Update app.py

app.py

@@ -3,31 +3,28 @@ import torchaudio
 from audioseal import AudioSeal
 import torch
 from io import BytesIO
+import numpy as np
+
+# Function to handle audio data as NumPy arrays
+def detect_watermark(audio_data, sample_rate):
+    # Convert NumPy array to tensor
+    waveform = torch.tensor(audio_data, dtype=torch.float32)
+    waveform = waveform.unsqueeze(0)  # Add batch dimension if necessary
 
-# Adjusted function to handle Gradio's file input correctly
-def detect_watermark(file_info):
-    # Gradio passes the uploaded file as a tuple of (filename, fileobject)
-    _, file_object = file_info
-    
-    # Use BytesIO object for compatibility with torchaudio.load
-    audio, sr = torchaudio.load(file_object)
-    audio = audio.unsqueeze(0)  # Add batch dimension
-    
     # Initialize and use the AudioSeal detector
     detector = AudioSeal.load_detector("audioseal_detector_16bits")
-    result, message = detector.detect_watermark(audio, message_threshold=0.5)
+    result, message = detector.detect_watermark(waveform, message_threshold=0.5)
     
     # Interpret and return the detection result
     detection_result = "AI-generated" if result else "genuine"
     return f"This audio is likely {detection_result} based on watermark detection."
 
-# Updated Gradio interface without the 'source' keyword argument
+# Define Gradio interface with adjusted input to handle NumPy arrays
 interface = gr.Interface(fn=detect_watermark,
-                         inputs=gr.Audio(type="file", label="Upload your audio"),
+                         inputs=[gr.Audio(label="Upload your audio", type="numpy"), gr.Number(label="Sample Rate", default=44100)],
                          outputs="text",
                          title="Deep Fake Defender: AI Voice Cloning Detection",
                          description="Upload an audio file to check if it's AI-generated or genuine.")
 
 if __name__ == "__main__":
     interface.launch()
-