"""
Text extraction utility
Extract only text excluding table regions to prevent duplication
"""
from typing import List, Dict
def extract_non_table_text_regions(page, table_bboxes: List[tuple]) -> List[Dict]:
"""
Extract text regions excluding table areas
Split page by table bboxes and extract text from each region
Args:
page: pdfplumber page object
table_bboxes: List of table bboxes [(x0, top, x1, bottom), ...]
Returns:
[{'top': float, 'text': str}, ...] Text by region
"""
if not table_bboxes:
# If no tables, extract full text
text = page.extract_text()
return [{'top': 0, 'text': text}] if text else []
# Sort bboxes by top coordinate
sorted_tables = sorted(table_bboxes, key=lambda b: b[1])
text_regions = []
page_height = page.height
page_width = page.width
# Region before first table
first_table = sorted_tables[0]
if first_table[1] > 0: # If top is greater than 0
above_bbox = (0, 0, page_width, first_table[1])
above_region = page.within_bbox(above_bbox)
text = above_region.extract_text()
if text and text.strip():
text_regions.append({'top': 0, 'text': text})
# Regions between tables
for i in range(len(sorted_tables) - 1):
current_table = sorted_tables[i]
next_table = sorted_tables[i + 1]
# Between current table end and next table start
between_bbox = (0, current_table[3], page_width, next_table[1])
if next_table[1] > current_table[3]: # If there's space
between_region = page.within_bbox(between_bbox)
text = between_region.extract_text()
if text and text.strip():
text_regions.append({'top': current_table[3], 'text': text})
# Region after last table
last_table = sorted_tables[-1]
if last_table[3] < page_height:
below_bbox = (0, last_table[3], page_width, page_height)
below_region = page.within_bbox(below_bbox)
text = below_region.extract_text()
if text and text.strip():
text_regions.append({'top': last_table[3], 'text': text})
return text_regions
