Update pdf_processing.py

pdf_processing.py

@@ -19,6 +19,8 @@ def process_pdf(pdf_path, output_folder):
     if not os.path.exists(output_folder):
         os.makedirs(output_folder)
 
+    output_files = []  # Dùng để lưu danh sách file kết quả
+
     for page_num, image in enumerate(images):
         print(f"Processing page {page_num + 1}/{len(images)}")
 
@@ -41,7 +43,10 @@ def process_pdf(pdf_path, output_folder):
 
         # Perform OCR on the remaining area of the image
         remaining_text = apply_ocr_remaining_area(image_without_tables)
-        save_remaining_text_to_txt(remaining_text, output_folder, page_num)
+        
+        # Save the remaining text to a file and thêm vào danh sách output_files
+        txt_file = save_remaining_text_to_txt(remaining_text, output_folder, page_num)
+        output_files.append(txt_file)
 
         # Process each cropped table and save data as CSV
         table_data_list = []
@@ -51,6 +56,13 @@ def process_pdf(pdf_path, output_folder):
             table_data = process_cropped_table(cropped_table, structure_model, device, page_num, idx, output_folder)
             table_data_list.append(table_data)
 
+            # Save each table data to CSV and thêm vào danh sách output_files
+            csv_filename = os.path.join(output_folder, f'page_{page_num + 1}_table_{idx + 1}.csv')
+            output_files.append(csv_filename)
+    
+    # Trả về danh sách các file output cho Gradio
+    return output_files
+
 # Function to process each cropped table and save to CSV
 def process_cropped_table(cropped_table, structure_model, device, page_num, table_index, output_folder):
     structure_transform = transforms.Compose([
@@ -104,4 +116,4 @@ def get_cell_coordinates_by_row(table_data):
             row_cells.append({'bbox': cell_bbox})
         cell_coordinates.append(row_cells)
 
-    return cell_coordinates
+    return cell_coordinates