fix

app.py

@@ -11,7 +11,6 @@ import base64
 import logging
 import sys
 import shutil
-import zipfile
 
 # Configure logging
 # Simplified logging for cloud deployment
@@ -282,18 +281,6 @@ def process_business_cards(images, model_name="gemini-2.5-flash", save_images=Tr
         
         logger.info(f"Successfully saved {len(saved_image_paths)} images")
         
-        # Create a ZIP file of saved images
-        if len(saved_image_paths) > 0:
-            zip_filename = f"business_cards_{timestamp}.zip"
-            zip_path = Path(zip_filename)
-            try:
-                with zipfile.ZipFile(zip_path, 'w') as zipf:
-                    for img_path in saved_image_paths:
-                        zipf.write(img_path, os.path.basename(img_path))
-                logger.info(f"Created ZIP file with {len(saved_image_paths)} images: {zip_filename}")
-            except Exception as e:
-                logger.error(f"Failed to create ZIP file: {e}")
-        
         # List directory contents for debugging
         try:
             images_list = list(images_dir.iterdir())
@@ -463,13 +450,25 @@ def process_business_cards(images, model_name="gemini-2.5-flash", save_images=Tr
         
         logger.info(f"Current run Excel file saved successfully: {current_filename}")
         
-        # Also save a copy in root directory for Hugging Face Spaces visibility
-        root_current_file = Path(f"current_run_{timestamp}.xlsx")
+        # Create a manifest file for the current run directory
+        manifest_filename = output_dir / f"current_run_{timestamp}_manifest.txt"
         try:
-            shutil.copy2(current_filename, root_current_file)
-            logger.info(f"Copied current run file to root: {root_current_file}")
+            with open(manifest_filename, "w") as f:
+                f.write(f"Current Run Directory: {output_dir}\n")
+                f.write(f"Images Directory: {images_dir}\n")
+                f.write(f"Current Run File: {current_filename}\n")
+                f.write(f"Cumulative File: {cumulative_filename}\n")
+                f.write(f"Total Cards in Database: {len(cumulative_df)}\n")
+                f.write(f"Total Images Saved: {len(saved_image_paths) if saved_image_paths else 0}\n")
+                f.write(f"Total API Calls Made: {len(image_batches)}\n")
+                f.write(f"Model Used: {model_name}\n")
+                f.write(f"Save Images: {save_images}\n")
+                f.write(f"Processing Date: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n")
+                f.write(f"Session ID: {timestamp}\n")
+                f.write(f"Session Summary: {summary}\n") # Use the summary variable from the previous run
+            logger.info(f"Manifest file created: {manifest_filename}")
         except Exception as e:
-            logger.error(f"Failed to copy current file to root: {e}")
+            logger.error(f"Failed to create manifest file: {e}")
     except Exception as e:
         logger.error(f"Failed to write current run Excel file: {e}")
         raise
@@ -497,13 +496,25 @@ def process_business_cards(images, model_name="gemini-2.5-flash", save_images=Tr
         
         logger.info(f"Cumulative Excel file saved successfully: {cumulative_filename}")
         
-        # Also save a copy in root directory for Hugging Face Spaces visibility
-        root_cumulative_file = Path("all_business_cards_total.xlsx")
+        # Create a manifest file for the cumulative directory
+        manifest_filename = output_dir / f"all_business_cards_total_{timestamp}_manifest.txt"
         try:
-            shutil.copy2(cumulative_filename, root_cumulative_file)
-            logger.info(f"Copied cumulative file to root: {root_cumulative_file}")
+            with open(manifest_filename, "w") as f:
+                f.write(f"All Business Cards Directory: {output_dir}\n")
+                f.write(f"Images Directory: {images_dir}\n")
+                f.write(f"Current Run File: {current_filename}\n")
+                f.write(f"Cumulative File: {cumulative_filename}\n")
+                f.write(f"Total Cards in Database: {len(cumulative_df)}\n")
+                f.write(f"Total Images Saved: {len(saved_image_paths) if saved_image_paths else 0}\n")
+                f.write(f"Total API Calls Made: {len(image_batches)}\n")
+                f.write(f"Model Used: {model_name}\n")
+                f.write(f"Save Images: {save_images}\n")
+                f.write(f"Processing Date: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n")
+                f.write(f"Session ID: {timestamp}\n")
+                f.write(f"Session Summary: {summary}\n") # Use the summary variable from the previous run
+            logger.info(f"Manifest file created: {manifest_filename}")
         except Exception as e:
-            logger.error(f"Failed to copy cumulative file to root: {e}")
+            logger.error(f"Failed to create manifest file: {e}")
     except Exception as e:
         logger.error(f"Failed to write cumulative Excel file: {e}")
         raise
@@ -525,13 +536,11 @@ def process_business_cards(images, model_name="gemini-2.5-flash", save_images=Tr
     summary += f"📁 Total cumulative file: {cumulative_filename.name}\n"
     summary += f"📊 Total cards in database: {len(cumulative_df)}\n\n"
     
-    # Add note about root directory files for HF Spaces
-    summary += "📌 Note: Files are also saved in the root directory for easy access in Hugging Face Spaces Files tab.\n"
-    summary += "   - current_run_[timestamp].xlsx\n"
-    summary += "   - all_business_cards_total.xlsx\n"
-    if save_images and 'saved_image_paths' in locals() and len(saved_image_paths) > 0:
-        summary += f"   - business_cards_{timestamp}.zip (contains all images)\n"
-    summary += "\n"
+    # Add note about file locations
+    summary += "📌 File Locations:\n"
+    summary += f"   - Excel files: business_card_exports/ folder\n"
+    summary += f"   - Images: business_cards/ folder\n"
+    summary += f"   - Check FOLDER_CONTENTS.txt for complete file listing\n\n"
     
     if errors:
         logger.warning(f"Encountered {len(errors)} errors during processing")
@@ -578,6 +587,47 @@ def process_business_cards(images, model_name="gemini-2.5-flash", save_images=Tr
     except Exception as e:
         logger.error(f"Error listing root directory: {e}")
     
+    # Create a directory listing file in root showing folder contents
+    try:
+        with open("FOLDER_CONTENTS.txt", "w") as f:
+            f.write("=== BUSINESS CARD ANALYZER - FOLDER CONTENTS ===\n")
+            f.write(f"Generated: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n\n")
+            
+            # List business_card_exports folder
+            f.write("📁 business_card_exports/\n")
+            f.write("-" * 50 + "\n")
+            if output_dir.exists():
+                files = sorted(output_dir.iterdir())
+                for file in files:
+                    if file.is_file():
+                        size = file.stat().st_size
+                        f.write(f"  📄 {file.name} ({size:,} bytes)\n")
+            else:
+                f.write("  [Folder does not exist]\n")
+            f.write(f"\nTotal files: {len(list(output_dir.glob('*')))}\n\n")
+            
+            # List business_cards folder
+            f.write("📁 business_cards/\n")
+            f.write("-" * 50 + "\n")
+            if images_dir.exists():
+                files = sorted(images_dir.iterdir())
+                for file in files:
+                    if file.is_file():
+                        size = file.stat().st_size
+                        f.write(f"  🖼️ {file.name} ({size:,} bytes)\n")
+            else:
+                f.write("  [Folder does not exist]\n")
+            f.write(f"\nTotal files: {len(list(images_dir.glob('*')))}\n\n")
+            
+            # Add latest processing summary
+            f.write("📊 Latest Processing Summary\n")
+            f.write("-" * 50 + "\n")
+            f.write(summary)
+            
+        logger.info("Created FOLDER_CONTENTS.txt in root directory")
+    except Exception as e:
+        logger.error(f"Failed to create folder contents file: {e}")
+    
     # Return string paths for Gradio File components
     return str(current_filename), str(cumulative_filename), summary, preview_df
 
@@ -600,9 +650,10 @@ with gr.Blocks(title="Business Card Data Extractor") as demo:
         - 📝 **Tracking**: Image file paths included in Excel database
         
         **📌 Finding Your Files in Hugging Face Spaces:**
-        - Click the "Files" tab at the top of this page
-        - Look for: `current_run_*.xlsx`, `all_business_cards_total.xlsx`, and `business_cards_*.zip`
-        - These files are copied to the root directory for easy access
+        - All files are saved in the `business_card_exports` and `business_cards` folders
+        - Check `FOLDER_CONTENTS.txt` in the Files tab to see a complete list of all files
+        - Excel files are in `business_card_exports/`
+        - Images are in `business_cards/` (if save option enabled)
         """
     )