About

Hi-Audio online platform is a collaborative web application for musicians and researchers in the MIR (Music Information Retrieval) domain, with a view to build a public database of music recordings from a wide variety of styles and different cultures. It allows:

• Creating musical compositions and collections with different levels of privacy.
• Uploading and recording audio tracks from the browser.
• Annotating audio tracks with relevant MIR information.
• Inviting collaborators to participate using different roles.

This repo contains information relative to the server side or back-end, for the client side (web application) see NOTE 1.

Local database preparation:

Install MySQL for macOS

brew install mysql

mysql.server start

mysql -u root -p

Install MySQL for Linux

sudo apt install mysql-server
sudo apt install python3-dev libmysqlclient-dev

# See NOTE 2 in case of errors

service mysql start

sudo mysql -u root -p

Install MySQL for Windows

# Install MySQL server from here: 
https://dev.mysql.com/downloads/installer/

# See NOTE 3 for more info:

mysql -u root -p

Create the local database

# Create a DB and add new user (mysqluser) at localhost. 
# Choose your own user name if you want to.
create database hiaudio ; 
CREATE USER 'mysqluser'@'localhost' IDENTIFIED BY 'password';
GRANT ALL PRIVILEGES ON hiaudio.* TO 'mysqluser'@'localhost';
FLUSH PRIVILEGES;

mysql > exit

Run the local server. Recommended Python version 3.10

Clone or download the repository

git clone https://github.com/idsinge/hiaudio_backend.git

cd hiaudio_backend

Create and activate Python environment (Linux and macOS)

python3 -m venv venv

. venv/bin/activate

Create and activate Python environment (Windows)

python -m venv venv

venv\Scripts\activate

Install requirements

pip install -r requirements.txt

Create .env file with the following content

# Google Values: https://console.cloud.google.com/apis/credentials
# SECRET_KEY is independent and can be self-elected
# JWT_SECRET_KEY: https://flask-jwt-extended.readthedocs.io/en/stable/options.html#JWT_SECRET_KEY
# OVH_EMAIL_PASSWD: https://www.ovh.com/manager/#/web/email_domain/
# ACOUSTIC_ID_API_KEY: https://acoustid.org/
GOOGLE_CLIENT_ID=*****
GOOGLE_CLIENT_SECRET=*****
SECRET_KEY=*****
JWT_SECRET_KEY=*****
OVH_EMAIL_PASSWD=*****
ACOUSTIC_ID_API_KEY=*****

Duplicate config.py.sample and rename it to config.py

# In config.py fill the following details:
MYSQL_HOST="localhost"
MYSQL_USER="mysqluser"
MYSQL_PASS="password"
MYSQL_DB="hiaudio"

# In config.py for Mail settings go to Email provider:
MAIL_SERVER = ""
MAIL_PORT = 0
MAIL_USERNAME = ""

Initialize the DB for the first time and run the app:

python initdb.py

# Run the local server 
python app.py

# Verify it's running
Open -> https://localhost:7007/

To make the frontend repo work together with the backend in local DEV mode/environment

Inside backend repo clone the frontend repo:

git clone https://github.com/idsinge/hiaudio_webapp.git

Then rename the folder hiaudio_webapp to webapp

Hint: during development it might be useful to temporarly ignore the contents of the public directory, this can be done with

# ignore public/ contents for git diff, grep, status, etc.
git ls-files -z public/ | xargs -0 git update-index --skip-worktree

# track the contents of public/ again (when commiting changes to it for example)
git ls-files -z public/ | xargs -0 git update-index --no-skip-worktree

COMPRESSION MODULE

In order to run the compression module locally, the env variable COMPRESSION_MODULE_ACTIVE at config.py needs to be set to True. It's required to execute the follwoing commands, the first for the installation of the pydub package (see NOTE 4) and the other to run the thread.

pip install pydub

python compress_thread.py

AUDIO PROCESSING MODULE

To use the Acoustic ID API for audio identification, the environment variable ACOUSTIC_ID_API_KEY needs to be set at .env. It's required to execute the follwoing commands, for the installation of pydub (see NOTE 4) and essentia-tensorflow in order to run the audio processing service.

Download the following models files and place them in the suggested locations.

1. Download and place the following models under models/ directory:

• genre_rosamerica-vggish-audioset-1.pb

2. Download, rename the following model to audio_mdl.pth and place it under process_audio_thread/pretrained_models/ directory:

• audio_mdl.pth

Installation and launch

pip install -r requirements_process_audio.txt

python process_audio_thread/process_audio_thread.py

NOTES:

1- Web App Repo

2- In case pip install mysqlclient fails, try the following commands:

sudo apt install build-essential

sudo apt-get install libmariadb-dev

sudo apt-get install pkg-config

3- MySQL Installation on Windows: https://www.w3schools.com/mysql/mysql_install_windows.asp

4- Install FFMPEG. pydub needs either sudo apt install ffmpeg (Linux) or brew install ffmpeg (Mac) in order to function correctly.