Closed gjost closed 5 years ago
Instead of making a whole new ddrindex
command and spending time figuring out how to transcode video into the IA formats, just point unpublished objects at selected dummy objects on IA:
audio: https://archive.org/download/ddr-csujad-28-1
video: https://archive.org/download/ddr-densho-1000-28-1
Fixed in commit 4cef203. Unpublished or missing objects are pointed at dummy IA records.
DDR A/V content (video, audio) is uploaded to Internet Archive. The DDR site is served by Densho but the media assets are served by IA. Example: https://archive.org/details/ddr-densho-1000-210-1
Problem: Objects whose A/V content has not yet been uploaded to IA cannot be previewed because the binaries are not there. Proposal: A mock Internet Archive API. This would be an XML file for each Entity and a set of derivative files that could be uploaded to a server that could be used for testing.
Background:
For our purposes, the "Internet Archive API" is the XML files that are created by IA when they ingest a file and its metadata. The data is provided by Densho via $LINK_TO_SCRIPT_HERE. The XML file lists the original file and its derivatives (various video formats, mp3, thumbnails) plus a SQLite3 .db file and the XML file itself.
Object URL https://archive.org/details/ddr-densho-1000-210-1 XML URL https://archive.org/download/ddr-densho-1000-210-1/ddr-densho-1000-210-1_files.xml
The DDR extracts info from the XML and includes selected parts of it when publishing to Elasticsearch:
This info is used in
ddr-public
to in segment templates segment.htmlsegment-audio.html
segment-video.html