Update pdftotext.py

pdftotext.py

@@ -48,24 +48,25 @@ def texts_from_pdf(pdfshareablelink):
         
         # Iterate over blocks, lines, and spans to extract lines of text
         for block in text_dict['blocks']:
-            for line in block['lines']:
-                for span in line['spans']:
-                    span_text = span['text'].strip()
-                    span_y = span['bbox'][1]  # Y-coordinate of the span (bbox[1] is the top y-coordinate)
-                    
-                    # Check if the current span belongs to the same line (based on y-coordinate)
-                    if current_y is None:
-                        current_y = span_y  # Initialize the first y-coordinate
+            if 'lines' in block:  # Check if 'lines' key exists
+                for line in block['lines']:
+                    for span in line['spans']:
+                        span_text = span['text'].strip()
+                        span_y = span['bbox'][1]  # Y-coordinate of the span (bbox[1] is the top y-coordinate)
                         
-                    if abs(current_y - span_y) < 2:  # Threshold to determine if it's the same line
-                        # If the y-coordinate is close enough, add to the current line
-                        current_line += " " + span_text
-                    else:
-                        # If it's a new line, append the current line and reset
-                        all_text += current_line.strip() + '\n'  # Add line to all_text with a newline
-                        current_line = span_text  # Start the new line with the current span
-                        current_y = span_y  # Update the y-coordinate for the new line
-    
+                        # Check if the current span belongs to the same line (based on y-coordinate)
+                        if current_y is None:
+                            current_y = span_y  # Initialize the first y-coordinate
+                            
+                        if abs(current_y - span_y) < 2:  # Threshold to determine if it's the same line
+                            # If the y-coordinate is close enough, add to the current line
+                            current_line += " " + span_text
+                        else:
+                            # If it's a new line, append the current line and reset
+                            all_text += current_line.strip() + '\n'  # Add line to all_text with a newline
+                            current_line = span_text  # Start the new line with the current span
+                            current_y = span_y  # Update the y-coordinate for the new line
+        
         # Append the last line of the page (if there's any)
         if current_line:
             all_text += current_line.strip() + '\n'