emory-libraries / TheKeep

Curation and access for digitized and born-digital archival audio, video, and disk image content
1 stars 2 forks source link

Video objects are missing a mimetype in fedora. #7

Closed alexBLR closed 6 years ago

alexBLR commented 7 years ago

Simple risearch sparql query to identify fedora datastreams with the generic application/octet-stream mimetype:

select ?ds where { ?ds <info:fedora/fedora-system:def/view#mimeType> "application/octet-stream" . FILTER regex(str(?ds), "VIDEO") }

I ran this query to get some counts - there are 7 video records with this problem in QA.

sparql query to get video ds id and mimetype:

select ?ds where { ?ds <info:fedora/fedora-system:def/view#mimeType> "application/octet-stream" . FILTER regex(str(?ds), "VIDEO") }

rosy1280 commented 7 years ago

is there a mimetype that they need? is there a way of automatically generating these?

rosy1280 commented 7 years ago
alexBLR commented 6 years ago

Identified two pids in prod where the mime type is missing. ['info:fedora/emory:ps7h2/VIDEO', 'info:fedora/emory:ps7j6/VIDEO']

alexBLR commented 6 years ago

 >>> query = """select ?ds where {?ds <info:fedora/fedora-system:def/view#mimeType> "application/octet-stream" FILTER regex(str(?ds), "VIDEO")}"""
>>> pids = [o['ds'] for o in repo.risearch.sparql_query(query)]
>>> pids
[]
erussey commented 6 years ago

@alexBLR : I'm a little confused about this one. In the original request for this bug (in 2016), Rebecca reported that there were 185 video records in production missing a mimetype. We've done no work on this bug, so I don't understand why you only found 2.

alexBLR commented 6 years ago

@erussey, I remember something about us running it before Rebecca left. I identified only 2 videos.

erussey commented 6 years ago

ok @alexBLR . I've verified both have been fixed. Testing complete.