Update app.py

app.py

@@ -1,48 +1,20 @@
-import torch
-import torchaudio
-import numpy as np
-import sounddevice as sd
-import queue
-import time
-import requests  # To trigger Flask endpoint
+import os
+from flask import Flask, render_template
 
-# Flask Server URL
-FLASK_URL = "http://127.0.0.1:5000/trigger_wake_word"
+app = Flask(__name__)
 
-# Load a pre-trained speech recognition model (Torch Hub)
-model = torchaudio.pipelines.WAV2VEC2_ASR_BASE_960H.get_model()
-waveform_queue = queue.Queue()
+# Get Hugging Face assigned port, default to 7860 if running locally
+PORT = int(os.getenv("PORT", 7860))
 
-# Capture audio stream
-def callback(indata, frames, time, status):
-    if status:
-        print(status)
-    waveform_queue.put(indata.copy())
+@app.route('/')
+def index():
+    """Serve the main page with voice recognition."""
+    return render_template('index.html')
 
-# Function to process audio and detect wake word
-def detect_wake_word():
-    while True:
-        waveform = waveform_queue.get()
-        waveform = torch.tensor(waveform).mean(dim=1)  # Convert stereo to mono
-        waveform = waveform.unsqueeze(0)  # Add batch dimension
+@app.route('/landing')
+def landing():
+    """Landing page that loads after wake-word detection."""
+    return render_template('landing.html')
 
-        # Convert waveform to text
-        with torch.inference_mode():
-            transcript = model(waveform)[0]
-
-        transcript = transcript.lower()
-        print(f"Detected: {transcript}")
-
-        if "hi biriyani hub" in transcript:
-            print("Wake word detected!")
-            requests.get(FLASK_URL)  # Trigger Flask route
-            break
-
-# Start capturing audio in a background thread
-def start_listening():
-    with sd.InputStream(callback=callback, channels=1, samplerate=16000):
-        detect_wake_word()
-
-# Run Wake Word Detection
-if __name__ == "__main__":
-    start_listening()
+if __name__ == '__main__':
+    app.run(host="0.0.0.0", port=PORT, debug=True)