fix lint

runtime/python/fastapi/client.py

@@ -38,7 +38,7 @@ def main():
         payload = {
             'tts_text': args.tts_text,
         }
-        files = [('prompt_wav', ('prompt_wav', open(args.prompt_wav,'rb'), 'application/octet-stream'))]
+        files = [('prompt_wav', ('prompt_wav', open(args.prompt_wav, 'rb'), 'application/octet-stream'))]
         response = requests.request("GET", url, data=payload, files=files, stream=True)
     else:
         payload = {
@@ -55,6 +55,7 @@ def main():
     torchaudio.save(args.tts_wav, tts_speech, target_sr)
     logging.info('get response')
 
+
 if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     parser.add_argument('--host',
@@ -81,7 +82,8 @@ if __name__ == "__main__":
                         default='../../../zero_shot_prompt.wav')
     parser.add_argument('--instruct_text',
                         type=str,
-                        default='Theo \'Crimson\', is a fiery, passionate rebel leader. Fights with fervor for justice, but struggles with impulsiveness.')
+                        default='Theo \'Crimson\', is a fiery, passionate rebel leader. \
+                                 Fights with fervor for justice, but struggles with impulsiveness.')
     parser.add_argument('--tts_wav',
                         type=str,
                         default='demo.wav')

runtime/python/fastapi/server.py

@@ -13,9 +13,6 @@
 # limitations under the License.
 import os
 import sys
-ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
-sys.path.append('{}/../../..'.format(ROOT_DIR))
-sys.path.append('{}/../../../third_party/Matcha-TTS'.format(ROOT_DIR))
 import argparse
 import logging
 logging.getLogger('matplotlib').setLevel(logging.WARNING)
@@ -24,6 +21,9 @@ from fastapi.responses import StreamingResponse
 from fastapi.middleware.cors import CORSMiddleware
 import uvicorn
 import numpy as np
+ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
+sys.path.append('{}/../../..'.format(ROOT_DIR))
+sys.path.append('{}/../../../third_party/Matcha-TTS'.format(ROOT_DIR))
 from cosyvoice.cli.cosyvoice import CosyVoice
 from cosyvoice.utils.file_utils import load_wav
 
@@ -36,34 +36,40 @@ app.add_middleware(
     allow_methods=["*"],
     allow_headers=["*"])
 
+
 def generate_data(model_output):
     for i in model_output:
         tts_audio = (i['tts_speech'].numpy() * (2 ** 15)).astype(np.int16).tobytes()
         yield tts_audio
 
+
 @app.get("/inference_sft")
 async def inference_sft(tts_text: str = Form(), spk_id: str = Form()):
     model_output = cosyvoice.inference_sft(tts_text, spk_id)
     return StreamingResponse(generate_data(model_output))
 
+
 @app.get("/inference_zero_shot")
 async def inference_zero_shot(tts_text: str = Form(), prompt_text: str = Form(), prompt_wav: UploadFile = File()):
     prompt_speech_16k = load_wav(prompt_wav.file, 16000)
     model_output = cosyvoice.inference_zero_shot(tts_text, prompt_text, prompt_speech_16k)
     return StreamingResponse(generate_data(model_output))
 
+
 @app.get("/inference_cross_lingual")
 async def inference_cross_lingual(tts_text: str = Form(), prompt_wav: UploadFile = File()):
     prompt_speech_16k = load_wav(prompt_wav.file, 16000)
     model_output = cosyvoice.inference_cross_lingual(tts_text, prompt_speech_16k)
     return StreamingResponse(generate_data(model_output))
 
+
 @app.get("/inference_instruct")
 async def inference_instruct(tts_text: str = Form(), spk_id: str = Form(), instruct_text: str = Form()):
     model_output = cosyvoice.inference_instruct(tts_text, spk_id, instruct_text)
     return StreamingResponse(generate_data(model_output))
 
-if __name__=='__main__':
+
+if __name__ == '__main__':
     parser = argparse.ArgumentParser()
     parser.add_argument('--port',
                         type=int,
@@ -74,4 +80,4 @@ if __name__=='__main__':
                         help='local path or modelscope repo id')
     args = parser.parse_args()
     cosyvoice = CosyVoice(args.model_dir)
-    uvicorn.run(app, host="127.0.0.1", port=args.port)
+    uvicorn.run(app, host="127.0.0.1", port=args.port)

runtime/python/grpc/client.py

@@ -96,7 +96,8 @@ if __name__ == "__main__":
                         default='../../../zero_shot_prompt.wav')
     parser.add_argument('--instruct_text',
                         type=str,
-                        default='Theo \'Crimson\', is a fiery, passionate rebel leader. Fights with fervor for justice, but struggles with impulsiveness.')
+                        default='Theo \'Crimson\', is a fiery, passionate rebel leader. \
+                                 Fights with fervor for justice, but struggles with impulsiveness.')
     parser.add_argument('--tts_wav',
                         type=str,
                         default='demo.wav')

runtime/python/grpc/server.py

@@ -13,9 +13,6 @@
 # limitations under the License.
 import os
 import sys
-ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
-sys.path.append('{}/../../..'.format(ROOT_DIR))
-sys.path.append('{}/../../../third_party/Matcha-TTS'.format(ROOT_DIR))
 from concurrent import futures
 import argparse
 import cosyvoice_pb2
@@ -25,11 +22,15 @@ logging.getLogger('matplotlib').setLevel(logging.WARNING)
 import grpc
 import torch
 import numpy as np
+ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
+sys.path.append('{}/../../..'.format(ROOT_DIR))
+sys.path.append('{}/../../../third_party/Matcha-TTS'.format(ROOT_DIR))
 from cosyvoice.cli.cosyvoice import CosyVoice
 
 logging.basicConfig(level=logging.DEBUG,
                     format='%(asctime)s %(levelname)s %(message)s')
 
+
 class CosyVoiceServiceImpl(cosyvoice_pb2_grpc.CosyVoiceServicer):
     def __init__(self, args):
         self.cosyvoice = CosyVoice(args.model_dir)
@@ -43,7 +44,9 @@ class CosyVoiceServiceImpl(cosyvoice_pb2_grpc.CosyVoiceServicer):
             logging.info('get zero_shot inference request')
             prompt_speech_16k = torch.from_numpy(np.array(np.frombuffer(request.zero_shot_request.prompt_audio, dtype=np.int16))).unsqueeze(dim=0)
             prompt_speech_16k = prompt_speech_16k.float() / (2**15)
-            model_output = self.cosyvoice.inference_zero_shot(request.zero_shot_request.tts_text, request.zero_shot_request.prompt_text, prompt_speech_16k)
+            model_output = self.cosyvoice.inference_zero_shot(request.zero_shot_request.tts_text,
+                                                              request.zero_shot_request.prompt_text,
+                                                              prompt_speech_16k)
         elif request.HasField('cross_lingual_request'):
             logging.info('get cross_lingual inference request')
             prompt_speech_16k = torch.from_numpy(np.array(np.frombuffer(request.cross_lingual_request.prompt_audio, dtype=np.int16))).unsqueeze(dim=0)
@@ -51,7 +54,9 @@ class CosyVoiceServiceImpl(cosyvoice_pb2_grpc.CosyVoiceServicer):
             model_output = self.cosyvoice.inference_cross_lingual(request.cross_lingual_request.tts_text, prompt_speech_16k)
         else:
             logging.info('get instruct inference request')
-            model_output = self.cosyvoice.inference_instruct(request.instruct_request.tts_text, request.instruct_request.spk_id, request.instruct_request.instruct_text)
+            model_output = self.cosyvoice.inference_instruct(request.instruct_request.tts_text,
+                                                             request.instruct_request.spk_id,
+                                                             request.instruct_request.instruct_text)
 
         logging.info('send inference response')
         for i in model_output:
@@ -59,6 +64,7 @@ class CosyVoiceServiceImpl(cosyvoice_pb2_grpc.CosyVoiceServicer):
             response.tts_audio = (i['tts_speech'].numpy() * (2 ** 15)).astype(np.int16).tobytes()
             yield response
 
+
 def main():
     grpcServer = grpc.server(futures.ThreadPoolExecutor(max_workers=args.max_conc), maximum_concurrent_rpcs=args.max_conc)
     cosyvoice_pb2_grpc.add_CosyVoiceServicer_to_server(CosyVoiceServiceImpl(args), grpcServer)