library management: "suggested library fixes"

[andrewrk]

Groove Basin should show a user interface with a list of "suggested fixes" to the user's library. The user could accept or reject these fixes. They would be things like this:

• duplicate detection and elimination
  • if a song is byte for byte the same (check md5's) then ignore the new song
  • use heuristics to guess if songs are probably the same (using tags). if
    we are reasonably confident that the songs are the same, delete the one
    with the lower quality.
  • if we're not confident enough, there will be an
    api that lists possible duplicates and actions to resolve them.
• update the duration metadata field with the exact correct duration
• moving songs to their canonical file path as specified by the user
• list files that cannot be played and offer to move or delete them. This includes DRM-protected files!
• delete empty folders in music library
• submit musicbrainz information for their tracks and feel good about it
• replace their own tags with musicbrainz information when possible
• write new song tags to disk
  • wrap audio streams in a container format in order to be able to write tags

[andrewrk]

I have a script running on my entire music collection which is testing saving song tags to disk and then ensuring if the audio data is byte-for-byte identical to before we saved the tags. When this scan completes sometime tomorrow I'll look at the results and then determine how robust saving tags is and whether we can integrate it into groove basin, or whether it needs some bug fixing first.