How to locate the folder of the track being played from USB drive?

[vbutani]

Many LMS users make a point of removing Musicbrainz tags due to issues they cause .

I tried Musicbrainz on Jellyfin earlier, it sort of messed up my collection!

 

[vbutani]

I think sqlite3 will be installed by default on Ubuntu:

sqlite3 -header -tabs files.db "select path, title, duration, bitrate, samplerate, creator, album, channels, disc, track, date, mime from details where mime not null;" > ~/tracks.txt

Although if @vbutani wants to go down this route it may be easier to perform the dedupe in sql.

This query works fine, checking tracks csv.

 

[simbun]

This query works fine, checking tracks csv.

Did you try:

select *
  from (
         select artist||' - '||regexp_replace(album, ' \(.*\)', '') as artist_album,
                regexp_replace(path, '(?i)\/(CD|DISC) *\d*', '') as path,
                sum(1) over (partition by artist||' - '||regexp_replace(album, ' \(.*\)', '')) as count_albums
           from (
                  select distinct artist,
                         album,
                         rtrim(path, replace(path, '/', '')) as path
                    from details
                    where mime not null
                )
           group by 1, 2
       )
where count_albums > 1;

EDIT: Updated to strip (xyz) from album and CD/Disc folders from paths to help with match.
This does require a regex extension to be loaded.

sqlite> .load ./regexp

Releases · nalgeon/sqlean

The ultimate set of SQLite extensions. Contribute to nalgeon/sqlean development by creating an account on GitHub.

github.com

I can put some basic instructions together if you don't know how to load the regex extension.

EDIT: I'll change it so it's a cli call like the other extract.

 

[rccarguy]

I tried Musicbrainz on Jellyfin earlier, it sort of messed up my collection!

I just deleted all musicbrainz tags, didn't have problems but I'm using LMS

 

[simbun]

The following code performs matching on lowercase albumartist and album (stripped of anything in parenthesis) and strips any trailing folders called CDXX or DiscXX from paths to avoid false positives.

# Create a temporary directory and change into it
mkdir temp && cd $_

# download the regex extension and unzip
wget "https://github.com/nalgeon/sqlean/releases/download/0.27.1/sqlean-linux-x86.zip"
unzip sqlean-linux-x86.zip

#####  copy the files.db into this directory ######

# run the following code to produce matches.txt
sqlite3 -header -tabs files.db ".load ./regexp" "select *
  from (
         select artist||' - '||regexp_replace(album, ' \(.*\)', '') as artist_album,
                regexp_replace(path, '(?i)\/(CD|DISC) *\d*', '') as path,
                sum(1) over (partition by artist||' - '||regexp_replace(album, ' \(.*\)', '')) as count_albums
           from (
                  select distinct lower(artist) as artist,
                         lower(album) as album,
                         rtrim(path, replace(path, '/', '')) as path
                    from details
                    where mime not null
                )
           group by 1, 2
       )
where count_albums > 1" > matches.txt

It's obviously a pretty crude match but it might save you some time.