fix: fix bailian stt

Author: wxg0103

apps/setting/models_provider/impl/aliyun_bai_lian_model_provider/model/stt.py

@@ -4,6 +4,7 @@
 
 import dashscope
 from dashscope.audio.asr import (Recognition)
+from pydub import AudioSegment
 
 from setting.models_provider.base_model_provider import MaxKBBaseModel
 from setting.models_provider.impl.base_stt import BaseSpeechToText
@@ -40,7 +41,6 @@ def speech_to_text(self, audio_file):
         dashscope.api_key = self.api_key
         recognition = Recognition(model=self.model,
                                   format='mp3',
-                                  sample_rate=16000,
                                   callback=None)
         with tempfile.NamedTemporaryFile(delete=False) as temp_file:
             # 将上传的文件保存到临时文件中
@@ -49,6 +49,13 @@ def speech_to_text(self, audio_file):
             temp_file_path = temp_file.name
 
         try:
+            audio = AudioSegment.from_file(temp_file_path)
+            if audio.channels != 1:
+                audio = audio.set_channels(1)
+            audio = audio.set_frame_rate(16000)
+
+            # 将转换后的音频文件保存到临时文件中
+            audio.export(temp_file_path, format='mp3')
             # 识别临时文件
             result = recognition.call(temp_file_path)
             text = ''