Update app.py

app.py

@@ -1,24 +1,22 @@
-import gradio as gr
 import os
 import tempfile
 from openai import OpenAI
 from tts_voice import tts_order_voice
 import edge_tts
-import tempfile
 import anyio
-
-
 import torch
 import torchaudio
 import gradio as gr
 from scipy.io import wavfile
 from scipy.io.wavfile import write
 
-
+# 创建 KNN-VC 模型
 knn_vc = torch.hub.load('bshall/knn-vc', 'knn_vc', prematched=True, trust_repo=True, pretrained=True, device='cpu')
 
+# 初始化 language_dict
 language_dict = tts_order_voice
 
+# 异步文字转语音函数
 async def text_to_speech_edge(text, language_code):
     voice = language_dict[language_code]
     communicate = edge_tts.Communicate(text, voice)
@@ -29,7 +27,7 @@ async def text_to_speech_edge(text, language_code):
 
     return "语音合成完成：{}".format(text), tmp_path
 
-
+# 声音更改函数
 def voice_change(audio_in, audio_ref):
     samplerate1, data1 = wavfile.read(audio_in)
     samplerate2, data2 = wavfile.read(audio_ref)
@@ -42,37 +40,29 @@ def voice_change(audio_in, audio_ref):
     torchaudio.save('output.wav', out_wav[None], 16000)
     return 'output.wav'
 
-
-def tts(text, model, voice, api_key,base_url):
-    if len(text)>300:
+# 文字转语音（OpenAI）
+def tts(text, model, voice, api_key):
+    if len(text) > 300:
         raise gr.Error('您输入的文本字符多于300个，请缩短您的文本')
     if api_key == '':
-        raise gr.Error('Please enter your OpenAI API Key')
-    else:
-        try:
-            client = OpenAI(api_key=api_key,base_url='https://lmzh.top/v1')
-
-            response = client.audio.speech.create(
-                model=model, # "tts-1","tts-1-hd"
-                voice=voice, # 'alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer'
-                input=text,
-            )
-
-        except Exception as error:
-            # Handle any exception that occurs
-            raise gr.Error("An error occurred while generating speech. Please check your API key and try again.")
-            print(str(error))
-
-    # Create a temp file to save the audio
+        raise gr.Error('请填写您的 OpenAI API Key')
+
+    try:
+        client = OpenAI(api_key=api_key, base_url='https://lmzh.top/v1')
+        response = client.audio.speech.create(
+            model=model,
+            voice=voice,
+            input=text,
+        )
+    except Exception as error:
+        raise gr.Error(f"生成语音时出错：{error}")
+
     with tempfile.NamedTemporaryFile(suffix=".mp3", delete=False) as temp_file:
         temp_file.write(response.content)
 
-    # Get the file path of the temp file
-    temp_file_path = temp_file.name
-        
-    return temp_file_path
-
+    return temp_file.name
 
+# Gradio 前端设计
 app = gr.Blocks()
 
 with app:
@@ -80,22 +70,21 @@ with app:
     gr.Markdown("### <center>🎶 地表最强文本转语音模型 + 3秒实时AI变声，支持中文！Powered by [OpenAI TTS](https://platform.openai.com/docs/guides/text-to-speech) and [KNN-VC](https://github.com/bshall/knn-vc) </center>")
     with gr.Tab("🤗 OpenAI TTS"):
         with gr.Row(variant='panel'):
-          api_key = gr.Textbox(type='password', label='OpenAI API Key', placeholder='请在此填写您的OpenAI API Key')
-          model = gr.Dropdown(choices=['tts-1','tts-1-hd'], label='请选择模型（tts-1推理更快，tts-1-hd音质更好）', value='tts-1')
-          voice = gr.Dropdown(choices=['alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer'], label='请选择一个说话人', value='alloy')
+            api_key = gr.Textbox(type='password', label='OpenAI API Key', placeholder='请在此填写您的OpenAI API Key')
+            model = gr.Dropdown(choices=['tts-1','tts-1-hd'], label='请选择模型（tts-1推理更快，tts-1-hd音质更好）', value='tts-1')
+            voice = gr.Dropdown(choices=['alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer'], label='请选择一个说话人', value='alloy')
         with gr.Row():
-          with gr.Column():
-            inp_text = gr.Textbox(label="请填写您想生成的文本（中英文皆可）", placeholder="想说却还没说的 还很多 攒着是因为想写成歌", lines=5)
-            btn_text = gr.Button("一键开启真实拟声吧", variant="primary")
-    
-          with gr.Column():
-            inp1 = gr.Audio(type="filepath", label="OpenAI TTS真实拟声", interactive=False)
-            inp2 = gr.Audio(type="filepath", label="请上传AI变声的参照音频（决定变声后的语音音色）")
-            btn1 = gr.Button("一键开启AI变声吧", variant="primary")
-          with gr.Column():
-            out1 = gr.Audio(type="filepath", label="AI变声后的专属音频")
-          btn_text.click(tts, [inp_text, model, voice, api_key, base_url], inp1)
-          btn1.click(voice_change, [inp1, inp2], out1)
+            with gr.Column():
+                inp_text = gr.Textbox(label="请填写您想生成的文本（中英文皆可）", placeholder="想说却还没说的 还很多 攒着是因为想写成歌", lines=5)
+                btn_text = gr.Button("一键开启真实拟声吧", variant="primary")
+            with gr.Column():
+                inp1 = gr.Audio(type="filepath", label="OpenAI TTS真实拟声", interactive=False)
+                inp2 = gr.Audio(type="filepath", label="请上传AI变声的参照音频（决定变声后的语音音色）")
+                btn1 = gr.Button("一键开启AI变声吧", variant="primary")
+            with gr.Column():
+                out1 = gr.Audio(type="filepath", label="AI变声后的专属音频")
+            btn_text.click(tts, [inp_text, model, voice, api_key], inp1)
+            btn1.click(voice_change, [inp1, inp2], out1)
     with gr.Tab("⚡ Edge TTS"):
         with gr.Row():
             input_text = gr.Textbox(lines=5, placeholder="想说却还没说的 还很多 攒着是因为想写成歌", label="请填写您想生成的文本（中英文皆可）")
@@ -110,10 +99,9 @@ with app:
             btn_vc = gr.Button("一键开启AI变声吧", variant="primary")
             out_vc = gr.Audio(type="filepath", label="AI变声后的专属音频")
 
-        btn_edge.click(text_to_speech_edge, [input_text, language], [output_text, output_audio])
+        btn_edge.click(lambda text, lang: anyio.run(text_to_speech_edge, text, lang), [input_text, language], [output_text, output_audio])
         btn_vc.click(voice_change, [output_audio, inp_vc], out_vc)
 
-        
     gr.Markdown("### <center>注意❗：请不要生成会对个人以及组织造成侵害的内容，此程序仅供科研、学习及个人娱乐使用。Get your OpenAI API Key [here](https://platform.openai.com/api-keys).</center>")
     gr.HTML('''
         <div class="footer">
@@ -121,4 +109,4 @@ with app:
         </div>
     ''')
 
-app.launch(share=True)
+app.launch(share=True, show_error=True)