Workflow suggestions for reassembling library from backups

I’m looking for a suitable workflow for reassembling and managing my music library after a hard drive failure (this time on a NAS). I assume I’m using beets wrong, because I’m getting some odd results. I’d be grateful for some feedback about what’s going on and how to improve my process.

Thanks!

Challenges:

  • Due to the piecemeal nature of existing backups, there is a significant amount of duplication in different formats, bitrates, etc.
  • Some music files have generic file names and no metadata. (e.g. “Track 01”)
  • Besides the music files, many of the album-level directories contain cover art and liner notes.
  • Some tracks have important info in the title, which must be preserved. (e.g. “radio edit”)
  • Many tracks are singletons (not complete albums)

Desired outcomes:

  • Music files organized on disk
  • Keep only the highest quality copy of duplicates
  • Preserve info in titles (“radio edit,” etc.)

I’ve done the following so far:

  • Processed tracks without metadata through MusicBrainz Picard to get guesses of titles. Then matched these against an old backup of my iTunes library XML file to make corrections where possible.
  • Placed all files in an “incoming” directory, and ran a beet import, set to move files appropriately.
    • Tried to use the auto-tagger on import, but it tended to remove data such as “radio edit.”
  • Wrote a script to check hashes of each file and remove exact duplicates.
  • Used the Duplicates plugin to remove duplicates based on metadata, keeping the higher bitrate versions.

I still need to deal with the following issues:

  • Still have duplicate artists. (e.g. “Death Cab for Cutie”, “Death Cab For Cutie”)

  • Still have duplicates albums with appended numbers. (e.g. “Plans [745]”, “Plans [746]”)

  • When I import new music (with move files option), images and liner notes are left behind.

  • When importing, beets moved many files into a few erroneous album directories. (e.g. my directory Library->Asa->Asa contains 97 files from dozens of different artists.)

  • Many regular track files have been erroneously moved into the “Compilations” directory. (e.g. “Library/Compilations/A Rush Of Blood To The Head [548]/00 Warning Sign.m4a” should be “Library/Coldplay/A Rush Of Blood To The Head/08 Warning Sign.m4a”)

  • I’ve manually moved some files and ran beet update. The update wants to make a number of identical changes with each run. (See Update repeatedly outputs the same changes - #5 by zwiebelspaetzle). beet write stops this, but presents it’s own issues (see below).

  • If I run beet write -p, beets wants to move more files into wrong locations. For example:

    Fleetwood Mac - Greatest Hits - As Long as You Follow
    albumartist: Fleetwood Mac → Al Green
    genre: Pop → R&B/Soul
    year: 1988 → 1973
    month: 11 → 04
    day: 22 → 01

Am I going about this all wrong?

Thanks!

That’s a big job no matter the tool you use.

As for [Radio Edit] portion, your defined file naming structure in the config file is going to overwrite that no matter what unless you import the file and match it to a release that is also labelled as “radio edit”. If “radio edit” is not an official release then perhaps import it as-is then modify the attributes manually. Alternatively, match the song to the official release (if want the metadata from the release) then after the fact modify the title and include “radio edit” manually.

To the duplicate artist issue, that again is probably related to variations in the band name from different sources. If one album is matched to a release on say Discogs and the second album is matched to a release on MusicBrainz each may spell the name of the band different. Manual intervention is the key or a forced template filter on import to ensure Title Case on all artist names but that is probably overkill.

There is a plugin for moving related files see the plugin from the docs here GitHub - sbarakat/beets-copyartifacts: A plugin for beets that moves non-music files during the import process. but notice the readme suggests you select from one of two successors since this project is deprecated.

I have issue with the numbers at the end of files and folders also, haven’t dug in to why it happens but I noticed if I have the file open while trying to move it usually causes it. Gotta figure how to clean it up myself.

That’s all I have to time to write now, sorry hope it helps with some of those issues.

@beets_vanuatu , thanks for the tips!