Update database.py

database.py

@@ -156,6 +156,104 @@ class DogDatabase:
             WHERE dog_id = ?
         """, (dog_id,))
         self.conn.commit()
+    def export_training_dataset(self, output_dir: str = "training_dataset") -> Dict:
+        """
+        Export database images to folder structure for fine-tuning
+        
+        Output structure:
+            training_dataset/
+            ├── dog_1/
+            │   ├── img_0000.jpg
+            │   └── ...
+            ├── dog_2/
+            └── metadata.json
+        
+        Returns:
+            dict with export statistics
+        """
+        from pathlib import Path
+        import json
+        from datetime import datetime
+        
+        output_path = Path(output_dir)
+        output_path.mkdir(parents=True, exist_ok=True)
+        
+        # Get all active dogs
+        dogs_df = self.get_all_dogs(active_only=True)
+        
+        if dogs_df.empty:
+            return {
+                'success': False,
+                'message': 'No dogs in database',
+                'total_dogs': 0,
+                'total_images': 0
+            }
+        
+        total_images = 0
+        export_info = []
+        
+        for _, dog in dogs_df.iterrows():
+            dog_id = dog['dog_id']
+            dog_name = dog['name'] or f"dog_{dog_id}"
+            
+            # Create folder for this dog
+            dog_folder = output_path / dog_name
+            dog_folder.mkdir(exist_ok=True)
+            
+            # Get all non-discarded images
+            images = self.get_dog_images(
+                dog_id=dog_id,
+                validated_only=False,
+                include_discarded=False
+            )
+            
+            if not images:
+                print(f"Warning: Dog {dog_id} has no images")
+                continue
+            
+            # Save each image as JPG
+            for idx, img_data in enumerate(images):
+                image = img_data['image']  # OpenCV array from BLOB
+                filename = f"img_{idx:04d}.jpg"
+                filepath = dog_folder / filename
+                
+                # Write to disk
+                cv2.imwrite(str(filepath), image)
+            
+            total_images += len(images)
+            
+            export_info.append({
+                'dog_id': dog_id,
+                'name': dog_name,
+                'image_count': len(images),
+                'folder': str(dog_folder)
+            })
+            
+            print(f"Exported {len(images)} images for {dog_name}")
+        
+        # Create metadata file
+        metadata = {
+            'export_date': datetime.now().isoformat(),
+            'total_dogs': len(dogs_df),
+            'total_images': total_images,
+            'output_directory': str(output_path),
+            'dogs': export_info
+        }
+        
+        metadata_path = output_path / 'metadata.json'
+        with open(metadata_path, 'w') as f:
+            json.dump(metadata, f, indent=2)
+        
+        print(f"\nExport complete: {len(dogs_df)} dogs, {total_images} images")
+        print(f"Location: {output_path}")
+        
+        return {
+            'success': True,
+            'total_dogs': len(dogs_df),
+            'total_images': total_images,
+            'output_path': str(output_path),
+            'dogs': export_info
+        }
     
     def merge_dogs(self, keep_id: int, merge_id: int) -> bool:
         """Merge two dogs, keeping keep_id"""