ucekmez / dejavu

Audio fingerprinting and recognition in Python

dejavu fork

Purpose

Installation

1. Download Dockerfile and docker-compose.yml files into the same folder

2. run

docker-compose up --build -d

Usage

Open http://localhost:7778 and start experimenting the library using jupyterlab

jupyter token can be seen by typing

docker logs afs

create a folder (e.g mp3) and put your training files into it

import warnings
warnings.filterwarnings(action='once')

from dejavu import Dejavu
from dejavu.recognize import FileRecognizer

djv = Dejavu(dburl='postgresql+psycopg2://ugur:123123@db/db')

djv.fingerprint_directory('/workspace/data/mp3', extensions=['mp3'], nprocesses=8)

then test it by typing

song = djv.recognize(FileRecognizer, 'test.mp3')

Extra :

how to convert any music file to mp3

from pydub import AudioSegment

audio = AudioSegment.from_file("somefile.wav", format='wav')
audio.export("result.mp3", format="mp3")