Update app.py

app.py

@@ -10,6 +10,7 @@ import os
 import random
 import tempfile
 import matplotlib.pyplot as plt
+import time  # To generate unique filenames
 
 # Load your Pix2Pix model (make sure the path is correct)
 model = load_model('./model_022600.h5', compile=False)
@@ -66,7 +67,7 @@ def save_spectrogram_image(spectrogram, name):
     plt.imshow(spectrogram, aspect='auto', origin='lower', cmap='gray')
     plt.axis('off')
     
-    # Create a spectrogram filename based on the uploaded image's name
+    # Save the spectrogram image using the unique name
     temp_image_path = f"{name}_spectrogram.png"
     plt.savefig(temp_image_path, bbox_inches='tight', pad_inches=0)
     
@@ -75,8 +76,8 @@ def save_spectrogram_image(spectrogram, name):
 
 # Process the input image and convert to audio
 def process_image(input_image):
-    # Extract the base name from the uploaded image filename (without extension)
-    image_name = os.path.splitext(os.path.basename(input_image.name))[0]
+    # Generate a unique name based on the current time
+    image_name = f"image_{int(time.time())}"
 
     def load_image(image, size=(256, 256)):
         image = image.resize(size)
@@ -102,13 +103,13 @@ def process_image(input_image):
     # Modify the spectrogram randomly
     img = modify_spectrogram(img)
     
-    # Save the modified spectrogram as an image, using the same name as the uploaded file
+    # Save the modified spectrogram as an image, using the unique name
     spectrogram_image_path = save_spectrogram_image(img, image_name)
     
     # Convert the spectrogram back to audio using librosa
     wav = librosa.feature.inverse.mel_to_audio(img, sr=44100, n_fft=2048, hop_length=512)
     
-    # Save the audio file to a location, using the same name as the uploaded file
+    # Save the audio file, using the unique name
     audio_file_path = f"{image_name}_generated_audio.wav"
     sf.write(audio_file_path, wav, samplerate=44100)
     
@@ -132,7 +133,7 @@ interface = gr.Interface(
     inputs=gr.Image(type="pil"),  # Input is an image
     outputs=[gr.Image(type="filepath"), gr.Audio(type="filepath")],  # Output both spectrogram image and audio file
     title="Image to Audio Generator with Spectrogram Display",
-    description="Upload an image, and get an audio file generated using Pix2Pix.",
+    description="Upload an image and get an audio file generated using Pix2Pix.",
 )
 
 # Launch the interface