Balance audio left/right based on window location

audio-window-shift.py

import pulsectl
import gi
gi.require_version('Wnck', '3.0')
from gi.repository import Wnck, GLib
screen = Wnck.Screen.get_default()
screen.force_update()  
width = screen.get_width()
pulse = pulsectl.Pulse('audio-window-shift')
sink = pulse.sink_list()[0]

width = screen.get_width()


def balance(uniform):
    range = [0.3,1.3]
    diff = range[1]-range[0]
    left = diff * uniform
    right = diff * (1-uniform)
    return [
        range[0] + right,
        range[0] + left,
    ]

def set_balance(window):
    for proc in pulse.sink_input_list():
        try:
            pid = int(proc.proplist['application.process.id'])
            if window.get_pid() != pid :
                continue
            volume = proc.volume;
            geo = window.get_geometry()
            center_norm = (geo.xp+(geo.widthp/2))/width
            [left,right] = balance(center_norm)
            volume.values[0] = left
            volume.values[1] = right
            print(geo,center_norm,pid,str(volume),str(proc))
            pulse.sink_input_volume_set(proc.index,volume)
        except Exception as e:
            print(e)


def setup_window(window):
    set_balance(window)
    window.connect("geometry-changed", lambda window,_:set_balance(window), None)

screen.connect("window-opened",lambda _,window: setup_window(window))
for window in screen.get_windows():
    setup_window(window)
# Connect the callback function to the "window-moved" signal

# Run the main loop to listen for signals
GLib.MainLoop().run()