Complete Refactor

Backend/api/app.py

@@ -0,0 +1,22 @@
+from flask import Flask
+from flask_socketio import SocketIO
+from src.utils.config import Config
+from src.utils.logger import setup_logger
+from api.routes import setup_routes
+from api.socket_handlers import setup_socket_handlers
+
+def create_app():
+    app = Flask(__name__)
+    app.config.from_object(Config)
+
+    setup_logger(app)
+    setup_routes(app)
+    setup_socket_handlers(app)
+
+    return app
+
+app = create_app()
+socketio = SocketIO(app)
+
+if __name__ == "__main__":
+    socketio.run(app, host='0.0.0.0', port=5000)

Backend/api/routes.py

@@ -0,0 +1,29 @@
+from flask import Blueprint, request, jsonify
+from src.services.transcription_service import TranscriptionService
+from src.services.tts_service import TextToSpeechService
+
+api = Blueprint('api', __name__)
+
+transcription_service = TranscriptionService()
+tts_service = TextToSpeechService()
+
+@api.route('/transcribe', methods=['POST'])
+def transcribe_audio():
+    audio_data = request.files.get('audio')
+    if not audio_data:
+        return jsonify({'error': 'No audio file provided'}), 400
+    
+    text = transcription_service.transcribe(audio_data)
+    return jsonify({'transcription': text})
+
+@api.route('/generate-response', methods=['POST'])
+def generate_response():
+    data = request.json
+    user_input = data.get('input')
+    if not user_input:
+        return jsonify({'error': 'No input provided'}), 400
+    
+    response_text = tts_service.generate_response(user_input)
+    audio_data = tts_service.text_to_speech(response_text)
+    
+    return jsonify({'response': response_text, 'audio': audio_data})

Backend/api/socket_handlers.py

@@ -0,0 +1,32 @@
+from flask import request
+from flask_socketio import SocketIO, emit
+from src.audio.processor import process_audio
+from src.services.transcription_service import TranscriptionService
+from src.services.tts_service import TextToSpeechService
+from src.llm.generator import load_csm_1b
+
+socketio = SocketIO()
+
+transcription_service = TranscriptionService()
+tts_service = TextToSpeechService()
+generator = load_csm_1b()
+
+@socketio.on('audio_stream')
+def handle_audio_stream(data):
+    audio_data = data['audio']
+    speaker_id = data['speaker']
+    
+    # Process the incoming audio
+    processed_audio = process_audio(audio_data)
+    
+    # Transcribe the audio to text
+    transcription = transcription_service.transcribe(processed_audio)
+    
+    # Generate a response using the LLM
+    response_text = generator.generate(text=transcription, speaker=speaker_id)
+    
+    # Convert the response text back to audio
+    response_audio = tts_service.convert_text_to_speech(response_text)
+    
+    # Emit the response audio back to the client
+    emit('audio_response', {'audio': response_audio, 'speaker': speaker_id})