Update app.py

app.py

@@ -1,16 +1,10 @@
 import os
-import io
 import torch
 import whisper
 import streamlit as st
 from groq import Groq
 from dotenv import load_dotenv
 from tempfile import NamedTemporaryFile
-from streamlit_webrtc import webrtc_streamer, WebRtcMode, ClientSettings
-import av
-import numpy as np
-import uuid
-import time
 
 # Load environment variables
 load_dotenv()
@@ -97,7 +91,7 @@ def transcribe_audio(audio_path, model_size="base"):
 # Generate Speech using the configured XTTS model
 def generate_speech(text, output_file, speaker_wav, language="en"):
     if not os.path.exists(speaker_wav):
-        raise FileNotFoundError("Reference audio file not found. Please upload or record a valid audio.")
+        raise FileNotFoundError("Reference audio file not found. Please upload a valid audio.")
     
     if language not in supported_languages:
         st.warning(f"Language {language} is not supported. Defaulting to English.")
@@ -105,6 +99,7 @@ def generate_speech(text, output_file, speaker_wav, language="en"):
     
     # Use the configured model directly
     try:
+        import time
         t_latent = time.time()
         gpt_cond_latent, speaker_embedding = model.get_conditioning_latents(
             audio_path=speaker_wav, 
@@ -131,30 +126,6 @@ def generate_speech(text, output_file, speaker_wav, language="en"):
     except Exception as e:
         return False, f"Error generating speech: {str(e)}"
 
-# Audio Frame Processing for WebRTC
-class AudioProcessor:
-    def __init__(self):
-        self.audio_frames = []
-        self.sample_rate = 24000  # XTTS expects 24kHz
-
-    def recv(self, frame):
-        sound = frame.to_ndarray()
-        self.audio_frames.append(sound)
-        return frame
-
-    def save_audio(self, file_path):
-        if not self.audio_frames:
-            return None
-            
-        # Concatenate audio frames
-        concat_audio = np.concatenate(self.audio_frames, axis=0)
-        
-        # Save as WAV file
-        import soundfile as sf
-        sf.write(file_path, concat_audio, self.sample_rate)
-        
-        return file_path
-
 # Streamlit App
 def main():
     st.set_page_config(page_title="Vocal AI", layout="wide")
@@ -174,86 +145,37 @@ def main():
     # TOS agreement
     agree_tos = st.sidebar.checkbox("I agree to the Coqui Public Model License (CPML)", value=False)
     
-    # User option for reference audio (Record or Upload)
-    ref_audio_choice = st.sidebar.radio("Reference Audio", ("Upload", "Record"))
-
-    ref_audio_path = None
-    reference_audio_processor = None
-
+    import uuid
+    
     col1, col2 = st.columns(2)
     
     with col1:
         st.header("Step 1: Provide Reference Voice")
-        if ref_audio_choice == "Upload":
-            reference_audio = st.file_uploader("Upload Reference Audio", type=["wav", "mp3", "ogg"])
-            if reference_audio:
-                with NamedTemporaryFile(delete=False, suffix=".wav") as temp_ref_audio:
-                    temp_ref_audio.write(reference_audio.read())
-                    ref_audio_path = temp_ref_audio.name
-                st.audio(ref_audio_path)
-        else:
-            st.write("Record your reference voice:")
-            reference_audio_processor = AudioProcessor()
-            webrtc_ctx = webrtc_streamer(
-                key="ref_audio",
-                mode=WebRtcMode.SENDRECV,
-                client_settings=ClientSettings(
-                    rtc_configuration={"iceServers": [{"urls": ["stun:stun.l.google.com:19302"]}]},
-                    media_stream_constraints={"audio": True, "video": False},
-                ),
-                audio_receiver_size=1024,
-                video_processor_factory=None,
-                audio_processor_factory=lambda: reference_audio_processor,
-            )
-            
-            if webrtc_ctx.state.playing and reference_audio_processor is not None:
-                st.info("Recording... Speak into your microphone.")
-            
-            if st.button("Save Reference Audio"):
-                if reference_audio_processor and reference_audio_processor.audio_frames:
-                    with NamedTemporaryFile(delete=False, suffix=".wav") as temp_ref_audio:
-                        reference_audio_processor.save_audio(temp_ref_audio.name)
-                        ref_audio_path = temp_ref_audio.name
-                    st.success("Reference audio saved!")
-                    st.audio(ref_audio_path)
-                else:
-                    st.error("No audio recorded. Please speak into your microphone.")
+        reference_audio = st.file_uploader("Upload Reference Audio", type=["wav", "mp3", "ogg"])
+        ref_audio_path = None
+        
+        if reference_audio:
+            with NamedTemporaryFile(delete=False, suffix=".wav") as temp_ref_audio:
+                temp_ref_audio.write(reference_audio.read())
+                ref_audio_path = temp_ref_audio.name
+            st.audio(ref_audio_path)
 
     with col2:
         st.header("Step 2: Ask Something")
         # User Input (Text or Audio)
-        input_type = st.radio("Choose Input Type", ("Text", "Audio"))
+        input_type = st.radio("Choose Input Type", ("Text", "Upload Audio"))
         user_input = None
-        user_audio_processor = None
 
         if input_type == "Text":
             user_input = st.text_area("Enter your question or prompt here")
         else:
-            st.write("Record your question:")
-            user_audio_processor = AudioProcessor()
-            webrtc_ctx_user = webrtc_streamer(
-                key="user_audio",
-                mode=WebRtcMode.SENDRECV,
-                client_settings=ClientSettings(
-                    rtc_configuration={"iceServers": [{"urls": ["stun:stun.l.google.com:19302"]}]},
-                    media_stream_constraints={"audio": True, "video": False},
-                ),
-                audio_receiver_size=1024,
-                video_processor_factory=None,
-                audio_processor_factory=lambda: user_audio_processor,
-            )
-            
-            if webrtc_ctx_user.state.playing and user_audio_processor is not None:
-                st.info("Recording... Ask your question")
-            
-            if st.button("Process Recording"):
-                if user_audio_processor and user_audio_processor.audio_frames:
-                    with NamedTemporaryFile(delete=False, suffix=".wav") as temp_user_audio:
-                        user_audio_processor.save_audio(temp_user_audio.name)
-                        user_input = transcribe_audio(temp_user_audio.name)
-                        st.write(f"Transcribed: {user_input}")
-                else:
-                    st.error("No audio recorded. Please speak into your microphone.")
+            user_audio = st.file_uploader("Upload your question as audio", type=["wav", "mp3", "ogg"])
+            if user_audio:
+                with NamedTemporaryFile(delete=False, suffix=".wav") as temp_user_audio:
+                    temp_user_audio.write(user_audio.read())
+                    st.audio(temp_user_audio.name)
+                    user_input = transcribe_audio(temp_user_audio.name)
+                    st.write(f"Transcribed: {user_input}")
 
     # Process and generate response
     if st.button("Generate AI Response in My Voice"):
@@ -262,11 +184,11 @@ def main():
             return
             
         if not ref_audio_path:
-            st.error("Please provide reference audio (upload or record).")
+            st.error("Please upload reference audio.")
             return
 
         if not user_input:
-            st.error("Please enter text or record a question.")
+            st.error("Please enter text or upload an audio question.")
             return
 
         with st.spinner("Processing..."):