If so, could you explain the rationale fot having multiple bots for the same task? If not, could you explain the differences? --Zhuyifei1999 (talk) 16:55, 9 May 2018 (UTC)
- If there is similar bots for one thing, there will be faster. There is not differences.--Frettie (talk) 18:57, 9 May 2018 (UTC)
- If there happens to be some bug in the script, will you be able to fix them? --Zhuyifei1999 (talk) 20:07, 9 May 2018 (UTC)
- I think yes, but if there is mistake in Mike's script, there will be corrected by Mike and i adopt this again. --Frettie (talk) 10:08, 11 May 2018 (UTC)
- If there is similar bots for one thing, there will be faster. There is not differences.--Frettie (talk) 18:57, 9 May 2018 (UTC)
- The general expectation is that you are able to correct mistakes in Mike's script if necessary. Please confirm. --Schlurcher (talk) 17:28, 11 May 2018 (UTC)
- As a practical test, I've made a few changes to the script recently, @Frettie: can you explain what I've changed and why? Thanks. Mike Peel (talk) 23:17, 11 May 2018 (UTC)
- I like your original python code, the targetcats, which was filled in cycle - its interesting. Now you are set seen categories (like a history view) and "what all i want to browse in future from my root category". And while there is some active category, which is not "seen" (because seen category is set only, which become "active" for next iteration), go through this category. Maybe the for cycle with subcategory generator will be slow if there will be too much categories (but its ok, too fast run is not good in this case).--Frettie (talk) 16:35, 12 May 2018 (UTC)
- The new category walker is mostly designed for speed, since checking if a given subcategory is included in a list of categories is much slower than checking if it's in a set - although I also preferred the older version as it was more straightforward! I also changed a few things in the 'addtemplate' function, though, can you see why? Thanks. Mike Peel (talk) 23:31, 14 May 2018 (UTC)
- Hi Mike, there is check if wikidata item is exist, if not, "skipping" next lines (return 0), and this is earlier. There is not processed list without category main topic by target link. Return 0/1 helps to compute count of modified categories. Sorry for late reaction.--Frettie (talk) 08:25, 21 May 2018 (UTC)
- The new category walker is mostly designed for speed, since checking if a given subcategory is included in a list of categories is much slower than checking if it's in a set - although I also preferred the older version as it was more straightforward! I also changed a few things in the 'addtemplate' function, though, can you see why? Thanks. Mike Peel (talk) 23:31, 14 May 2018 (UTC)
- I like your original python code, the targetcats, which was filled in cycle - its interesting. Now you are set seen categories (like a history view) and "what all i want to browse in future from my root category". And while there is some active category, which is not "seen" (because seen category is set only, which become "active" for next iteration), go through this category. Maybe the for cycle with subcategory generator will be slow if there will be too much categories (but its ok, too fast run is not good in this case).--Frettie (talk) 16:35, 12 May 2018 (UTC)
- As a practical test, I've made a few changes to the script recently, @Frettie: can you explain what I've changed and why? Thanks. Mike Peel (talk) 23:17, 11 May 2018 (UTC)
- The general expectation is that you are able to correct mistakes in Mike's script if necessary. Please confirm. --Schlurcher (talk) 17:28, 11 May 2018 (UTC)
- I generally don't mind multiple bots for the same task if they are programmed to some extend independently. This will have the merit that they can supplement each other. As I understand this is a second bot running exactly the same script. There is less benefit in having this as long as the other bot is well maintained and running. During the request for Commons:Bots/Requests/Pi_bot_1 we allowed for a higher edit rate than normal to accommodate for the number of edits. So, please provide additional rational except they will be faster together to support this request. Also note that during Commons:Bots/Requests/Pi_bot_1 there was some consideration on the server load these edits might put on Wikidata (not Commons). With two independent bots acting this will be more challenging to monitor. --Schlurcher (talk) 23:33, 9 May 2018 (UTC)
- Yes, there will by challenging to monitor this, but (maybe) it's better, to edit some known categories (like czech themes).--Frettie (talk) 10:08, 11 May 2018 (UTC)
- Do you see any concerns with edits from User:Pi bot that require deeper understanding of the categories? --Schlurcher (talk) 17:28, 11 May 2018 (UTC)
- To clarify, it was the server load increase due to refreshing pages on Commons that use Wikidata information that I was worried about, so that's somewhere between server load on commons and on wikidata. The main limitation I've found with the bot code so far is that it takes time to work through the categories and find ones to add the infobox to (since it checks each category in series), so it helps to have multiple copies running with different target categories. Plus, it helps to have someone else here that knows how to run the code in case I'm not available in the future (since it'll be useful to keep running this after the main run to catch new categories that the infobox can be added to). Thanks. Mike Peel (talk) 23:17, 11 May 2018 (UTC)
- I generally support multiple bot on the same task for positive redundancy, but only if the code runs slowly and is certain to avoid conflicts. Just running it twice to double the speed doesn't appear as good idea to me. If I'm mistaken, please advise. --Krd 05:42, 27 May 2018 (UTC)
- @Frettie: ? --Krd 07:54, 12 June 2018 (UTC)
- @Krd: Twice with other root category - sleeping is default in pywikibot.--Frettie (talk) 10:17, 12 June 2018 (UTC)
I'm going to call this approved now. Please take care that no problems arise. Thank you. --Krd 05:50, 14 June 2018 (UTC)