Intro

About

Name

The name was chosen like such: "spotify" => "despotifier" => "despot". It symbolizes the main author's personal despise for spotify.

What it can do now

• Library issue detection
  • Missing metadata
  • Different tags in one folder
  • Different extension tracks in one folder

Notes

It can allocate a significant amount of RAM if you have a large collection. (~0.5GiB for my ~1.6TiB collection)

• outdated RAM usage info

Database model

root                str     //path to library root directory
update_time         float   //last time the DB was updated
statistics: {               //precomputed statistical data
    max_track_peak: float   //maximum value of compensated peak for track_gain
    max_album_peak: float   //maximum value of compensated peak for album_gain
    total_length:   float   //combined playtime for all tracks
    track_counts: {
        total:      int
        clipping:   int     //tracks that have peaks exceeding 1.0
        uploaded_orig:  int
        uploaded_opus:  int
        extension: {
            flac:   int
            mp3:    int
            ****:   int
        }
        depth: {
            16:     int
            24:     int
            **:     int
        }
        rate: {
            44100:  int
            48000:  int
            *****:  int
        }
        lacking_metadata: {
            critical:   int
            wanted:     int
        },
        artwork: {
            embedded:   int
            external:   int
            both:       int
            # any: embedded+external-both
        }
    }
}
last_uploaded_track: {      //path to last uploaded tracks
    original:       str
    opus:           str
}
releases: {
    *path_N*: {             //path relative to DB root
        tracks: {
            *filename_N*: {
                mtime:      float   //seconds since epoch to find modified files and directories
                tags:   {}
                length:     float   //length (seconds)
                samples:    int     //length (samples)
                depth:      int     //bit depth
                rate:       int     //sampling rate
                link_orig:  str     //link to message in lossless channel
                link_opus:  str     //link to message in opus channel
            }
        }
        files: {
            *filename_N*: {
                mtime:      float   //seconds since epoch to find modified files and directories
                link_orig:  str     //link to message in lossless channel
                link_opus:  str     //link to message in opus channel
            }
        }
        images: {
            *filename_N*: {
                mtime:      float   //seconds since epoch to find modified files and directories
                link_orig:  str     //link to message in lossless channel
                link_opus:  str     //link to message in opus channel
            }
        }
        link_orig:  str     //link to message in lossless channel
        link_opus:  str     //link to message in opus channel
    }
}

History

12-05-2023

It was decided to replace recursive tree with a release-based storage This has several benefits:

1. get rid of recursion (more predictability, no need to transfer lots of data through recursion)
2. higher performance
3. lower memory consumption
4. smaller database size
5. simpler database updates

TODO

• add telegram-formatted link insertion in tracklist generated by format_release_string

OFFTOP

• get the target loudness that fits at least 95% of my music, but ideally >99.9%
  • finish the library scanner functionality and use python json
    • sort tracks by loudness, find average for info, test -18 and -23 targets

Note that there will be a Ukrainian readme-uk.md