Update app.py

app.py

@@ -24,22 +24,24 @@ object_bboxes_list = []
 # Function to add or update the prompt in the list
 def submit_prompt(prompt):
     if object_classes_list:
-        object_classes_list[0] = prompt
+        object_classes_list[0] = prompt  # Overwrite the first element if it exists
     else:
-        object_classes_list.insert(0, prompt)
+        object_classes_list.insert(0, prompt)  # Add to the beginning if the list is empty
 
     if not object_bboxes_list:
-        object_bboxes_list.insert(0, "0,0,512,512")
+        object_bboxes_list.insert(0, "0,0,512,512")  # Add the default bounding box if the list is empty
 
     combined_list = [[cls, bbox] for cls, bbox in zip(object_classes_list, object_bboxes_list)]
-    return combined_list, gr.update(interactive=False)
+    return combined_list, gr.update(interactive=False)  # Make the prompt input non-editable
 
 # Function to add a new object with validation
 def add_object(object_class, bbox):
     try:
         x1, y1, x2, y2 = map(int, bbox.split(","))
-        if x2 < x1 or y2 < y1 or x1 < 0 or y1 < 0 or x2 > 512 or y2 > 512:
-            return "Error: Invalid coordinates.", []
+        if x2 < x1 or y2 < y1:
+            return "Error: x2 cannot be less than x1 and y2 cannot be less than y1.", []
+        if x1 < 0 or y1 < 0 or x2 > 512 or y2 > 512:
+            return "Error: Coordinates must be between 0 and 512.", []
         object_classes_list.append(object_class)
         object_bboxes_list.append(bbox)
         combined_list = [[cls, bbox] for cls, bbox in zip(object_classes_list, object_bboxes_list)]
@@ -47,21 +49,52 @@ def add_object(object_class, bbox):
     except ValueError:
         return "Error: Invalid input format. Use x1,y1,x2,y2.", []
 
+# Function to generate images based on added objects
+def generate_image(prompt, guidance_scale, num_inference_steps, randomize_seed, seed):
+    img_width, img_height = 512, 512
+    r_image = np.zeros((img_height, img_width, 3), dtype=np.uint8)
+    list_cond_image = []
+
+    for bbox in object_bboxes_list:
+        x1, y1, x2, y2 = map(int, bbox.split(","))
+        cond_image = np.zeros_like(r_image, dtype=np.uint8)
+        cond_image[y1:y2, x1:x2] = 255
+        list_cond_image.append(Image.fromarray(cond_image).convert('RGB'))
+
+    if randomize_seed or seed is None:
+        seed = np.random.randint(0, MAX_SEED)
+
+    generator = torch.manual_seed(seed)
+
+    image = pipe(
+        prompt=prompt,
+        layo_prompt=object_classes_list,
+        guess_mode=False,
+        guidance_scale=guidance_scale,
+        num_inference_steps=num_inference_steps,
+        image=list_cond_image,
+        fuse_type="avg",
+        width=512,
+        height=512
+    ).images[0]
+
+    return image, seed
+
 # Gradio UI with custom CSS
 css = """
 #custom-prompt {
     width: 400px;  /* Set the width of the prompt input box */
 }
 #custom-button {
-    width: 150px;  /* Set the width of the submit button */
-    height: 38px;  /* Set the height of the submit button */
+    width: 120px;  /* Set the width of the submit button */
+    height: 40px;  /* Set the height of the submit button */
 }
 """
 
 with gr.Blocks(css=css) as demo:
     gr.Markdown("# Text-to-Image Generator with Object Addition")
 
-    # Put prompt and submit button in the same row and apply custom styles
+    # Put prompt and submit button in the same row with custom styles
     with gr.Row():
         prompt = gr.Textbox(label="Prompt", placeholder="Enter your prompt here", elem_id="custom-prompt")
         submit_button = gr.Button("Submit Prompt", elem_id="custom-button")
@@ -72,7 +105,56 @@ with gr.Blocks(css=css) as demo:
         value=[]
     )
 
-    # Add remaining UI components...
+    with gr.Row():
+        object_class_input = gr.Textbox(label="Object Class", placeholder="Enter object class (e.g., Object_1)")
+        bbox_input = gr.Textbox(label="Bounding Box (x1,y1,x2,y2)", placeholder="Enter bounding box coordinates")
+
+    add_button = gr.Button("Add Object")
+
+    # Advanced settings in a collapsible accordion
+    with gr.Accordion("Advanced Settings", open=False):
+        seed = gr.Slider(label="Seed", minimum=0, maximum=MAX_SEED, step=1, value=0)
+        randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
+
+        with gr.Row():
+            guidance_scale = gr.Slider(
+                label="Guidance scale",
+                minimum=0.0,
+                maximum=10.0,
+                step=0.1,
+                value=7.5
+            )
+            num_inference_steps = gr.Slider(
+                label="Number of inference steps",
+                minimum=1,
+                maximum=50,
+                step=1,
+                value=50
+            )
+
+    generate_button = gr.Button("Generate Image")
+    result = gr.Image(label="Generated Image")
+
+    # Submit the prompt and update the display
+    submit_button.click(
+        fn=submit_prompt,
+        inputs=prompt,
+        outputs=[objects_display, prompt]
+    )
+
+    # Add object and update display
+    add_button.click(
+        fn=add_object,
+        inputs=[object_class_input, bbox_input],
+        outputs=[objects_display]
+    )
+
+    # Generate image based on added objects
+    generate_button.click(
+        fn=generate_image,
+        inputs=[prompt, guidance_scale, num_inference_steps, randomize_seed, seed],
+        outputs=[result, seed]
+    )
 
 if __name__ == "__main__":
     demo.launch()