Crate bliss_audio

bliss audio library

bliss is a library for making “smart” audio playlists.

The core of the library is the Song object, which relates to a specific analyzed song and contains its path, title, analysis, and other metadata fields (album, genre…). Analyzing a song is as simple as running Song::from_path("/path/to/song").

The analysis field of each song is an array of f32, which makes the comparison between songs easy, by just using e.g. euclidean distance (see distance for instance).

Once several songs have been analyzed, making a playlist from one Song is as easy as computing distances between that song and the rest, and ordering the songs by distance, ascending.

If you want to implement a bliss plugin for an already existing audio player, the library::Library struct is a collection of goodies that should prove useful (it contains utilities to store analyzed songs in a self-contained database file, to make playlists directly from the database, etc). blissify for both an example of how the library::Library struct works, and a real-life demo of bliss implemented for MPD.

Examples

Analyze & compute the distance between two songs

use bliss_audio::decoder::Decoder as DecoderTrait;
use bliss_audio::decoder::ffmpeg::FFmpeg as Decoder;
use bliss_audio::playlist::euclidean_distance;
use bliss_audio::BlissResult;

fn main() -> BlissResult<()> {
    let song1 = Decoder::song_from_path("/path/to/song1")?;
    let song2 = Decoder::song_from_path("/path/to/song2")?;

    println!(
        "Distance between song1 and song2 is {}",
        euclidean_distance(&song1.analysis.as_arr1(), &song2.analysis.as_arr1())
    );
    Ok(())
}

Make a playlist from a song, discarding failed songs

use bliss_audio::decoder::Decoder as DecoderTrait;
use bliss_audio::decoder::ffmpeg::FFmpeg as Decoder;
use bliss_audio::{
    playlist::{closest_to_songs, euclidean_distance},
    BlissResult, Song,
};


fn main() -> BlissResult<()> {
    let paths = vec!["/path/to/song1", "/path/to/song2", "/path/to/song3"];
    let mut songs: Vec<Song> = Decoder::analyze_paths(&paths).filter_map(|(_, s)| s.ok()).collect();

    // Assuming there is a first song
    let first_song = songs.first().unwrap().to_owned();

    closest_to_songs(&[first_song], &mut songs, &euclidean_distance);

    println!("Playlist is:");
    for song in songs {
        println!("{}", song.path.display());
    }
    Ok(())
}

Modules

• cue
  CUE-handling module.
• decoder
  Module holding all the nitty-gritty decoding details.
• library
  Module containing utilities to properly manage a library of Songs, for people wanting to e.g. implement a bliss plugin for an existing audio player. A good resource to look at for inspiration is blissify’s source code.
• playlist
  Module containing various functions to build playlists, as well as various distance metrics.

Structs

• Analysis
  Object holding the results of the song’s analysis.
• Song
  Simple object used to represent a Song, with its path, analysis, and other metadata (artist, genre…)

Enums

• AnalysisIndex
  Indexes different fields of an Analysis.
• BlissError
  Umbrella type for bliss error types

Constants

• FEATURES_VERSION
  Stores the current version of bliss-rs’ features. It is bumped every time one or more feature is added, updated or removed, so plug-ins can rescan libraries when there is a major change.
• NUMBER_FEATURES
  The number of features used in Analysis

Type Aliases

• BlissResult
  bliss error type