Closed pdpinch closed 6 years ago
I think the mapping of S3 to cloudfront is part of #411 -- please add a comment and/or edit the description.
This issue was for any necessary data model and front-end changes.
Each VideoS3
object (used for video files, thumbnails, and captions) has a bucket_name
and s3_object_key
property. These can be set to the appropriate TechTV bucket and path when creating an object, for example:
VideoThumbnail.objects.get_or_create(
s3_object_key='techtv/thumbnails/<video_key>/001.jpg',
bucket_name='odl-video-services-techtv-thumbnails-ci'
video=video
)
Each cloudfront URL is based on the S3 object key, so assuming the cloudfront distribution maps unique URL paths for techtv files to S3 buckets, everything should just work.
A complication is that each VideoFile
is assumed to be either the original uploaded file or an m3u8 playlist of transcoded files, which won't be the case for TechTV videos. That will be addressed as part of #413.
Does that mean we can just close this?
Yes, I'll close it.
techtv/thumbnails/9d6aa5cf9b6e4d07a0dfdcd9aaf633f0/video_thumbnail_00001.jpg
instead ofthumbnails/9d6aa5cf9b6e4d07a0dfdcd9aaf633f0/video_thumbnail_00001.jpg
(via code)