jyapayne/vocode-python
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

add assembly ai integration

Browse source
This commit is contained in:
Kian 2023-02-27 10:19:44 -08:00
commit f39af91505
4 changed files with 14 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

4
vocode/helpers.py
View file

@ -11,7 +11,7 @@ def _get_device_prompt(device_infos: list[dict]) -> str:
Choice: """.format( Choice: """.format(
"\n".join(f"{index}: {device['name']}" for index, device in enumerate(device_infos))) "\n".join(f"{index}: {device['name']}" for index, device in enumerate(device_infos)))
def create_microphone_input_and_speaker_output(use_first_available_device=False) -> tuple[MicrophoneInput, SpeakerOutput]: def create_microphone_input_and_speaker_output(use_first_available_device=False, mic_sampling_rate=None) -> tuple[MicrophoneInput, SpeakerOutput]:
pa = pyaudio.PyAudio() pa = pyaudio.PyAudio()
num_devices = pa.get_device_count() num_devices = pa.get_device_count()
devices = list(map(pa.get_device_info_by_index, range(num_devices))) devices = list(map(pa.get_device_info_by_index, range(num_devices)))
@ -24,7 +24,7 @@ def create_microphone_input_and_speaker_output(use_first_available_device=False)
input_device_info = input_device_infos[int(input(_get_device_prompt(input_device_infos)))] input_device_info = input_device_infos[int(input(_get_device_prompt(input_device_infos)))]
output_device_info = output_device_infos[int(input(_get_device_prompt(output_device_infos)))] output_device_info = output_device_infos[int(input(_get_device_prompt(output_device_infos)))]
logger.info("Using microphone input device: %s", input_device_info['name']) logger.info("Using microphone input device: %s", input_device_info['name'])
microphone_input = MicrophoneInput(pa, input_device_info) microphone_input = MicrophoneInput(pa, input_device_info, sampling_rate=mic_sampling_rate)
logger.info("Using speaker output device: %s", output_device_info['name']) logger.info("Using speaker output device: %s", output_device_info['name'])
speaker_output = SpeakerOutput(pa, output_device_info) speaker_output = SpeakerOutput(pa, output_device_info)
return microphone_input, speaker_output return microphone_input, speaker_output

6
vocode/input_device/microphone_input.py
View file

@ -10,10 +10,10 @@ class MicrophoneInput(BaseInputDevice):
DEFAULT_SAMPLING_RATE = 44100 DEFAULT_SAMPLING_RATE = 44100
DEFAULT_CHUNK_SIZE = 2048 DEFAULT_CHUNK_SIZE = 2048
def __init__(self, pa: pyaudio.PyAudio, device_info: dict, chunk_size: int = DEFAULT_CHUNK_SIZE): def __init__(self, pa: pyaudio.PyAudio, device_info: dict, sampling_rate: int = None, chunk_size: int = DEFAULT_CHUNK_SIZE):
self.device_info = device_info self.device_info = device_info
sampling_rate = int(self.device_info.get('defaultSampleRate', self.DEFAULT_SAMPLING_RATE)) sampling_rate = sampling_rate or (self.device_info.get('defaultSampleRate', self.DEFAULT_SAMPLING_RATE))
super().__init__(sampling_rate, AudioEncoding.LINEAR16, chunk_size) super().__init__(int(sampling_rate), AudioEncoding.LINEAR16, chunk_size)
self.pa = pa self.pa = pa
self.stream = pa.open( self.stream = pa.open(
format=pyaudio.paInt16, format=pyaudio.paInt16,

8
vocode/models/transcriber.py
View file

@ -8,6 +8,7 @@ class TranscriberType(str, Enum):
BASE = "base" BASE = "base"
DEEPGRAM = "deepgram" DEEPGRAM = "deepgram"
GOOGLE = "google" GOOGLE = "google"
ASSEMBLY_AI = "assembly_ai"
class TranscriberConfig(TypedModel, type=TranscriberType.BASE): class TranscriberConfig(TypedModel, type=TranscriberType.BASE):
sampling_rate: int sampling_rate: int
@ -28,4 +29,9 @@ class DeepgramTranscriberConfig(TranscriberConfig, type=TranscriberType.DEEPGRAM
class GoogleTranscriberConfig(TranscriberConfig, type=TranscriberType.GOOGLE): class GoogleTranscriberConfig(TranscriberConfig, type=TranscriberType.GOOGLE):
model: Optional[str] = None model: Optional[str] = None
should_warmup_model: bool = False should_warmup_model: bool = False
class AssemblyAITranscriberConfig(TranscriberConfig, type=TranscriberType.ASSEMBLY_AI):
model: Optional[str] = None
should_warmup_model: bool = False
version: Optional[str] = None

4
vocode/output_device/speaker_output.py
View file

@ -7,9 +7,9 @@ class SpeakerOutput(BaseOutputDevice):
DEFAULT_SAMPLING_RATE = 44100 DEFAULT_SAMPLING_RATE = 44100
def __init__(self, pa: pyaudio.PyAudio, device_info: dict, audio_encoding: AudioEncoding = AudioEncoding.LINEAR16): def __init__(self, pa: pyaudio.PyAudio, device_info: dict, sampling_rate: int = None, audio_encoding: AudioEncoding = AudioEncoding.LINEAR16):
self.device_info = device_info self.device_info = device_info
sampling_rate = int(self.device_info.get('defaultSampleRate', self.DEFAULT_SAMPLING_RATE)) sampling_rate = sampling_rate or int(self.device_info.get('defaultSampleRate', self.DEFAULT_SAMPLING_RATE))
super().__init__(sampling_rate, audio_encoding) super().__init__(sampling_rate, audio_encoding)
self.pa = pa self.pa = pa
self.stream = self.pa.open( self.stream = self.pa.open(