feat: Enable MCP

Hello! This is an automated PR adding MCP compatibility to your AI App 🤖.

This PR introduces two improvements:
1. Adds docstrings to the functions in the app file that are directly connected to the Gradio UI, for the downstream LLM to use.
2. Enables the Model-Compute-Platform by adding `mcp_server=True` to the `.launch()` call.

No other logic has been changed. Please review and merge if it looks good!Learn more about MCP compatibility in Spaces here: https://huggingface.co/changelog/add-compatible-spaces-to-your-mcp-tools

app.py

@@ -19,6 +19,15 @@ css = "#col-container {margin:0 auto; max-width:960px;}"
 
 # Background generation via Replicate
 def _gen_bg(prompt: str):
+    """
+    Generate a background image using Replicate's imagen-4-fast model.
+    
+    Args:
+        prompt: Text description for the background to generate
+        
+    Returns:
+        PIL.Image: Generated background image in RGB format
+    """
     url = replicate.run(
         "google/imagen-4-fast",
         input={"prompt": prompt or "cinematic background", "aspect_ratio": "1:1"},
@@ -28,6 +37,21 @@ def _gen_bg(prompt: str):
 
 # Main processing function
 def process_image_and_text(subject_image, adapter_dict, prompt, _unused1, _unused2, size=ADAPTER_SIZE, rank=10.0):
+    """
+    Process subject and adapter images with text prompt to generate inpainted result.
+    
+    Args:
+        subject_image: PIL.Image of the subject to be placed
+        adapter_dict: Either a PIL.Image or dict with 'image' and 'mask' keys for background/sketch
+        prompt: Text description for the generation
+        _unused1: Unused parameter (placeholder)
+        _unused2: Unused parameter (placeholder)
+        size: Target size for processing (default: ADAPTER_SIZE)
+        rank: Rank parameter for the model (default: 10.0)
+        
+    Returns:
+        tuple: (output_image, raw_image) - both are the same PIL.Image result
+    """
     seed, guidance_scale, steps = 42, 2.5, 28
 
     adapter_image = adapter_dict["image"] if isinstance(adapter_dict, dict) else adapter_dict
@@ -111,7 +135,7 @@ with gr.Blocks(css=css, title="ZenCtrl Inpainting") as demo:
 
     gr.HTML(header_html)
     gr.Markdown(
-        "**Generate context-aware images of your subject with ZenCtrl’s inpainting playground.** Upload a subject + optional mask, write a prompt, and hit **Generate**.   \n"
+        "**Generate context-aware images of your subject with ZenCtrl's inpainting playground.** Upload a subject + optional mask, write a prompt, and hit **Generate**.   \n"
         "Open *Advanced Settings* for an AI-generated background.  \n\n"
         "**Note:** The model was trained mainly on interior scenes and other *rigid* objects. Results on people or highly deformable items may contain visual distortions."
     )
@@ -146,12 +170,22 @@ with gr.Blocks(css=css, title="ZenCtrl Inpainting") as demo:
 
     def _load_and_show(subj_path, bg_path, prompt_text):
         """
+        Load example images and prompt for display in the interface.
+        
         Takes the three values coming from an Examples row
         and returns FOUR objects – one for every output widget:
           1. subject PIL image               -> subj_img
           2. dict for the sketch component   -> ref_img
           3. prompt string                   -> promptbox
           4. pre-rendered result PIL         -> output_img
+          
+        Args:
+            subj_path: Path to subject image file
+            bg_path: Path to background image file
+            prompt_text: Example prompt text
+            
+        Returns:
+            tuple: (subject_image, sketch_dict, prompt, output_image)
         """
         out_path = subj_path.replace(".png", "_out.png")          # your saved result
         return (
@@ -203,4 +237,4 @@ with gr.Blocks(css=css, title="ZenCtrl Inpainting") as demo:
 
 # ---------------- Launch ---------------------------------------
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch(mcp_server=True)