shoutpy

mail feedback: Daniel Holth <dholth@fastmail.fm>

shoutpy uses Boost.Python to expose libshout 2.1 or higher to Python. It includes a separately usable C++ wrapper that makes libshout considerably nicer to program from C++.

See also oggpy. oggpy is a C++ and Boost.Python wrapper for Ogg, Vorbis AND most of Theora. With oggpy and shoutpy you could write a reencoding icecast source in Python.

See the doxygen generated documentation here. It explains some things about stock libshout2 that were unclear from the official documentation.

In Python:


#!/usr/bin/python
#
# Simple streaming to icecast.
#
# Daniel Holth

import sys
import shoutpy

s           = shoutpy.Shout()
s.user      = "source"
s.password  = "hackme"
s.mount     = "/shoutpy.ogg"
s.port      = 8001
s.format    = shoutpy.FORMAT_VORBIS

s.open()

for arg in sys.argv[1:]:
    f = open(arg, "rb")

    buf = f.read(4096)

    while buf:
        s.send(buf)
        buf = f.read(512)
        s.sync()

    f.close()

s.close()

In C++:


/*  Simple multi-file streaming to a server.
 *
 *  Daniel Holth <dholth@fastmail.fm>
 */

#include "shoutcc.h"
#include <stdio.h>

#define BUFSIZE 4096

void usage(char *name)
{
    printf("Usage:\n"
           "\t%s file1.ogg [file2.ogg]...\n", name);
}


int main(int argc, char **argv)
{
    static unsigned char buffer[BUFSIZE];
    Shout s;

    try {
        try {
            s.set_user("source");
            s.set_password("hackme");
            s.set_mount("/");
            s.set_port(8001);
            s.set_format(SHOUT_FORMAT_VORBIS);

            s.open();
        } 
        catch(ShoutErr error) {
            printf("Error connecting to server %s:%d with  username %s, password %s.\n",
                   s.get_host(), s.get_port(), s.get_user(), s.get_password());
            throw error;
        }

        if(argc > 1) {
            int delay;
            for(int i=1; i<argc; i++) {
                size_t bytes = 0;
                FILE *f = fopen(*(argv+i), "rb");

                while(!feof(f)) {
                    bytes = fread(buffer, sizeof(char), BUFSIZE, f);
                    s.send(buffer, bytes);

                    delay = s.delay();
                    if(delay > 1000) {
                        printf("delay: %d\n", s.delay());
                    }

                    s.sync();
                }
            }
            s.close();
        } 
        else {
            usage(*argv);
        }
    }
    catch(ShoutErr error) {
        printf("Shout Error: %s (code %d)\n", error.err.c_str(), error.ern);
        exit(1);
    }

    return 0;
}

Download

News

26 Nov. 2006: shoutpy goes 1.0.0! Thanks to all who helped it to mature over these years. Adds support for nonblocking I/O introduced in libshout 2.1. Deprecated building with boost.build; just use setup.py.

10 April 2004: Version 0.5.2. Added C++ example to distribution. Added a Jamfile so you can optionally build shoutpy with boost.build.

7 April 2004: Cleaned up setup.py; now it should build out of the box on gentoo (if boost.python and libshout are already installed).

27 Jan. 2004: New example.

18 Jan. 2004: Version 0.5. Now including Python module docstrings and doxygen C++ documentation.

15 Jan. 2004, evening (EST): Version 0.4 released. New feature: it can stream audio to an icecast server. ;-)