Update app.py

app.py

@@ -5,7 +5,6 @@ from PIL import Image
 import numpy as np
 from torchvision import transforms
 
-# Set up device
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
 # Initialize the inpainting model
@@ -18,65 +17,38 @@ except Exception as e:
 # Load a test image
 def load_test_image():
     try:
-        return Image.open("path/to/your/test_image.png")  # Replace with a valid path
+        # Provide the absolute path to a test image
+        return Image.open("/absolute/path/to/your/test_image.png")
     except Exception as e:
         print(f"Error loading test image: {e}")
         return None
 
 def process_image(prompt, image, style, upscale_factor, inpaint):
-    try:
-        # Use a test image if no image is received
+    if image is None:
+        image = load_test_image()
         if image is None:
-            image = load_test_image()
-            if image is None:
-                return None, "No image received and failed to load test image."
+            return None, "No image received and failed to load test image."
 
+    try:
         print(f"Received image type: {type(image)}")
-        
         if isinstance(image, np.ndarray):
-            print(f"Image shape: {image.shape}")
             image = Image.fromarray(image)
         elif isinstance(image, torch.Tensor):
-            print(f"Image tensor shape: {image.shape}")
             image = transforms.ToPILImage()(image)
-        elif isinstance(image, Image.Image):
-            print("Image is already in PIL format.")
-        else:
+        elif not isinstance(image, Image.Image):
             return None, f"Unsupported image format: {type(image)}."
-
-        # Check if the image is valid
-        if not isinstance(image, Image.Image):
-            return None, "Error: Image format conversion failed."
-
-        # Log the input parameters
-        print(f"Prompt: {prompt}")
-        print(f"Style: {style}")
-        print(f"Upscale Factor: {upscale_factor}")
-        print(f"Inpaint: {inpaint}")
-
-        # Example placeholder logic for using the pipeline
-        if inpaint and inpaint_model:
-            result = inpaint_model(prompt=prompt, image=image, guidance_scale=7.5)
-        else:
-            result = inpaint_model(prompt=prompt, guidance_scale=7.5) if inpaint_model else None
-
-        # Check if the result is valid
-        if result and hasattr(result, 'images') and len(result.images) > 0:
-            return result.images[0], None  # Return image and no error
-        else:
-            return None, "Error: No image returned from model."  # Return no image and an error message
-
+        
+        return image, None
     except Exception as e:
         error_message = f"Error in process_image function: {e}"
         print(error_message)
-        return None, error_message  # Return no image and the error message
+        return None, error_message
 
-# Define the Gradio interface
 with gr.Blocks() as demo:
     with gr.Row():
         with gr.Column():
             prompt_input = gr.Textbox(label="Enter your prompt")
-            image_input = gr.Image(label="Image (for inpainting)", type="pil")  # Ensure type is PIL
+            image_input = gr.Image(label="Image (for inpainting)", type="pil")
             style_input = gr.Dropdown(choices=["Fooocus Style", "SAI Anime"], label="Select Style")
             upscale_input = gr.Slider(minimum=1, maximum=4, step=1, label="Upscale Factor")
             inpaint_input = gr.Checkbox(label="Enable Inpainting")
@@ -88,8 +60,7 @@ with gr.Blocks() as demo:
             generate_button.click(
                 process_image,
                 inputs=[prompt_input, image_input, style_input, upscale_input, inpaint_input],
-                outputs=[output_image, error_output]  # Handle both image and error output
+                outputs=[output_image, error_output]
             )
 
-# Launch the interface
 demo.launch()