[说话人认证]SPKFunctor完成本地说话人embs/wav加载。

data/denoise.py

@@ -0,0 +1,13 @@
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+
+
+ans = pipeline(
+    Tasks.acoustic_noise_suppression,
+    model='iic/speech_frcrn_ans_cirm_16k')
+
+wav_file = 'speaker_wav/HaiaoDuan.wav'    
+output_path = 'denoise_output/HaiaoDuan_denoise_output.wav'
+result = ans(
+    wav_file,
+    output_path=output_path)

data/record.py

@@ -0,0 +1,35 @@
+"""
+本地录音，保存为wav格式，存储在data/speaker_wav目录下
+"""
+import pyaudio
+import wave
+
+def record_audio(filename, duration=5, format=pyaudio.paInt16, channels=1, rate=16000):
+    """
+    本地录音，保存为wav格式，存储在data/speaker_wav目录下
+    """
+    p = pyaudio.PyAudio()
+    stream = p.open(format=format, channels=channels, rate=rate, input=True, frames_per_buffer=1024)
+    
+    print("按下回车键开始录音...")
+    input()
+    frames = []
+    for i in range(0, int(rate / 1024 * duration)):
+        data = stream.read(1024)
+        frames.append(data)
+    print("录音结束")
+    stream.stop_stream()
+    stream.close()
+    p.terminate()
+    wav_file = wave.open(filename, 'wb')
+    wav_file.setnchannels(channels)
+    wav_file.setsampwidth(p.get_sample_size(format))
+    wav_file.setframerate(rate)
+    wav_file.writeframes(b''.join(frames))
+    wav_file.close()
+
+if __name__ == "__main__":
+    record_audio(
+        "data/speaker_wav/test.wav",
+        duration=5
+    )

data/speakers.json

@@ -0,0 +1,400 @@
+[
+    {
+        "speaker_id": "b7e2c8e2-1f3a-4c2a-9e7a-2c1d4e8f9a3b",
+        "speaker_name": "ZiyangZhang",
+        "wav_path": "/home/lyg/Code/funasr/data/speaker_wav/ZiyangZhang.wav",
+        "speaker_embs": [
+            -0.4249887466430664,
+            -0.12976674735546112,
+            1.6118208169937134,
+            1.3348901271820068,
+            0.1423041820526123,
+            0.16940945386886597,
+            -0.042910803109407425,
+            0.9634712934494019,
+            0.9677271246910095,
+            1.1112406253814697,
+            -2.0086846351623535,
+            1.729629635810852,
+            -0.3664000928401947,
+            2.4323978424072266,
+            -1.587996244430542,
+            -1.0803641080856323,
+            0.08011860400438309,
+            1.6515964269638062,
+            -1.1337167024612427,
+            -0.5088973045349121,
+            -1.0002555847167969,
+            0.11426643282175064,
+            -0.8616334199905396,
+            -0.006051262840628624,
+            0.44800689816474915,
+            0.6659525632858276,
+            -0.9864538908004761,
+            2.1259539127349854,
+            -0.49345871806144714,
+            -0.14384664595127106,
+            0.0742349922657013,
+            0.25577273964881897,
+            1.0516602993011475,
+            1.7297064065933228,
+            -0.44126248359680176,
+            1.3971654176712036,
+            0.04305446520447731,
+            -2.261837959289551,
+            -0.355578750371933,
+            -0.8388981819152832,
+            0.8178591728210449,
+            0.016942109912633896,
+            0.8212596774101257,
+            1.108891248703003,
+            -0.5182072520256042,
+            -0.07741295546293259,
+            0.9407528042793274,
+            0.026407398283481598,
+            -0.6210324168205261,
+            -2.0659642219543457,
+            0.13895569741725922,
+            -1.3570973873138428,
+            2.236407995223999,
+            -0.29706746339797974,
+            1.9819035530090332,
+            1.3580390214920044,
+            -0.5505754351615906,
+            0.7189999222755432,
+            -0.3190038502216339,
+            1.1075336933135986,
+            -1.4158482551574707,
+            0.20138776302337646,
+            0.8354343175888062,
+            0.1671304553747177,
+            -0.56927490234375,
+            1.057538390159607,
+            -0.2868591248989105,
+            0.005044424440711737,
+            0.49878695607185364,
+            -0.7493277192115784,
+            2.4639663696289062,
+            0.5516767501831055,
+            -0.2763596177101135,
+            -0.8769170641899109,
+            -1.296872615814209,
+            -0.5233777165412903,
+            -0.10551001876592636,
+            -0.5955559611320496,
+            -0.6046199202537537,
+            0.22645621001720428,
+            1.12480890750885,
+            -0.3678736388683319,
+            -1.1580262184143066,
+            -0.3625229299068451,
+            0.8251489996910095,
+            0.3464623987674713,
+            2.261840581893921,
+            -0.11341957747936249,
+            -0.6645990610122681,
+            0.8480257987976074,
+            -0.47770705819129944,
+            0.8085628747940063,
+            -0.26823946833610535,
+            -0.25040531158447266,
+            1.0610276460647583,
+            -0.14239133894443512,
+            -1.309299349784851,
+            -1.0987954139709473,
+            -0.1301683634519577,
+            -0.05199439451098442,
+            -0.07838833332061768,
+            -0.21310138702392578,
+            0.29347339272499084,
+            1.0793802738189697,
+            -1.813226342201233,
+            -1.1362330913543701,
+            -0.13013578951358795,
+            0.6647212505340576,
+            -0.34312230348587036,
+            0.5921282172203064,
+            0.26284533739089966,
+            0.9369505047798157,
+            0.1739131063222885,
+            0.7924790978431702,
+            0.3412249982357025,
+            0.16646981239318848,
+            -0.32468467950820923,
+            -0.5835385918617249,
+            0.05923287197947502,
+            1.191710352897644,
+            -0.3653518557548523,
+            -0.8665252923965454,
+            0.7419591546058655,
+            -1.7234965562820435,
+            0.3421083092689514,
+            -0.24517370760440826,
+            -0.8724228143692017,
+            -0.11004912108182907,
+            -0.10676378011703491,
+            -1.0688399076461792,
+            0.4397974908351898,
+            -0.9902229309082031,
+            -0.2676651179790497,
+            1.4346729516983032,
+            0.34571582078933716,
+            0.9091840386390686,
+            0.41458258032798767,
+            -0.7863419055938721,
+            0.6952191591262817,
+            0.8847752809524536,
+            0.15871241688728333,
+            -0.10740098357200623,
+            -0.5305340886116028,
+            1.0536329746246338,
+            -1.337695837020874,
+            0.23358777165412903,
+            -0.19285082817077637,
+            -0.5339606404304504,
+            -0.6768214106559753,
+            1.6815600395202637,
+            -0.36710524559020996,
+            -0.22888287901878357,
+            -0.2714850902557373,
+            -0.0895417258143425,
+            0.3480932116508484,
+            -0.19148986041545868,
+            0.44108960032463074,
+            0.03198949620127678,
+            -0.3665091097354889,
+            -0.6040502786636353,
+            0.37234461307525635,
+            -0.07462035119533539,
+            -0.18109525740146637,
+            -0.19882601499557495,
+            0.33298638463020325,
+            0.039957765489816666,
+            0.6185765266418457,
+            1.5921381711959839,
+            0.04164457693696022,
+            -0.7556226849555969,
+            -1.0537445545196533,
+            0.36932048201560974,
+            -0.2881639897823334,
+            -1.3762420415878296,
+            -0.6029151678085327,
+            -1.3592504262924194,
+            0.6726564168930054,
+            0.06349147856235504,
+            -0.4627697765827179,
+            1.1113581657409668,
+            -1.1767970323562622,
+            0.3900119662284851,
+            -0.3050364851951599,
+            -0.2807784676551819,
+            -0.7237444519996643,
+            -0.039161279797554016,
+            0.5845404267311096,
+            -0.4385261833667755,
+            -0.3988557755947113,
+            -1.235430359840393,
+            -0.648483395576477,
+            1.084520936012268
+        ]
+    },
+    {
+        "speaker_id": "b7e2c8e2-1f3a-4c2a-9e7a-2c1d4e8f9a3b",
+        "speaker_name": "HaiaoDuan",
+        "wav_path": "/home/lyg/Code/funasr/data/speaker_wav/HaiaoDuan.wav",
+        "speaker_embs": [
+            -1.3490606546401978,
+            -0.9654964208602905,
+            0.6671794652938843,
+            2.3401081562042236,
+            -1.374346137046814,
+            0.24404077231884003,
+            0.08137784898281097,
+            0.10915698111057281,
+            0.8208633065223694,
+            -1.0312862396240234,
+            1.721955418586731,
+            -0.16976028680801392,
+            -1.0259445905685425,
+            -0.9134035706520081,
+            -1.3709611892700195,
+            -0.6821202635765076,
+            1.0825326442718506,
+            1.4931895732879639,
+            -0.06801076978445053,
+            -0.5044959187507629,
+            -1.3154232501983643,
+            -1.1049765348434448,
+            0.6122218370437622,
+            1.1061663627624512,
+            -0.2288999855518341,
+            -0.03568289428949356,
+            -0.9260172247886658,
+            1.1030527353286743,
+            -0.7439772486686707,
+            1.4323620796203613,
+            0.2221372127532959,
+            -0.8355774283409119,
+            0.6758987307548523,
+            0.8520456552505493,
+            -0.0186605341732502,
+            -0.981821596622467,
+            0.11743613332509995,
+            -0.3539535701274872,
+            -0.33924832940101624,
+            -0.510174036026001,
+            0.6893219351768494,
+            -0.10966216027736664,
+            -1.5873743295669556,
+            1.7041956186294556,
+            -0.9844599366188049,
+            -1.368901252746582,
+            0.44316115975379944,
+            -2.406590700149536,
+            0.9880101680755615,
+            0.8344699740409851,
+            0.22896111011505127,
+            -1.4464795589447021,
+            2.222980260848999,
+            -0.22508130967617035,
+            0.8659772276878357,
+            0.7801474928855896,
+            1.824644923210144,
+            -0.2455991804599762,
+            -0.06682202965021133,
+            0.07106778025627136,
+            -1.8072712421417236,
+            0.7733234763145447,
+            0.20490191876888275,
+            -1.119908094406128,
+            -1.2623472213745117,
+            0.34426289796829224,
+            0.7909225821495056,
+            0.47128093242645264,
+            -0.9976771473884583,
+            -0.6703121662139893,
+            0.7459381818771362,
+            1.0664807558059692,
+            0.659284770488739,
+            -0.49438077211380005,
+            0.1974140703678131,
+            -0.07557231187820435,
+            -1.324866533279419,
+            -1.2217090129852295,
+            -1.0160834789276123,
+            0.7517350912094116,
+            0.06301767379045486,
+            0.8621189594268799,
+            -1.033493161201477,
+            -0.18051855266094208,
+            -0.2633781135082245,
+            0.5859690308570862,
+            1.5803791284561157,
+            -0.7071301341056824,
+            -0.016185184940695763,
+            -0.5259001851081848,
+            -0.6252623796463013,
+            1.4383807182312012,
+            0.6068354845046997,
+            0.39534664154052734,
+            0.22612401843070984,
+            -1.541978120803833,
+            -2.575181484222412,
+            -0.9924071431159973,
+            1.9649298191070557,
+            -1.1940282583236694,
+            -0.6481325030326843,
+            -1.5226261615753174,
+            1.6535273790359497,
+            0.7740333676338196,
+            -1.8780876398086548,
+            0.627184271812439,
+            1.0915889739990234,
+            1.694388508796692,
+            -0.47886598110198975,
+            -0.04895557090640068,
+            0.3620351552963257,
+            0.640113115310669,
+            -0.4149058163166046,
+            -0.18083086609840393,
+            -0.30447620153427124,
+            0.022528085857629776,
+            -0.6550383567810059,
+            -0.3812088668346405,
+            -0.478842169046402,
+            0.6615785360336304,
+            0.49959492683410645,
+            -0.249789759516716,
+            1.7448066473007202,
+            -0.9037050008773804,
+            -0.7441433668136597,
+            0.5949154496192932,
+            -1.1230697631835938,
+            -0.2552490830421448,
+            0.4216223657131195,
+            -0.5870983004570007,
+            0.7283152937889099,
+            -0.13834434747695923,
+            -1.3267407417297363,
+            1.1050132513046265,
+            1.731435775756836,
+            0.3724023103713989,
+            0.830539882183075,
+            -1.032881736755371,
+            0.8204181790351868,
+            0.05735205113887787,
+            0.5442802906036377,
+            -0.7974395751953125,
+            0.18374553322792053,
+            -0.17642715573310852,
+            -0.051413919776678085,
+            -0.2413552850484848,
+            -0.43316808342933655,
+            -0.2594863772392273,
+            1.5363879203796387,
+            0.5056991577148438,
+            -1.3894445896148682,
+            -1.2057586908340454,
+            -0.48546579480171204,
+            -0.2659154236316681,
+            0.9767322540283203,
+            -1.97313392162323,
+            -0.3016327917575836,
+            -0.6123557686805725,
+            0.288481205701828,
+            0.2976057827472687,
+            0.08243764936923981,
+            0.6122551560401917,
+            -0.6019028425216675,
+            -0.10548368841409683,
+            -0.016991911455988884,
+            1.75961172580719,
+            0.6418831944465637,
+            0.3137458264827728,
+            0.25365981459617615,
+            -0.45389246940612793,
+            0.238858163356781,
+            0.2631453275680542,
+            1.1121031045913696,
+            -0.9991472363471985,
+            -0.8904637694358826,
+            -1.1346020698547363,
+            -1.1918814182281494,
+            -1.1205440759658813,
+            -1.486283779144287,
+            1.0530670881271362,
+            -0.583172082901001,
+            0.26391518115997314,
+            1.2654175758361816,
+            -0.8430055975914001,
+            0.21697403490543365,
+            -0.30710718035697937,
+            2.191946506500244,
+            -0.19980488717556,
+            -0.5966204404830933,
+            0.04923265427350998,
+            -0.8815436959266663,
+            0.9289136528968811
+        ]
+    }
+]

data/speakers.json.backup

@@ -0,0 +1,14 @@
+[
+    {
+        "speaker_id": "b7e2c8e2-1f3a-4c2a-9e7a-2c1d4e8f9a3b",
+        "speaker_name": "ZiyangZhang",
+        "wav_path": "/home/lyg/Code/funasr/data/speaker_wav/ZiyangZhang.wav",
+        "speaker_embs": ""
+    },
+    {
+        "speaker_id": "b7e2c8e2-1f3a-4c2a-9e7a-2c1d4e8f9a3b",
+        "speaker_name": "HaiaoDuan",
+        "wav_path": "/home/lyg/Code/funasr/data/speaker_wav/HaiaoDuan.wav",
+        "speaker_embs": ""
+    }
+]

main.py

@@ -1,30 +1,12 @@
-from funasr import AutoModel
+from src.server import app
+import uvicorn
+from datetime import datetime
+from src.utils.logger import get_module_logger, setup_root_logger
 
-chunk_size = 200  # ms
+time = format(datetime.now(), "%Y-%m-%d %H:%M:%S")
-model = AutoModel(model="fsmn-vad", model_revision="v2.0.4", disable_update=True)
+setup_root_logger(level="DEBUG", log_file=f"logs/fastapiserver_{time}.log")
+logger = get_module_logger(__name__)
 
-import soundfile
 
-wav_file = "tests/vad_example.wav"
+if __name__ == "__main__":
-speech, sample_rate = soundfile.read(wav_file)
+    uvicorn.run(app, host="0.0.0.0", port=8000)
-chunk_stride = int(chunk_size * sample_rate / 1000)
-
-cache = {}
-total_chunk_num = int(len((speech) - 1) / chunk_stride + 1)
-for i in range(total_chunk_num):
-    speech_chunk = speech[i * chunk_stride : (i + 1) * chunk_stride]
-    is_final = i == total_chunk_num - 1
-    res = model.generate(
-        input=speech_chunk,
-        cache=cache,
-        is_final=is_final,
-        chunk_size=chunk_size,
-        disable_pbar=True,
-    )
-    if len(res[0]["value"]):
-        print(res)
-
-print(f"len(speech): {len(speech)}")
-print(f"len(speech_chunk): {len(speech_chunk)}")
-print(f"total_chunk_num: {total_chunk_num}")
-print(f"generateconfig: chunk_size: {chunk_size}, chunk_stride: {chunk_stride}")

src/functor/spk_functor.py

@@ -11,6 +11,7 @@ import json
 import torch
 import threading
 import numpy
+import os
 
 # 日志
 from src.utils.logger import get_module_logger
@@ -60,6 +61,7 @@ class SPKFunctor(BaseFunctor):
 
         def add_speaker(self, speaker: SpeakerCreate) -> None:
             self._spk_data.append(speaker)
+            logger.debug("添加说话人: %s", speaker)
 
         def verify(self, emb: numpy.ndarray) -> Dict:
             # 将输入的numpy embedding转换为tensor
@@ -116,6 +118,9 @@ class SPKFunctor(BaseFunctor):
         self._input_queue: Queue = None  # 输入队列
         self._audio_config: AudioBinary_Config = None  # 音频配置
 
+        logger.debug("加载本地说话人数据")
+        self.load_spk_data_local()
+
     def load_spk_data_local(
         self,
         spk_data_path: str = 'data/speakers.json',
@@ -125,10 +130,38 @@ class SPKFunctor(BaseFunctor):
         """
         with open(spk_data_path, 'r') as f:
             spk_data = json.load(f)
-            for spk in spk_data:
+            for i, spk in enumerate(spk_data):
+                logger.debug("加载本地说话人数据: %s", spk)
+                if spk['speaker_embs'] == "" and spk['wav_path'] != "":
+                    logger.debug("尝试转换本地wav为embs: %s", spk['wav_path'])
+                    try:
+                        # 读取数据为numpy数组
+                        import soundfile as sf
+                        import numpy as np
+                        wav_data, sr = sf.read(spk['wav_path'], dtype='int16')
+                        # 确保是单通道
+                        if wav_data.ndim > 1:
+                            wav_data = wav_data[:, 0]
+                        # 转为numpy数组后送入pipeline
+                        spk['speaker_embs'] = self._sv_pipeline([wav_data], output_emb=True)['embs'][0]
+                        logger.debug("转换本地wav为embs: length=%s type=%s", len(spk['speaker_embs']), type(spk['speaker_embs']))
+                    except Exception as e:
+                        logger.error("转换本地wav为embs失败: %s", e)
+                else:
+                    logger.debug("加载本地说话人数据: %s", spk)
+                # 将spk的speaker_embs转换为numpy
+                spk['speaker_embs'] = numpy.array(spk['speaker_embs'])
                 self._spk_verify.add_speaker(SpeakerCreate(**spk))
-        
+                spk['speaker_embs'] = spk['speaker_embs'].tolist()
+                spk_data[i] = spk
+            # 保存更新后的数据
 
+            try:
+                with open(spk_data_path, 'w') as f:
+                    json.dump(spk_data, f, indent=4)
+            except Exception as e:
+                logger.error("保存更新后的数据失败: %s", e)
+        
     def reset_cache(self) -> None:
         """
         重置缓存, 用于任务完成后清理缓存数据, 准备下次任务

src/functor/vad_functor.py

@@ -159,6 +159,7 @@ class VADFunctor(BaseFunctor):
                 self._audio_cache = numpy.concatenate((self._audio_cache, data))
             elif isinstance(self._audio_cache, list):
                 self._audio_cache.append(data)
+
         if self._audiobinary_cache is None:
             self._audiobinary_cache = data
         else:
@@ -175,17 +176,28 @@ class VADFunctor(BaseFunctor):
         处理数据
         使用model进行生成, 并使用_do_callback进行回调
         """
+        if data is None:
+            result = self._model["vad"].generate(
+                input=self._audio_cache,
+                cache=self._model_cache,
+                chunk_size=self._audio_config.chunk_size,
+                is_final=True,
+            )
+            self._do_callback(result[0]["value"])
+            return
+
         self._predeal_data(data)
         if len(self._audio_cache) >= self._audio_config.chunk_stride:
             result = self._model["vad"].generate(
                 input=self._audio_cache,
                 cache=self._model_cache,
                 chunk_size=self._audio_config.chunk_size,
+                max_end_silence_time = 300,
                 is_final=False,
             )
             if len(result[0]["value"]) > 0:
                 self._do_callback(result[0]["value"])
-                # logger.debug(f"VADFunctor结果: {result[0]['value']}")
+                logger.debug(f"VADFunctor结果: {result[0]['value']}")
             self._audio_cache = None
 
     def _run(self):
@@ -202,11 +214,11 @@ class VADFunctor(BaseFunctor):
         while self._is_running:
             try:
                 data = self._input_queue.get(True, timeout=1)
-                if data is None:
+                # logger.debug("[VADFunctor]获取到的数据length: %s", len(data))
-                    break
-                logger.debug("[VADFunctor]获取到的数据length: %s", len(data))
                 self._process(data)
                 self._input_queue.task_done()
+                if data is None:
+                    break
             # 当队列为空时, 间隔1s检测是否进入停止事件。
             except Empty:
                 if self._stop_event:
@@ -252,67 +264,4 @@ class VADFunctor(BaseFunctor):
         self._thread.join()
         with self._status_lock:
             self._is_running = False
-        return not self._thread.is_alive()
+        return not self._thread.is_alive()
-
-
-# class VAD:
-
-#     def __init__(
-#         self,
-#         VAD_model=None,
-#         audio_config: AudioBinary_Config = None,
-#         callback: Callable = None,
-#     ):
-#         # vad model
-#         self.VAD_model = VAD_model
-#         if self.VAD_model is None:
-#             self.VAD_model = AutoModel(
-#                 model="fsmn-vad", model_revision="v2.0.4", disable_update=True
-#             )
-#         # audio config
-#         self.audio_config = audio_config
-#         # vad result
-#         self.vad_result = VADResponse(time_chunk_index_callback=callback)
-#         # audio binary poll
-#         self.audio_chunk = AudioChunk(audio_config=self.audio_config)
-#         self.cache = {}
-
-#     def push_binary_data(
-#         self,
-#         binary_data: bytes,
-#     ):
-#         # 压入二进制数据
-#         self.audio_chunk.add_chunk(binary_data)
-#         # 处理音频块
-#         res = self.VAD_model.generate(
-#             input=binary_data,
-#             cache=self.cache,
-#             chunk_size=self.audio_config.chunk_size,
-#             is_final=False,
-#         )
-#         # print("VAD generate", res)
-#         if len(res[0]["value"]):
-#             self.vad_result += VADResponse.from_raw(res)
-
-#     def set_callback(
-#         self,
-#         callback: Callable,
-#     ):
-#         self.vad_result.time_chunk_index_callback = callback
-
-#     def process_vad_result(self, callback: Callable = None):
-#         # 处理VAD结果
-#         callback = (
-#             callback
-#             if callback is not None
-#             else self.vad_result.time_chunk_index_callback
-#         )
-#         self.vad_result.process_time_chunk(
-#             lambda x: callback(
-#                 AudioBinary_Chunk(
-#                     start_time=x["start_time"],
-#                     end_time=x["end_time"],
-#                     chunk=self.audio_chunk.get_chunk(x["start_time"], x["end_time"]),
-#                 )
-#             )
-#         )

src/models/spk.py

@@ -1,5 +1,5 @@
 """
-src/schemas/speaker.py
+src/models/spk.py
 ------------------------
 此模块定义与说话人（speakers）表对应的 Pydantic 模型，用于 API 数据验证和序列化。
 
@@ -14,10 +14,12 @@ from datetime import datetime
 from typing import Optional, List
 from uuid import UUID
 from pydantic import BaseModel, Field
-from .base import BaseSchema
+from src.utils import get_module_logger
+
+logger = get_module_logger(__name__)
 
 # 基础模型，定义说话人的核心属性
-class SpeakerBase(BaseSchema):
+class SpeakerBase(BaseModel):
     speaker_id: UUID = Field(
         ..., 
         description="说话人唯一标识符"

src/pipeline/ASRpipeline.py

@@ -224,7 +224,7 @@ class ASRPipeline(PipelineBase):
         while self._is_running and not self._stop_event:
             try:
                 data = self._input_queue.get(timeout=self._queue_timeout)
-                logger.debug("[ASRpipeline]获取到的数据length: %s", len(data))
+                # logger.debug("[ASRpipeline]获取到的数据length: %s", len(data))
                 # 检查是否是结束信号
                 if data is None:
                     logger.info("收到结束信号，管道准备停止")

src/runner/ASRRunner.py

@@ -105,7 +105,7 @@ class ASRRunner(RunnerBase):
                 data = self._sender.recv()
                 if data is None:
                     break
-                logger.debug("[ASRRunner][SAR-%s]接收到的数据length: %s", self._name, len(data))
+                # logger.debug("[ASRRunner][SAR-%s]接收到的数据length: %s", self._name, len(data))
                 self._input_queue.put(data)
             self.stop()
 

src/utils/mock_websocket.py

@@ -47,7 +47,7 @@ class MockWebSocketClient:
         """Puts data into the receive queue for the `recv` method to consume."""
         if data is None:
             return
-        logger.debug("Mock WS put_for_recv length: %s", len(data))
+        # logger.debug("Mock WS put_for_recv length: %s", len(data))
         self.receive_queue.put(data)
 
     @property

src/websockets/endpoint/asr_endpoint.py

@@ -1,32 +0,0 @@
-"""
- -*- coding: utf-8 -*-
- 此模块是ASR的websocket端点, 使用FastAPI的websocket端点
-"""
-
-from fastapi import WebSocket, APIRouter
-
-router = APIRouter()
-
-from src.runner.ASRRunner import ASRRunner
-
-ASRRunner_instance = ASRRunner()
-
-from src.core import ModelLoader
-model_loader = ModelLoader()
-args = {
-    "asr_model": "paraformer-zh",
-    "asr_model_revision": "v2.0.4",
-    "vad_model": "fsmn-vad",
-    "vad_model_revision": "v2.0.4",
-    "spk_model": "cam++",
-    "spk_model_revision": "v2.0.2",
-    "audio_update": False,
-}
-models = model_loader.load_models(args)
-
-@router.websocket("/asr_full")
-async def asr_endpoint(websocket: WebSocket):
-    await websocket.accept()
-    while True:
-        data = await websocket.receive_text()
-        print(data)

test_main.py

@@ -10,6 +10,10 @@ from tests.runner.asr_runner_test import test_asr_runner
 setup_root_logger(level="INFO", log_file="logs/test_main.log")
 logger = get_module_logger(__name__)
 
+# 清空logs/test_main.log文件
+with open("logs/test_main.log", "w") as f:
+    f.truncate()
+
 # from tests.functor.vad_test import test_vad_functor
 # logger.info("开始测试VAD函数器")
 # test_vad_functor()

tests/runner/asr_runner_test.py

@@ -10,6 +10,10 @@ from src.core.model_loader import ModelLoader
 from src.models import AudioBinary_Config
 from src.utils.mock_websocket import MockWebSocketClient
 
+from src.utils.logger import get_module_logger
+
+logger = get_module_logger(__name__)
+
 def test_asr_runner():
     """
     End-to-end test for ASRRunner.
@@ -32,17 +36,18 @@ def test_asr_runner():
         "audio_update": False,
     }
     models = model_loader.load_models(args)
-    audio_data, sample_rate = soundfile.read("tests/vad_example.wav")
+    audio_file_path = "tests/XT_ZZY_denoise.wav"
-
+    audio_data, sample_rate = soundfile.read(audio_file_path)
+    logger.info(f"加载数据: {audio_file_path} , audio_data_length: {len(audio_data)}, audio_data_type: {type(audio_data)}, sample_rate: {sample_rate}")
     # 2. Configure audio
     audio_config = AudioBinary_Config(
         chunk_size=200,  # ms
-        chunk_stride=1600,  # 10ms stride for 16kHz
+        chunk_stride=1000,  # 10ms stride for 16kHz
         sample_rate=sample_rate,
         sample_width=2,  # 16-bit
-        channels=1,
+        channels=2,
     )
-    audio_config.chunk_stride = int(audio_config.chunk_stride * sample_rate / 1000)
+    audio_config.chunk_stride = int(audio_config.chunk_size * sample_rate / 1000)
 
     # 3. Setup ASRRunner
     asr_runner = ASRRunner()
@@ -70,6 +75,6 @@ def test_asr_runner():
         mock_ws.put_for_recv(chunk)
 
     # 6. Wait for results and assert
-    time.sleep(10)
+    time.sleep(30)
     # Signal end of audio stream by sending None
     mock_ws.put_for_recv(None)