Implement filearchive in Pywikibot Site
Closed, ResolvedPublicFeature
Actions

Assigned To

None

Authored By

	Multichill
	Aug 9 2019, 12:48 PM

Description

~~The very old task T23345 got implemented~~ Sorry, mixed up old versions of a not deleted file ( https://www.mediawiki.org/wiki/Manual:Oldimage_table ) and deleted images ( https://www.mediawiki.org/wiki/Manual:Filearchive_table ).

We can now search the filearchive in the api by hash, see https://commons.wikimedia.org/w/api.php?action=query&list=filearchive&faprop=sha1&fasha1=c0f895d6a4973ef711d9749fb9b83ce65ac0c53f . That means we can finally add the site function to access this in pywikibot. Function in API is described at https://www.mediawiki.org/wiki/API:Filearchive and https://commons.wikimedia.org/w/api.php?action=help&recursivesubmodules=1#query+filearchive . Implementation could be done the same way as the allimages, see https://github.com/wikimedia/pywikibot/blob/master/pywikibot/site.py#L4434

Details

	Subject	Repo	Branch	Lines +/-
	[IMPR] Add ListGenerator for filearchive to site.py	pywikibot/core	master	+108 -0

Customize query in gerrit

Related Objects
Search...

		Status	Subtype	Assigned	Task
		Resolved	Feature	None	T230196 Implement filearchive in Pywikibot Site
		Resolved		Don-vip	T60993 Make limited information from filearchive available to everyone

Event Timeline

Multichill created this task.Aug 9 2019, 12:48 PM

Restricted Application added subscribers: pywikibot-bugs-list, Aklapper. · View Herald TranscriptAug 9 2019, 12:48 PM

Multichill updated the task description. (Show Details)Aug 9 2019, 12:53 PM

Multichill mentioned this in T23345: Make old images searchable by hash.Aug 9 2019, 12:55 PM

When I hit the api I'm logged into my own (sysop) account so I get something back. Bot account can't see deleted stuff so doesn't work. Looking at https://commons.wikimedia.org/wiki/Special:ListGroupRights the rights are either see everything deleted or see nothing. A more fine grained right would be needed to be able to search by sha1, but not see the actual file content and maybe intentionally hidden metadata. Sounds complicated, dropping this one to low.

T60993 got implemented, probably now this is possible.

api.PageGenerator does not work because that module cannot be used as a generator. ListGenerator can be used instead but it retrieves the property items instead of the FilePage which would be given by a PageGenerator. What would be the best solution for this request:

retrieving the FilePages for the given request or
retrieving the list items for the given properties or
retrieving the FilePages for the given request where other parts are given as a data attribute