websocket

游雁

2023-03-23 38d655e562a7090e8470cf6be705bfb3a2d2e8b9

 funasr/runtime/python/websocket/ASR_server.py

@@ -1,4 +1,10 @@
# server.py   注意本例仅处理单个clent发送的语音数据，并未对多client连接进行判断和处理
import asyncio
import websockets
import time
from queue import Queue
import threading
import argparse

from modelscope.pipelines import pipeline
from modelscope.utils.constant import Tasks
from modelscope.utils.logger import get_logger
@@ -7,12 +13,6 @@
logger = get_logger(log_level=logging.CRITICAL)
logger.setLevel(logging.CRITICAL)

import asyncio
import websockets
import time
from queue import Queue
import threading
import argparse

parser = argparse.ArgumentParser()
parser.add_argument("--host",
@@ -36,7 +36,7 @@

parser.add_argument("--punc_model",
                    type=str,
                    default="",
                    default="damo/punc_ct-transformer_zh-cn-common-vad_realtime-vocab272727",
                    help="model from modelscope")
parser.add_argument("--ngpu",
                    type=int,
@@ -49,23 +49,36 @@
voices = Queue()
speek = Queue()

# 创建一个VAD对象
vad_pipline = pipeline(
# vad
inference_pipeline_vad = pipeline(
    task=Tasks.voice_activity_detection,
    model=args.vad_model,
    model_revision="v1.2.0",
    model_revision=None,
    output_dir=None,
    batch_size=1,
    mode='online',
    ngpu=args.ngpu,
)
param_dict_vad = {'in_cache': dict()}
  
# 创建一个ASR对象
param_dict = dict()
# asr
param_dict_asr = {}
# param_dict["hotword"] = "小五 小五月"  # 设置热词，用空格隔开
inference_pipeline2 = pipeline(
inference_pipeline_asr = pipeline(
    task=Tasks.auto_speech_recognition,
    model=args.asr_model,
    param_dict=param_dict,
    param_dict=param_dict_asr,
    ngpu=args.ngpu,
)

param_dict_punc = {'cache': list()}
inference_pipeline_punc = pipeline(
    task=Tasks.punctuation,
    model=args.punc_model,
    model_revision=None,
    ngpu=args.ngpu,
)

print("model loaded")


@@ -85,17 +98,20 @@


def vad(data):  # 推理
    global vad_pipline
    global vad_pipline, param_dict_vad
    #print(type(data))
    segments_result = vad_pipline(audio_in=data)
    #print(segments_result)
    # print(param_dict_vad)
    segments_result = inference_pipeline_vad(audio_in=data, param_dict=param_dict_vad)
    # print(segments_result)
    # print(param_dict_vad)
    speech_start = False
    speech_end = False
    if len(segments_result) == 0 or len(segments_result["text"] > 1):
        return False
    elif segments_result["text"][0][0] != -1:
    
    if len(segments_result) == 0 or len(segments_result["text"]) > 1:
        return speech_start, speech_end
    if segments_result["text"][0][0] != -1:
        speech_start = True
    elif segments_result["text"][0][1] != -1:
    if segments_result["text"][0][1] != -1:
        speech_end = True
    return speech_start, speech_end

@@ -106,8 +122,9 @@
        while not speek.empty():
            audio_in = speek.get()
            speek.task_done()
            rec_result = inference_pipeline2(audio_in=audio_in)
            print(rec_result)
            rec_result = inference_pipeline_asr(audio_in=audio_in)
            rec_result_punc = inference_pipeline_punc(text_in=rec_result['text'], param_dict=param_dict_punc)
            print(rec_result_punc)
            time.sleep(0.1)
        time.sleep(0.1)    

@@ -135,20 +152,21 @@
                frames.append(data)
                RECORD_NUM += 1
            speech_start_i, speech_end_i = vad(data)
            # print(speech_start_i, speech_end_i)
            if speech_start_i:
                speech_start = speech_start_i
                # if not speech_detected:
                print("检测到人声...")
                # print("检测到人声...")
                # speech_detected = True  # 标记为检测到语音
                frames = []
                frames.extend(buffer)  # 把之前2个语音数据快加入
                # silence_count = 0  # 重置静音次数
            elif speech_end_i or RECORD_NUM > 300:
            if speech_end_i or RECORD_NUM > 300:
                # silence_count += 1  # 增加静音次数
                # speech_end = speech_end_i
                speech_start = False
                # if RECORD_NUM > 300: #这里 50 可根据需求改为合适的数据快数量
                print("说话结束或者超过设置最长时间...")
                # print("说话结束或者超过设置最长时间...")
                audio_in = b"".join(frames)
                #asrt = threading.Thread(target=asr,args=(audio_in,))
                #asrt.start()
@@ -170,16 +188,4 @@
s.start()

asyncio.get_event_loop().run_until_complete(start_server)
asyncio.get_event_loop().run_forever()


 





 

        

asyncio.get_event_loop().run_forever()