Using PYO with other audio output engines

[snuq]

So unfortunately PYO's reliance on portaudio for output resulted in some serious issues on my linux install (unable to output through bluetooth, unable to record output in obs, frequent stuttering issues, occasional complete refusal to output audio until restarted computer, etc...), so I decided to try and use an alternate library for outputting PYO's audio.

After quite a bit of frustration, I was able to get a working setup using Sounddevice ( https://github.com/spatialaudio/python-sounddevice/ ). This method should work with pretty much any sound output library for python, tho some modification will be needed of course.

I am posting the relevant code here in case others have the same issue, or if anyone might have any ideas for improving it:

class AudioDriver:
    audio = None
    audio_mixer = None
    channels = 2
    output_frequency = 96000
    buffer_size = 1024
    audio_stream = None
    buffer_inputter = None
    data_table = None
    buffers = None

    def start_audio(self):
        #Start and set up PYO i 'manual' mode
        self.audio = Server(sr=self.output_frequency, audio='manual', nchnls=self.channels, buffersize=self.buffer_size, duplex=0)
        self.audio.boot()
        self.audio.start()
        self.output_frequency = self.audio.getSamplingRate()
        self.buffer_size = self.audio.getBufferSize()

        #Since we arent using the server output, this class will be used for all outputs, pyo objects must manually use audio_mixer.addInput/delInput when they play/stop
        self.audio_mixer = Mixer(outs=1, chnls=self.channels)
        self.audio_mixer.out()

        #Get the buffer stream data from the Mixer object
        self.data_table = DataTable(size=self.buffer_size, chnls=self.channels)
        self.buffer_inputter = TableFill(self.audio_mixer[0], self.data_table)
        self.buffers = []
        for channel in range(self.channels):
            self.buffers.append(numpy.asarray(self.data_table.getBuffer(channel)))
        
        #Set up sounddevice with a callback function
        self.audio_stream = sounddevice.Stream(samplerate=self.output_frequency, blocksize=self.buffer_size, channels=self.channels, dtype=numpy.float32, latency='low', callback=self.audio_process, finished_callback=self.stop_audio)
        self.audio_stream.start()

    def audio_process(self, indata, outdata, frames, time, status):
        #called by sounddevice at the start of each buffer, process PYO and copy self.audio_mixer's output into Sounddevice's buffer
        self.audio.process()
        for channel in range(self.channels):
            outdata[:, channel] = self.buffers[channel]

    def stop_audio(self):
        #Shuts down the audio engine
        self.audio.shutdown()
        self.audio_stream.stop()

A way to get the Server's audio output data directly after calling process() would remove the need for a lot of this code, and would mean that using standard .out() functions on pyo objects would behave normally, but I have not been able to find a way to do this.

is there some way to get a buffer output from a pyo Stream object? or maybe some server function i missed in the documentation? maybe this feature can be added?