Update app.py

app.py

@@ -5,6 +5,8 @@ from transformers import AutoConfig
 from transformers import GPT2Tokenizer, GPT2LMHeadModel
 from itertools import chain
 
+import os 
+
 import tempfile
 from typing import Optional
 from TTS.config import load_config
@@ -99,34 +101,9 @@ def get_chat_response(name,history=[], input_txt = "Hello , what is your name?")
 
 MODEL_NAME= "tts_models/multilingual/multi-dataset/your_tts"
 
-def tts(text: str, speaker_idx: str=None):
-    if len(text) > MAX_TXT_LEN:
-        text = text[:MAX_TXT_LEN]
-        print(f"Input text was cutoff since it went over the {MAX_TXT_LEN} character limit.")
-    print(text, model_name)
-    # download model
-    model_path, config_path, model_item = manager.download_model(f"tts_models/{MODEL_NAME}")
-    vocoder_name: Optional[str] = model_item["default_vocoder"]
-    # download vocoder
-    vocoder_path = None
-    vocoder_config_path = None
-    if vocoder_name is not None:
-        vocoder_path, vocoder_config_path, _ = manager.download_model(vocoder_name)
-    # init synthesizer
-    synthesizer = Synthesizer(
-        model_path, config_path, None, None, vocoder_path, vocoder_config_path,
-    )
-    # synthesize
-    if synthesizer is None:
-        raise NameError("model not found")
-    wavs = synthesizer.tts(text, speaker_idx)
-    # return output
-    with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
-        synthesizer.save_wav(wavs, fp)
-        return fp.name
-
-
-def greet(character,message,history,voice):
+
+
+def greet(character,audio,message,history,voice):
   
   #gradios set_state/get_state had problems on embedded html!
   history = history or {"character": character, "message_history" : [] }
@@ -139,8 +116,9 @@ def greet(character,message,history,voice):
 
 
   response = get_chat_response(character,history=history["message_history"],input_txt=message)
-  voice = tts(response)
-  
+  os.system('tts --text "'+response+'" --model_name tts_models/multilingual/multi-dataset/your_tts --speaker_wav '+audio+' --language_idx "es"') 
+  voice = "tts_output.wav"
+
   history["message_history"].append((message, response))
   
   
@@ -177,7 +155,7 @@ article = "<p style='text-align: center'><a href='https://www.linkedin.com/pulse
 
 history =   {"character": "None", "message_history" : [] }
 interface= gr.Interface(fn=greet, 
-                        inputs=[gr.Audio(source="microphone", type="filepath"),gr.inputs.Dropdown(personality_choices) ,"text", "state"], 
+                        inputs=[gr.inputs.Dropdown(personality_choices),gr.Audio(source="microphone", type="filepath") ,"text", "state"], 
                         outputs=["html","state",gr.Audio(type="filepath")],
                         
                         css=css, title=title, description=description,article=article )