Update InitialMarkups.py

InitialMarkups.py

@@ -528,7 +528,7 @@ def print_tree_with_numbers(headers, indent=0):
         print_tree_with_numbers(header["children"], indent + 1)
 
 
-def highlight_boxes(doc, highlights, fixed_width=500):  # Set your desired width here
+def highlight_boxes(doc, highlights, stringtowrite, fixed_width=500):  # Set your desired width here
     for page_num, bbox in highlights.items():
         page = doc.load_page(page_num)
         page_width = page.rect.width
@@ -551,7 +551,7 @@ def highlight_boxes(doc, highlights, fixed_width=500):  # Set your desired width
             annot.update()
 
             # Add right-aligned freetext annotation inside the fixed-width box
-            text = "[To be billed]"
+            text = '['+stringtowrite +']'
             annot1 = page.add_freetext_annot(
                 new_rect,
                 text,
@@ -739,7 +739,10 @@ def extract_section_under_header(pdf_path):
                     if combined_line_norm and combined_line_norm in paths[-2]:
                         print(combined_line_norm)
                         headertoContinue2 = combined_line_norm
-
+                    if  'installation' in paths[-2].lower() or 'execution' in paths[-2].lower() or 'miscellaneous items' in paths[-2].lower() :
+                      stringtowrite='Not to be billed'
+                    else:
+                      stringtowrite='To be billed'
                     # Optimized header matching
                     existsfull = (
                         ( combined_line_norm in allchildrenheaders_set or
@@ -836,7 +839,7 @@ def extract_section_under_header(pdf_path):
                                         "Author": "ADR",
                                         "Creation Date": formatted_time,
                                         "Layer": "Initial",
-                                        "Code": "to be added",
+                                        "Code": stringtowrite,
                                         "head above 1":  paths[-2],
                                         "head above 2":  paths[0]
                                     }
@@ -932,7 +935,7 @@ def extract_section_under_header(pdf_path):
                                             "Author": "ADR",
                                             "Creation Date": formatted_time,
                                             "Layer": "Initial",
-                                            "Code": "to be added",
+                                            "Code": stringtowrite,
                                             "head above 1":  paths[-2],
                                             "head above 2":  paths[0]
                                         }
@@ -973,7 +976,7 @@ def extract_section_under_header(pdf_path):
                                   for page_num, bbox in current_bbox.items():
                                       bbox[3] = last_y1s.get(page_num, bbox[3])
                                       page_highlights[page_num] = bbox
-                                  highlight_boxes(docHighlights, page_highlights)
+                                  highlight_boxes(docHighlights, page_highlights,stringtowrite)
 
                                   break_collecting = True
                                   break
@@ -1009,7 +1012,11 @@ def extract_section_under_header(pdf_path):
             for page_num, bbox in current_bbox.items():
                 bbox[3] = last_y1s.get(page_num, bbox[3])
                 page_highlights[page_num] = bbox
-            highlight_boxes(docHighlights, page_highlights)
+            if  'installation' in paths[-2].lower() or 'execution' in paths[-2].lower() or 'miscellaneous items' in paths[-2].lower() :
+                stringtowrite='Not to be billed'
+            else:
+                stringtowrite='To be billed'
+            highlight_boxes(docHighlights, page_highlights,stringtowrite)
 
     # docHighlights.save("highlighted_output.pdf", garbage=4, deflate=True)