Update app.py

app.py

@@ -1,5 +1,4 @@
 import streamlit as st
-import speech_recognition as sr
 from transformers import pipeline
 from gtts import gTTS
 import requests
@@ -8,6 +7,9 @@ import os
 # Initialize the text generation pipeline with GPT-2
 generator = pipeline('text-generation', model='gpt2')
 
+# Initialize Whisper model for transcription
+transcriber = pipeline("automatic-speech-recognition", model="openai/whisper-base")
+
 # Function to generate a Trump-like response
 def generate_response(prompt):
     # This is your original `generate_response` function
@@ -25,7 +27,6 @@ UBERDUCK_API_SECRET = 'pk_8e4b3ad9-d26a-49ca-ac83-4ddfe4dace85'  # Replace with
 
 # Function to generate audio using Uberduck API
 def generate_audio_uberduck(text):
-    # This is your original Uberduck API function
     url = "https://api.uberduck.ai/speak"
     headers = {
         "Accept": "application/json",
@@ -34,7 +35,7 @@ def generate_audio_uberduck(text):
     }
     json_data = {
         "speech": text,
-        "voice": "donald-trump"  # Use "donald-trump" voice in Uberduck (if available)
+        "voice": "donald-trump"
     }
     
     # Send a request to Uberduck's API to generate audio
@@ -57,17 +58,11 @@ def generate_audio_gtts(text):
     tts.save(audio_path)
     return audio_path
 
-# Function to transcribe audio to text
+# Function to transcribe audio to text using Whisper
 def transcribe_audio(file_path):
-    recognizer = sr.Recognizer()
-    try:
-        with sr.AudioFile(file_path) as source:
-            audio = recognizer.record(source)
-            text = recognizer.recognize_google(audio)
-            return text
-    except Exception as e:
-        st.write("Error transcribing audio:", str(e))
-        return None
+    # Transcribe audio using Whisper model
+    transcription = transcriber(file_path)
+    return transcription['text']
 
 # Streamlit app UI
 st.title("Trump-like Voice Assistant")
@@ -93,7 +88,6 @@ if uploaded_audio is not None:
         st.write("Trump-like Assistant:", trump_response)
         
         # Generate audio response (choose either Uberduck or gTTS)
-        # Uncomment one of the following two lines depending on your preferred TTS provider
         audio_output_path = generate_audio_uberduck(trump_response)  # Uberduck
         # audio_output_path = generate_audio_gtts(trump_response)     # gTTS as fallback