final new ultra fixes

app.py

@@ -672,7 +672,11 @@ with gr.Blocks(fill_height=True, title="Dressify - Advanced Outfit Recommendatio
     # Global dataset size control
     with gr.Row():
         gr.Markdown("#### 🎯 **Global Dataset Size Control**")
-        gr.Markdown("This affects BOTH dataset preparation AND training for ultra-fast testing")
+        gr.Markdown("**Note**: Initial bootstrap downloads full dataset (required). Use 'Apply' button to limit splits for testing.")
+    
+    with gr.Row():
+        gr.Markdown("#### 📊 **Current Behavior**")
+        gr.Markdown("• **Bootstrap**: Downloads full dataset (53K outfits) - this is necessary\n• **Training**: Uses limited samples based on your selection\n• **Apply Button**: Regenerates splits with your selected size limit")
     
     with gr.Row():
         global_dataset_size = gr.Dropdown(
@@ -684,7 +688,7 @@ with gr.Blocks(fill_height=True, title="Dressify - Advanced Outfit Recommendatio
         
         # Apply dataset size button
         apply_size_btn = gr.Button("🔄 Apply Dataset Size & Regenerate Splits", variant="primary")
-        size_status = gr.Textbox(label="Dataset Size Status", value="Dataset size: 2000 samples", interactive=False)
+        size_status = gr.Textbox(label="Dataset Size Status", value="Dataset size: 2000 samples (click Apply to regenerate splits)", interactive=False)
         
         def apply_dataset_size(size: str):
             """Apply global dataset size and regenerate splits."""
@@ -693,20 +697,27 @@ with gr.Blocks(fill_height=True, title="Dressify - Advanced Outfit Recommendatio
                     return f"✅ Using full dataset ({size}) - no size limit applied"
                 
                 # Call the dataset preparation with size limit
-                from scripts.prepare_polyvore import main as prepare_polyvore
+                import subprocess
                 import os
                 
                 # Set environment variable for dataset size
                 os.environ["DATASET_SIZE_LIMIT"] = size
                 
-                # Regenerate splits with size limit
-                result = prepare_polyvore(
-                    data_root="/home/user/app/data/Polyvore",
-                    output_dir="/home/user/app/data/Polyvore/splits",
-                    max_samples=int(size)
-                )
+                # Regenerate splits with size limit using subprocess
+                cmd = [
+                    "python", "scripts/prepare_polyvore.py",
+                    "--root", "/home/user/app/data/Polyvore",
+                    "--out", "/home/user/app/data/Polyvore/splits",
+                    "--max_samples", size
+                ]
                 
-                return f"✅ Successfully regenerated splits with {size} samples limit"
+                result = subprocess.run(cmd, capture_output=True, text=True, check=False)
+                
+                if result.returncode == 0:
+                    return f"✅ Successfully regenerated splits with {size} samples limit"
+                else:
+                    return f"❌ Failed to regenerate splits: {result.stderr}"
+                    
             except Exception as e:
                 return f"❌ Failed to apply dataset size: {str(e)}"
         

scripts/prepare_polyvore.py

@@ -358,6 +358,7 @@ def main() -> None:
     ap = argparse.ArgumentParser()
     ap.add_argument("--root", type=str, required=True, help="Polyvore dataset root")
     ap.add_argument("--out", type=str, default=None, help="Output directory for splits (default: <root>/splits)")
+    ap.add_argument("--max_samples", type=int, default=None, help="Maximum number of samples to use (for testing)")
     ap.add_argument("--max_triplets", type=int, default=200000)
     ap.add_argument("--neg_per_pos", type=int, default=1)
     ap.add_argument("--force_random_split", action="store_true", help="Force random split creation (not recommended)")
@@ -423,6 +424,15 @@ def main() -> None:
             print("🔧 Please ensure official splits are available in nondisjoint/ or disjoint/ folders.")
             return
 
+    # Apply dataset size limit if specified
+    if args.max_samples:
+        print(f"🎯 Limiting dataset to {args.max_samples} samples for testing...")
+        for split in splits:
+            if splits[split]:
+                # Take only the first max_samples outfits
+                splits[split] = splits[split][:args.max_samples]
+                print(f"   📊 {split}: Limited to {len(splits[split])} outfits")
+
     # Generate training data for each split
     for split, outfits in splits.items():
         if not outfits: