Closes #97, #98
8.2 KiB
Supysonic
Supysonic is a Python implementation of the Subsonic server API.
Current supported features are:
- browsing (by folders or tags)
- streaming of various audio file formats
- transcoding
- user or random playlists
- cover arts (
cover.jpg
files in the same folder as music files) - starred tracks/albums and ratings
- Last.FM scrobbling
Supysonic currently targets the version 1.8.0 of the Subsonic API. For more details, go check the API implementation status.
Table of contents
Installation
Supysonic can run as a standalone application (not recommended for a "production" server) or as a WSGI application (on Apache for instance). To install it, run:
$ python setup.py install
Prerequisites
You'll need these to run Supysonic:
- Python 2.7 or >= 3.5
- Flask
- PonyORM
- Python Imaging Library
- requests
- mutagen
- watchdog (if you want to use the watcher)
You can install all of them using pip
:
$ pip install .
You may also need a database specific package:
- MySQL:
pip install pymysql
orpip install mysqlclient
- PostgreSQL:
pip install psycopg2-binary
Database initialization
Supysonic needs a database to run. It can either be a SQLite, MySQL-compatible or PostgreSQL database.
Supysonic does not automatically create the database and tables it needs to
work. Thus the database and tables must be created prior to running the
application. Please refer to the documentation of the DBMS you've chosen on how
to create a database and how to use a command-line client. If you want to use
PostgreSQL you'll have to add the citext
extension to the database once
created. This can be done when connected to the database as the superuser with
the folowing SQL command:
supysonic=# CREATE EXTENSION citext;
Table creation scripts are provided in the schema
folder for SQLite,
MySQL and PostgreSQL. Just feed them to any client you're able to use.
If you absolutely have no clue about databases, you can go with SQLite.
You'll just need the sqlite3
command-line tool. Install it and create the
database and tables with the following commands:
$ apt install sqlite3
$ sqlite3 /some/path/to/a/supysonic.db < schema/sqlite.sql
Remember the path you've used for the database file
(/some/path/to/a/supysonic.db
in the example above), you'll need it in the
configuration file.
Note that using SQLite for large libraries might not be the brightest idea as it tends to struggle with larger datasets.
Configuration
Once you have a database, you'll need to create a configuration file. It must be saved under one of the following paths:
/etc/supysonic
~/.supysonic
~/.config/supysonic/supysonic.conf
A roughly documented sample configuration file is provided as config.sample
.
The minimal configuration using the SQLite database created on the example above whould be:
[base]
database_uri = sqlite:////some/path/to/a/supysonic.db
For a more details on the configuration, please refer to documentation.
Running the application
As a standalone debug server
It is possible to run Supysonic as a standalone server, but it is only recommended to do so if you are hacking on the source. A standalone won't be able to serve more than one request at a time.
To start the server, just run the cgi-bin/server.py
script.
$ python cgi-bin/server.py
By default, it will listen on the loopback interface (127.0.0.1
) on port
5000, but you can specify another address on the command line, for instance on
all the IPv6 interfaces:
$ python cgi-bin/server.py ::
As an Apache WSGI application
Supysonic can run as a WSGI application with the cgi-bin/supysonic.wsgi
file. To run it within an Apache2 server, first you need to install the WSGI
module and enable it.
$ apt install libapache2-mod-wsgi
$ a2enmod wsgi
Next, edit the Apache configuration to load the application. Here's a basic example of what it looks like:
WSGIScriptAlias /supysonic /path/to/supysonic/cgi-bin/supysonic.wsgi
<Directory /path/to/supysonic/cgi-bin>
WSGIApplicationGroup %{GLOBAL}
WSGIPassAuthorization On
Order deny,allow
Allow from all
</Directory>
You might also need to run Apache using the system default locale, as the one
it uses might cause problems while scanning the library from the web UI. To do
so, edit the /etc/apache2/envvars
file, comment the line export LANG=C
and
uncomment the . /etc/default/locale
line. Then you can restart Apache:
$ systemctl restart apache2
With that kind of configuration, the server address will look like http://server/supysonic/
Other options
If you use another HTTP server, such as nginx or lighttpd, or prefer to use
FastCGI or CGI over WSGI, FastCGI and CGI scripts are also provided in the
cgi-bin
folder, respectively as supysonic.fcgi
and supysonic.cgi
. You
might need to edit those file to suit your system configuration.
Here are some quick docs on how to configure your server for FastCGI or CGI.
Quickstart
To start using Supysonic, you'll first have to specify where your music library is located and create a user to allow calls to the API.
Let's start by creating a new admin user this way:
$ supysonic-cli user add MyUserName -a -p MyAwesomePassword
To add a new folder to your music library, you can do something like this:
$ supysonic-cli folder add MyLibrary /home/username/Music
Once you've added a folder, you will need to scan it:
$ supysonic-cli folder scan MyLibrary
You should now be able to enjoy your music with the client of your choice!
For more details on the command-line usage, take a look at the documentation.
Watching library changes
Instead of manually running a scan every time your library changes, you can run a watcher that will listen to any library change and update the database accordingly.
The watcher is bin/supysonic-watcher
, it is a non-exiting process and doesn't
print anything to the console. If you want to keep it running in background,
either use the old nohup
or screen
methods, or start it as a simple
systemd unit (unit file not included).
It needs some additional dependencies which can be installed with the following command:
$ pip install -e .[watcher]
Upgrading
Some commits might introduce changes in the database schema. When that's the
case migration scripts will be provided in the schema/migration
folder,
prefixed by the date of commit that introduced the changes. Those scripts
shouldn't be used when initializing a new database, only when upgrading from a
previous schema.
There could be both SQL scripts or Python scripts. The Python scripts require
arguments that are explained when the script is invoked with the -h
flag.
If a migration script isn't provided for a specific database engine, it simply
means that no migration is needed for this engine.