Update app.py

app.py

@@ -3,27 +3,31 @@ import base64
 import io
 from huggingface_hub import InferenceClient
 from gtts import gTTS
+from audiorecorder import audiorecorder
 import speech_recognition as sr
 
-def capture_audio():
+def recognize_speech(audio_data, show_messages=True):
     recognizer = sr.Recognizer()
+    audio_recording = sr.AudioFile(audio_data)
+
+    with audio_recording as source:
+        audio = recognizer.record(source)
+
+    try:
+        audio_text = recognizer.recognize_google(audio, language="es-ES")
+        if show_messages:
+            st.subheader("Texto Reconocido:")
+            st.write(audio_text)
+            st.success("Reconocimiento de voz completado.")
+    except sr.UnknownValueError:
+        st.warning("No se pudo reconocer el audio. ¿Intentaste grabar algo?")
+        audio_text = ""
+    except sr.RequestError:
+        st.error("No he recibido ningun audio. Por favor, inténtalo de nuevo.")
+        audio_text = ""
+
+    return audio_text
 
-    with sr.Microphone() as source:
-        st.write("Di algo...")
-        recognizer.adjust_for_ambient_noise(source)
-        audio = recognizer.listen(source, timeout=5)
-
-        try:
-            text = recognizer.recognize_google(audio, language="es-ES")
-            st.write("Has dicho: " + text)
-            return audio, text
-        except sr.UnknownValueError:
-            st.write("No se pudo entender el audio")
-        except sr.RequestError as e:
-            st.write("Error al solicitar los resultados: {0}".format(e))
-
-    return None, None
-    
 def format_prompt(message, history):
     prompt = "<s>"
 
@@ -73,27 +77,31 @@ def main():
     if "history" not in st.session_state:
         st.session_state.history = []
 
-    st.beta_container()
-    audio_data, audio_text = capture_audio()
-
-    if not st.session_state.history:
-        pre_prompt = "Te Llamarás Chaman 4.0 y tus respuestas serán sumamente breves."
-        output, _ = generate(pre_prompt, history=st.session_state.history)
-        st.session_state.history.append((pre_prompt, output))
+    audio_data = audiorecorder("Habla para grabar", "Deteniendo la grabación...")
 
-    if audio_text:
-        output, audio_file = generate(audio_text, history=st.session_state.history)
+    if not audio_data.empty():
+        st.audio(audio_data.export().read(), format="audio/wav")
+        audio_data.export("audio.wav", format="wav")
+        audio_text = recognize_speech("audio.wav")
 
+        if not st.session_state.history:
+            pre_prompt = "Te Llamarás Chaman 4.0 y tus respuestas serán sumamente breves."
+            output, _ = generate(pre_prompt, history=st.session_state.history)
+            st.session_state.history.append((pre_prompt, output))
+        
         if audio_text:
-            st.session_state.history.append((audio_text, output))
-
-        if audio_file is not None:
-            st.markdown(
-                f"""
-                <audio autoplay="autoplay" controls="controls" src="data:audio/mp3;base64,{base64.b64encode(audio_file.read()).decode()}" type="audio/mp3" id="audio_player"></audio>
-                """,
-                unsafe_allow_html=True
-            )
+            output, audio_file = generate(audio_text, history=st.session_state.history)  
+
+            if audio_text:  
+                st.session_state.history.append((audio_text, output))  
+
+            if audio_file is not None:
+                st.markdown(
+                    f"""
+                    <audio autoplay="autoplay" controls="controls" src="data:audio/mp3;base64,{base64.b64encode(audio_file.read()).decode()}" type="audio/mp3" id="audio_player"></audio>
+                    """,
+                    unsafe_allow_html=True
+                )
 
 if __name__ == "__main__":
     main()