Update app.py

app.py

@@ -1,9 +1,9 @@
 import streamlit as st
 from huggingface_hub import InferenceClient
-from gtts import gTTS
 import base64
 from pydub import AudioSegment
 from pydub.playback import play
+import pyttsx3
 
 client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
 pre_prompt = ""
@@ -24,16 +24,12 @@ def format_prompt(message, history):
     prompt += f"[INST] {message} [/INST]"
     return prompt
 
-def text_to_speech(text, speed=1.3):
-    tts = gTTS(text=text, lang='es')
-    audio_file_path = 'output.mp3'
-    tts.save(audio_file_path)
-    sound = AudioSegment.from_mp3(audio_file_path)
-    sound = sound.speedup(playback_speed=speed)
-    sound.export(audio_file_path, format="mp3")
-
-    return audio_file_path
-
+def text_to_speech(text):
+    engine = pyttsx3.init()
+    engine.save_to_file(text, "output_pyttsx3.mp3")
+    engine.runAndWait()
+    return "output_pyttsx3.mp3"
+    
 def generate(user_input, history, temperature=None, max_new_tokens=512, top_p=0.95, repetition_penalty=1.0):
     global pre_prompt_sent
     temperature = float(temperature) if temperature is not None else 0.9
@@ -70,12 +66,7 @@ if "history" not in st.session_state:
 
 user_input = st.text_input(label="", value="")
 output, audio_bytes = generate(user_input, history=st.session_state.history)
-
-if user_input:
-    st.session_state.history.append((user_input, output))
-
 st.text_area("Respuesta", value=output, height=400, key="output_text", disabled=True)
-st.text_input(label="Escribe aquí:", value=user_input, key="user_input", disabled=True)
 
 st.markdown(
     f"""