Update app.py

app.py

@@ -43,7 +43,7 @@ _ = utils.load_checkpoint("logs/leo/G_4000.pth", net_g, None)
 random_emotion_root = "wavs"
 emotion_dict = json.load(open("configs/leo.json", "r"))
 
-def tts(txt, emotion, roma=False, length_scale=1):
+def tts(txt, emotion, temp_file_path):
     """emotion为参考情感音频路径或random_sample（随机抽取）"""
     if roma:
         stn_tst = get_text_byroma(txt, hps)
@@ -68,27 +68,17 @@ def tts(txt, emotion, roma=False, length_scale=1):
         else:
             print("emotion参数不正确")
 
+        audio = net_g.infer(x_tst, x_tst_lengths, sid=sid, noise_scale=0.667, noise_scale_w=0.8, length_scale=1.2, emo=emo)[0][0,0].data.float().numpy()
+
         # Save the numpy array as a temporary file
-        temp_file = tempfile.NamedTemporaryFile(suffix=".wav", delete=False)
-        temp_file_path = temp_file.name
         write(temp_file_path, hps.data.sampling_rate, audio)
 
-    # Display the audio
-    ipd.display(ipd.Audio(temp_file_path, rate=hps.data.sampling_rate, normalize=False))
-
-    # Delete the temporary file
-    os.remove(temp_file_path)
+        # Display the audio
+        ipd.display(ipd.Audio(temp_file_path, rate=hps.data.sampling_rate, normalize=False))
 
-    audio = net_g.infer(x_tst, x_tst_lengths, sid=sid, noise_scale=0.667, noise_scale_w=0.8, length_scale=1.2, emo=emo)[0][0,0].data.float().numpy()
-    return audio
+        # Delete the temporary file
+        os.remove(temp_file_path)
 
-def generate_audio(txt, emotion):
-    audio = tts(txt, emotion)
     return audio
 
-input_text = gr.inputs.Textbox(label="输入文本")
-input_emotion = gr.inputs.Dropdown(choices=["random_sample"] + os.listdir(random_emotion_root), label="参考情感音频")
-output_audio = gr.outputs.Audio(type="numpy", label="合成音频")
-
-iface = gr.Interface(fn=generate_audio, inputs=[input_text, input_emotion], outputs=output_audio)
-iface.launch()
+def generate