Wikidata:Requests for permissions/Bot/NicereddyBot 6
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
- Approved --Lymantria (talk) 07:48, 28 March 2021 (UTC)[reply]
NicereddyBot 6 edit
NicereddyBot (talk • contribs • new items • new lexemes • SUL • Block log • User rights log • User rights • xtools)
Operator: Nicereddy (talk • contribs • logs)
Task/s: Add StrategyWiki ID (P9075) to games based on data import from PCGamingWiki (Q17013880) (via PCGamingWiki ID (P6337)). Add Internet Game Database numeric game ID (P9043) to games as qualifiers on Internet Game Database game ID (P5794) using the IGDB API.
Code: I write my bot in Ruby. The first script for StrategyWiki is very similar to the script I wrote for importing WineHQ IDs a while ago. The second is a bit different from most of my other stuff, but still pretty standard overall: StrategyWiki script, IGDB Numeric ID Script
Function details: There are 1800 StrategyWiki pages linked on PCGamingWiki, and we have around 8000 PCGW IDs in Wikidata. If I had to estimate, I'd say somewhere between 1200-1500 StrategyWiki pages will be imported into Wikidata via my script, with fairly high accuracy since the data is all curated manually by the PCGW editor community. For IGDB numeric game IDs, I anticipate that I'll add qualifiers for >99% of all existing IGDB game IDs, so it'll be roughly equal to the number of Internet Game Database game ID (P5794) usages (currently 7700). The only ones that wouldn't get numeric IDs would be invalid IGDB game IDs, but there shouldn't be many of those.
I already tested both scripts using my own user account (see my Contributions page) to import 30-50ish IDs each. The IGDB numeric ID test import was done on January 23rd between and 4:35AM and 4:41AM UTC, and the StrategyWiki test import was done on January 23rd between 5:00AM and 5:08AM UTC.
You'll note that I had imported some bad data at first (at about 4:15AM) on Alundra (Q861322) specifically. This was because the script originally didn't handle items where there was more than one IGDB ID. I resolved that issue and now apply the correct numeric ID to each claim correctly.
I also had a problem with the StrategyWiki import (before 5:00AM UTC) and then reverted the changes to each page manually. I don't anticipate any problems like that anymore since I fixed the bug that was causing that (PCGW's API returns StrategyWiki pages with the underscores replaced with spaces... for some reason, so I just do a simple replacement of spaces to underscores in my script).
Thanks :) --Nicereddy (talk) 05:25, 23 January 2021 (UTC)[reply]
- Hadn't seen the ruby API before. Code looks good. My only suggestions would be to add a reference (e.g. stated in PCGamingWiki) and an edit summary. Otherwise pretty clear support here. Honestly at just 1500 additions I would say you don't need to seek approval (people do larger quickstatement imports without any approval) . BrokenSegue (talk) 06:00, 23 January 2021 (UTC)[reply]
- Thanks for the proposal! If others are looking for edits: Example StrategyWiki edit and Example IGDB edit.
- Couple of suggestions:
- Since you’re adding qualifiers the IGDB statements, if you happen to have it in the API query you’re doing anyway, could you add subject named as (P1810) as qualifier? I find it handy to find matching errors in the future (although for IGDB it’s not super-critical because the name can be inferred from the slug fairly easily).
- As BrokenSegue says, can you add a reference? I guess with stated in (P248) − something like:
StrategyWiki ID |
| ||||||||||
add value |
Thanks! Jean-Fred (talk) 10:00, 25 January 2021 (UTC)[reply]
- @Jean-Frédéric, BrokenSegue: I've updated my script to include the reference information now. See the edits on User:NicereddyBot from 19:15 to 19:19 on February 6th, 2021 for examples. (link to Contributions page). The Ruby gem doesn't support edit summaries as far as I can tell, so I can't add those unfortunately. - Nicereddy (talk) 19:24, 6 February 2021 (UTC)[reply]
- ok, Support BrokenSegue (talk) 01:06, 7 February 2021 (UTC)[reply]
Could this bot also import:
- Origin ID (P8261)
- Ubisoft Store game ID (P8268)
- itch.io URL (P7294)
- MobyGames game ID (former scheme) (P1933)
@Nicereddy: --Trade (talk) 20:56, 21 February 2021 (UTC)[reply]
- @Trade: It should be able to , although I doubt it'd get many (if any) MobyGames records. The other three should be doable, though they're not as straightforward as some of the other properties I've been importing thus far since they're not available via Semantic MediaWiki and not in a consistent format inside the infobox. Nicereddy (talk) 05:49, 22 February 2021 (UTC)[reply]
- Can you link to the test edits? --- Jura 09:54, 22 February 2021 (UTC)[reply]
- Comment Example edit: Special:Diff/1353686037/1355929642 Jean-Fred (talk) 17:11, 22 February 2021 (UTC)[reply]
- Support Thanks for adding references! Jean-Fred (talk) 17:11, 22 February 2021 (UTC)[reply]
- Comment Just a minor thing: Looking at Special:Diff/1353686037/1355929642, it seems to be using reference URL (P854) instead of the dedicated property (here: PCGamingWiki ID (P6337)). Both should lead to the same page providing the mapping. Can you fix it and do 50 test edits? --- Jura 12:14, 23 February 2021 (UTC)[reply]
- @Jura1: Is there a reason to use the dedicated property over the reference URL? I'm not really aware of any standard practice in that regard. Nicereddy (talk) 04:40, 24 February 2021 (UTC)[reply]
- Not really, it depends whether someone created/proposed one or not. Help:Sources#Databases describes the reference format. --- Jura 08:51, 24 February 2021 (UTC)[reply]
- Just since I don't think it's a super important change to make and it's been a month, I'd prefer not to change that at this point. Nicereddy (talk) 03:18, 25 February 2021 (UTC)[reply]
- Well, it's a bot. If it can't correctly format statements, it shouldn't be approved. That you hadn't checked Help:Sources in over a month is a bit odd. --- Jura 07:07, 25 February 2021 (UTC)[reply]
- Just noticed that it was even you who had proposed the property: Wikidata:Property_proposal/PCGamingWiki_ID. If, in the meantime, you found that there is a problem with it and that it should be deleted, please request its deletion. --- Jura 07:10, 25 February 2021 (UTC)[reply]
- @Jura1: The property is used on over 9000 items, it works fine. But I don't see a reason to use it in the references field here, or at least why that should block the bot from being approved at this point. Nicereddy (talk) 00:27, 7 March 2021 (UTC)[reply]
- @Jura1: I'd think that the part of Help:Sources you refer to speaks about databases, while PCGamingWiki (Q17013880) is an instance of website (Q35127). IMHO the format of source is acceptable. Lymantria (talk) 06:26, 8 March 2021 (UTC)[reply]
- @Jura1: The property is used on over 9000 items, it works fine. But I don't see a reason to use it in the references field here, or at least why that should block the bot from being approved at this point. Nicereddy (talk) 00:27, 7 March 2021 (UTC)[reply]
- Just since I don't think it's a super important change to make and it's been a month, I'd prefer not to change that at this point. Nicereddy (talk) 03:18, 25 February 2021 (UTC)[reply]
- Not really, it depends whether someone created/proposed one or not. Help:Sources#Databases describes the reference format. --- Jura 08:51, 24 February 2021 (UTC)[reply]
- @Jura1: Is there a reason to use the dedicated property over the reference URL? I'm not really aware of any standard practice in that regard. Nicereddy (talk) 04:40, 24 February 2021 (UTC)[reply]
┌────────────────────────────────────────────────────────────────────────────────────────────────────┘
- Interesting question. I think external-id properties are only meant for stable identifiers, such as found in databases, online or offline. So either the property is made for such and should be used, if it's not a stable identifier, we should delete the property. If the operator considers that Help:Sources doesn't apply to them, this should be discussed elsewhere. Unfortunatly, we already have too many references to cleanup (see Wikidata:Bot requests) and adding more isn't helpful. Anyways, given that the operator hasn't done the required 50 test edits in almost 2 months and we can't observe that the bot operates correctly, I can't support it at this point and I think we can close this request as stale. --- Jura 08:45, 8 March 2021 (UTC)[reply]
- Support, the important part is to add the identifier; the formatting of the reference for this identifier is a minor detail (most external-id don't have nor need a reference, adding one is already far better that the usual standard) that shouldn't block the bot to run. Cheers, VIGNERON (talk) 11:22, 8 March 2021 (UTC)[reply]
- @Nicereddy: In this stage I would like to see test edits by your bot. Lymantria (talk) 06:22, 24 March 2021 (UTC)[reply]
- @Lymantria: There were the test-edits from early February, such as Special:Diff/1353686037/1355929642 no ? Jean-Fred (talk) 11:43, 24 March 2021 (UTC)[reply]
- Right, I see. It's the discussion following to that about the way the source was formatted that caused delay. I will approve the request in a couple of days, given that no further objections will be raised. Lymantria (talk) 11:49, 24 March 2021 (UTC)[reply]
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
- Jura doesn't agree. Multichill (talk) 10:52, 25 April 2021 (UTC)[reply]
New section edit
@Lymantria: Does that mean I should make a cleanup request or will it formatt references correctly? Why haven't we seen the mandatory test edits? --- Jura 10:04, 28 March 2021 (UTC)[reply]
- We have seen the mandatory test edits, and the reference format you demand is disputed to be incorrect. Lymantria (talk) 10:41, 28 March 2021 (UTC) P.S. Every once in a while you add sections to closed requests. Except for obvious mistakes by the closing user, I don't think that is a correct way to act.[reply]
- Can you provide quotes from references to support your claims? (i.e. bot approval policy, policy on referencing) --- Jura 11:07, 28 March 2021 (UTC)[reply]
- That already has been discussed, with two users not agreeing with your interpretation of the Help:Sources preventing approval of this task. The general statement of Help:Sources says in its second paragraph "Typically the property used for sources is one of two options: vermeld in (P248) (referring to publications and media) and URL voor bron (P854) (used for websites and online databases)". That part is fulfilled. I tried to point you to some ambiguity in the guideline, which did not convince you. Fair enough, but my judgement is that this issue should not block approval. Lymantria (talk) 13:56, 28 March 2021 (UTC)[reply]
- @Lymantria: if the bot task is controversal, I don't think it should be done. Bot approval pages aren't the place to sort out questions about referencing, especially if, as your quote, it's fairly clear that the bot doesn't follow it. --- Jura 20:03, 2 April 2021 (UTC)[reply]
- Well, if it wasn't before, now it's clear that we disagree. Lymantria (talk) 20:09, 2 April 2021 (UTC)[reply]
- That already has been discussed, with two users not agreeing with your interpretation of the Help:Sources preventing approval of this task. The general statement of Help:Sources says in its second paragraph "Typically the property used for sources is one of two options: vermeld in (P248) (referring to publications and media) and URL voor bron (P854) (used for websites and online databases)". That part is fulfilled. I tried to point you to some ambiguity in the guideline, which did not convince you. Fair enough, but my judgement is that this issue should not block approval. Lymantria (talk) 13:56, 28 March 2021 (UTC)[reply]
- Can you provide quotes from references to support your claims? (i.e. bot approval policy, policy on referencing) --- Jura 11:07, 28 March 2021 (UTC)[reply]