Wikidata:Bot requests/Archive/2018/3

From Wikidata
Jump to navigation Jump to search

This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

items for segments

For many anthology film (Q336144), it can be worth creating an item for each segment (sample: Q16672466#P527). Such items can include details on director/cast/etc as applicable (sample: Q26156116).

The list of anthology films includes already existing items.

This task is similar to #duos_without_parts above.
--- Jura 10:14, 22 April 2017 (UTC)

What source could the bot use? Matěj Suchánek (talk) 07:25, 25 August 2017 (UTC)
Good question. I checked the first 12 films on list above and about half had a WP article detailing the episodes.
For automated import, the structure of these articles might not be sufficiently standard:
  • section header per episode (ru), (pl)
  • table with names of episodes (de),
  • section with more details for each episode (es), (ru), (ca)
Maybe a user would need to gather the list for each film and a tool would create segment items. I guess I could try that on a spreadsheet.
--- Jura 08:49, 25 August 2017 (UTC)
  • I think it's still worth doing.
    --- Jura 10:43, 1 March 2018 (UTC)

Import lighthouses from enwiki

Per discussion at Wikidata_talk:WikiProject_Lighthouses#enwiki_bot_import.3F, please import the remaining lighthouses at

http://petscan.wmflabs.org/?psid=1187483 (from w:Category:Pages using infobox Lighthouse needing Wikidata item)

by creating new items and adding the new qid to the enwiki template. There is a mapping of properties at Wikidata:WikiProject_Lighthouses/tools#Mapping_of_infobox_properties_for_lighthouses. Some fields may not be suitable for bot import and could be skipped.

All these lighthouses are in more general articles about the region/island/place. Articles may include another infobox about that.
--- Jura 06:17, 30 July 2017 (UTC)

  • I think it's still worth doing.
    --- Jura 10:43, 1 March 2018 (UTC)

Item documentation

For every item which has something on its talk page a bot could usefully prepend {{Item documentation}} as in this edit.

It would be good if one of the active maintenance bots could then do this as new talk pages are created. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:54, 4 March 2018 (UTC)

Remove "Wikimedia list article"@en from items that don't have P31=Q13406463

SELECT *
WHERE
{
    ?item schema:description "Wikimedia list article"@en 
    MINUS { ?item wdt:P31 wd:Q13406463 }
}

Try it!

Currently 12514 items. Sample items: Q99676, Q115702, Q117073, Q117208, Q127301.

The above either got a description before P31 was changed or prematurely. It would be good if maintenance was done on these. The same could probably be done for other languages considered important.

This is similar to Wikidata:Bot_requests/Archive/2018/02#Nations at games.
--- Jura 14:41, 7 March 2018 (UTC)

I have a couple of suggestions for that query:
1) Some items are instances of a subclass of Q13406463, but not Q13406463 itself (e.g. list of people with the family name Carey (Q23044947) has P31:Wikimedia list of persons (Q19692233)). It might not be wise to remove the description from these.
2) If the English label starts with the word 'List', even if it doesn't have P31/P279*=Q13406463, the item is most likely a list (e.g. List of Trinidadian football transfers 2013–14 (Q16258717)). I browsed through some of these, and it seems that they are likely have problems, but those are best solved manually.
If we take the two above mentioned points into account, we have the following query:
SELECT ?item ?itemLabel
WHERE
{
    ?item schema:description "Wikimedia list article"@en .
    MINUS { ?item wdt:P31/wdt:P279* wd:Q13406463 }
    FILTER NOT EXISTS {
      ?item rdfs:label ?enLabel.
      FILTER(LANG(?enLabel) = 'en') 
      FILTER(STRSTARTS(lcase(STR(?enLabel)), 'list'))
    }
    SERVICE wikibase:label { bd:serviceParam wikibase:language "en" }
}
Try it!
The query currently returns 4254 items. --Shinnin (talk) 19:12, 25 March 2018 (UTC)
Good point .. I hadn't recalled lists having subclasses. So, yes, the 4000 "lists of persons" don't need the description removed. At some point, it seems that subclasses get carried away and include things like "Belgium in the Eurovision Song Contest" (60 instances), "aspect of history", etc. P279 at events in a specific year or time period (Q18340514) seems to be the source of that. Looking one of the samples from your query [1] reminded me where some of the incorrect P31 came from. Seems that cleanup hasn't gone all the way through. I will try to sort out the subclasses. I think the query should work (for English). Sorting out the subclasses will add more.
--- Jura 20:06, 25 March 2018 (UTC)