WikibaseIntegrator cannot write to empty MediaInfo files

[lubianat]

EDIT: Previous title was Some mediainfo IDs are not returning results --> The MW API returned that the entity was missing., changed to reflect updates.

Hi!

I continue working with mediainfo, and now I am running into this issue:

Code:

from wikibaseintegrator import wbi_login, WikibaseIntegrator, wbi_enums
from wikibaseintegrator.wbi_config import config as wbi_config


wbi_config['MEDIAWIKI_API_URL'] = 'https://commons.wikimedia.org/w/api.php'
wbi_config['SPARQL_ENDPOINT_URL'] = 'https://query.wikidata.org/sparql'
wbi_config['WIKIBASE_URL'] = 'https://commons.wikimedia.org'
wbi_config['USER_AGENT'] = 'TiagoLubiana (https://meta.wikimedia.org/wiki/User:TiagoLubiana)'

wbi = WikibaseIntegrator()

mediainfo_id = 'M44422187'

media = wbi.mediainfo.get(entity_id=mediainfo_id)

The issue only happens for some MediaInfo ids.

E.g. M91125720 works fine, but M44422187 does not.

The website does not return the .ttl for these entries too, e.g.

https://commons.wikimedia.org/wiki/Special:EntityData/M44422187.ttl

leads to

but the entry resolves just fine:

https://commons.wikimedia.org/entity/M44422187

I realize this is not a WikibaseIntegrator issue, and I'll try and report it on Phabricator. Leaving it for information.

[lubianat]

Well, it looks like the files do not have structured data yet (as Lucas Werkmeister noted on Telegram). Maybe it is just a matter of how the API is returning this error.

[lubianat]

This is a noticeble inconsistency with QIDs — there are no empty QIDs as far as I can tell.

[lubianat]

@LeMyst Thankfully there was a simple workaround:

           except Exception as e:
                if "The MW API returned that the entity was missing." in str(e):
                    media = wbi.mediainfo.new(id=mediainfo_id)

I did not find this in the documentation though, and maybe the message could be a bit more clear on the WikibaseIntegrator end?

I will open a ticket on Phab anyways

[LeMyst]

Hello @lubianat
Thanks for opening an issue.

I don't think it's a WikibaseIntegrator related issue, we correctly have a MissingEntityException and that's what we want.
I think there was a temporary issue on commons side related to MediaInfos, because today the link https://commons.wikimedia.org/wiki/Special:EntityData/M44422187.ttl works correctly for me and the code you provided works correctly for me too.

Do you think WBI need some improvement for this case (more than the Exception)