Looking for feedback: Is there a need for a test-library generator?

[AndreasGB]

Motivation

To troubleshoot a bug, I just had the need for a large media library of an item type I don't use, so I wrote a quick python script that creates a fake library. I now wonder whether I should develop this into a small utility that can generate libraries for all item types. Is there a need for this / would it be helpful for your development workflow?

Proposed approach

Prepare lists of plausible items of each item type

• E.g. random 1000 songs taken from MusicBrainz, Movies from TMDB, etc.
• One video and audio file as placeholder for the actual content

Then have the script use the lists to:

• Create appropriate folder structures for the item type
• Sym- or Hard-link each item's media file to the single placeholder
  • Hard-links would be necessary for using the library via bind-mount in a docker container
• Create appropriate .nfo files for the importer to pick up the linked files
• Fix up possible required additional changes via the API
  • Music requires artists to be set on the song for which no .nfo files can be written

[nielsvanvelzen]

(moved to jellyfin-meta)

I think having better test data would be nice, but it should probably contain real audio/video/photo/book files in various codecs for even better testing. Additionally having it on a team-hosted server (access for Jellyfin team only) might be easier because you don't need to setup a server locally.