fix: google speech_recognition

app.py

@@ -2,6 +2,7 @@
 import streamlit as st
 import whisper
 import speech_recognition as sr
+from pydub import AudioSegment
 import os
 
 # Function to transcribe audio using OpenAI Whisper
@@ -19,6 +20,13 @@ def transcribe_speech_recognition(file_path):
     result = r.recognize_google(audio)
     return result
 
+# Function to convert mp3 file to wav
+def convert_mp3_to_wav(mp3_path):
+    audio = AudioSegment.from_mp3(mp3_path)
+    wav_path = mp3_path.replace('.mp3', '.wav')
+    audio.export(wav_path, format="wav")
+    return wav_path
+
 # Streamlit App
 st.title('Transcriptor de Audio')
 
@@ -27,12 +35,10 @@ uploaded_file = st.file_uploader("Sube tu archivo de audio para transcribir", ty
 if uploaded_file is not None:
     file_details = {"FileName":uploaded_file.name, "FileType":uploaded_file.type, "FileSize":uploaded_file.size}
     st.write(file_details)
-    
-    # Make sure the temp directory exists
-    if not os.path.exists('temp'):
-        os.makedirs('temp')
 
-    with open(os.path.join("temp",uploaded_file.name), "wb") as f:
+    # Save uploaded file to temp directory
+    file_path = os.path.join("temp", uploaded_file.name)
+    with open(file_path, "wb") as f:
         f.write(uploaded_file.getbuffer())
     
     st.write("Archivo de audio cargado correctamente. Por favor, selecciona el método de transcripción.")
@@ -40,11 +46,14 @@ if uploaded_file is not None:
 
     if transcription_method == 'OpenAI Whisper':
         model_name = st.selectbox('Escoge el modelo de Whisper', ('base', 'small', 'medium', 'large', 'tiny'))
+    elif transcription_method == 'Google Speech API' and file_path.endswith('.mp3'):
+        # Convert mp3 to wav if Google Speech API is selected and file is in mp3 format
+        file_path = convert_mp3_to_wav(file_path)
 
     if st.button('Transcribir'):
         if transcription_method == 'OpenAI Whisper':
-            transcript = transcribe_whisper(model_name, os.path.join("temp",uploaded_file.name))
+            transcript = transcribe_whisper(model_name, file_path)
         else:
-            transcript = transcribe_speech_recognition(os.path.join("temp",uploaded_file.name))
+            transcript = transcribe_speech_recognition(file_path)
 
         st.write(transcript)