o if @sdZddlZddlmZddlZddlmZddlm Z m Z dZ zddl Z Wn e yFZzedeeded edZ[wwGd d d ZGd d d e ZdS)zSilero Voice Activity Detection (VAD) implementation for Pipecat. This module provides a VAD analyzer based on the Silero VAD ONNX model, which can detect voice activity in audio streams with high accuracy. Supports 8kHz and 16kHz sample rates. N)Optional)logger) VADAnalyzer VADParamsg@z Exception: zAIn order to use Silero VAD, you need to `pip install pipecat-ai`.zMissing module(s): c@s@eZdZdZdddZdefddZdd d Zdefd d Zd S)SileroOnnxModelzONNX runtime wrapper for the Silero VAD model. Provides voice activity detection using the pre-trained Silero VAD model with ONNX runtime for efficient inference. Handles model state management and input validation for audio processing. TcCs`t}d|_d|_|rdtvrtj|dg|d|_ntj||d|_|ddg|_dS)zInitialize the Silero ONNX model. Args: path: Path to the ONNX model file. force_onnx_cpu: Whether to force CPU execution provider. CPUExecutionProvider) providers sess_options)r @>N) onnxruntimeSessionOptionsinter_op_num_threadsintra_op_num_threadsget_available_providersInferenceSessionsession reset_states sample_rates)selfpathforce_onnx_cpuoptsrL/home/ubuntu/.local/lib/python3.10/site-packages/pipecat/audio/vad/silero.py__init__*s zSileroOnnxModel.__init__srcCs|t|dkr t|d}t|dkrtd|||jvr+td|jd|t|ddkr:td||fS) z)Validate and preprocess input audio data.rrz*Too many dimensions for input audio chunk zSupported sampling rates: z (or multiple of 16000)g@?@zInput audio chunk is too short)npndim expand_dims ValueErrordimrshape)rxrrrr_validate_input?s   zSileroOnnxModel._validate_inputrcCs:tjd|dfdd|_tj|dfdd|_d|_d|_dS)zReset the internal model states. Args: batch_size: Batch size for state initialization. Defaults to 1. rfloat32dtyperN)rzeros_state_context_last_sr_last_batch_size)r batch_sizerrrrOs zSileroOnnxModel.reset_statesc CsN|||\}}|dkrdnd}t|d|kr&tdt|ddt|d}|dkr3dnd }|js=|||jrJ|j|krJ|||jrW|j|krW||t|jd sitj||fd d |_tj |j|fd d }|dvr||j tj |dd d}|j d|}|\}} | |_ nt|d| df|_||_||_|S)z*Process audio input through the VAD model.r zProvided number of samples is z< (Supported values: 256 for 8000 sample rate, 512 for 16000)r@ rr(r))axis)r r int64)inputstaterN.)r&rr$r"r/rr.r-r+ concatenater,arrayrrun) rr%r num_samplesr0 context_size ort_inputsort_outsoutr9rrr__call__Zs6   zSileroOnnxModel.__call__N)T)r) __name__ __module__ __qualname____doc__rintr&rrBrrrrr"s    rcsheZdZdZddddeedeeffddZdeffdd Zd efd d Z d e fd dZ Z S)SileroVADAnalyzeraVoice Activity Detection analyzer using the Silero VAD model. Implements VAD analysis using the pre-trained Silero ONNX model for accurate voice activity detection. Supports 8kHz and 16kHz sample rates with automatic model state management and periodic resets. N sample_rateparamsrJrKc stj||dtdd}d}zddl}t|||}Wn=ty^ddl m }z| || }|}Wdn1sBwYWnty[t|||}YnwYnwt |dd |_ d|_td dS) zInitialize the Silero VAD analyzer. Args: sample_rate: Audio sample rate (8000 or 16000 Hz). If None, will be set later. params: VAD parameters for detection thresholds and timing. rIzLoading Silero VAD model...zsilero_vad.onnxzpipecat.audio.vad.datarN) resourcesT)rzLoaded Silero VAD)superrrdebugimportlib_resourcesstrfilesjoinpath BaseException importlibrLrr_model_last_reset_time)rrJrK model_name package_path impresourcesmodel_file_pathf __class__rrrs,     zSileroVADAnalyzer.__init__cs0|dkr|dkrtd|dt|dS)zSet the sample rate for audio processing. Args: sample_rate: Audio sample rate (must be 8000 or 16000 Hz). Raises: ValueError: If sample rate is not 8000 or 16000 Hz. r r z?Silero VAD sample rate needs to be 16000 or 8000 (sample rate: )N)r"rMset_sample_rate)rrJr\rrr_s   z!SileroVADAnalyzer.set_sample_ratereturncCs|jdkrdSdS)zGet the number of audio frames required for VAD analysis. Returns: Number of frames required (512 for 16kHz, 256 for 8kHz). r r1r2)rJ)rrrrnum_frames_requiredsz%SileroVADAnalyzer.num_frames_requiredc Csz6t|tj}tj|tjdtjd}|||jd}t}||j}|t kr4|j ||_|WSt yQ}zt d|WYd}~dSd}~ww)zCalculate voice activity confidence for the given audio buffer. Args: buffer: Audio buffer to analyze. Returns: Voice confidence score between 0.0 and 1.0. r)g@rz'Error analyzing audio with Silero VAD: N)r frombufferint16astyper(rUrJtimerV_MODEL_RESET_STATES_TIMEr Exceptionrerror)rbuffer audio_int16 audio_float32new_confidence curr_time diff_timeerrrvoice_confidences   z"SileroVADAnalyzer.voice_confidence) rCrDrErFrrGrrr_rafloatrp __classcell__rrr\rrHs &%rH)rFretypingrnumpyrlogururpipecat.audio.vad.vad_analyzerrrrfr ModuleNotFoundErrorrorhrgrrHrrrrs"    `