helpers

Extraction Helpers

Shared utility functions for extractor geometry and text operations.

Files

• geometry.py
  • bbox center/bounds utilities
  • region selection helpers (top_region_spans, bottom_region_spans)
• span_queries.py
  • anchor search and proximity queries:
    • find spans containing keyword
    • spans to right / below anchors
    • nearest joined text snippets
• text_parsing.py
  • normalization and parsing helpers:
    • normalize_text
    • extract_money_from_text
    • extract_date_from_text
    • extract_name_like_text

Why This Exists

Without shared helper modules, extractor logic duplicates geometry and regex code across document types. These helpers keep extractor classes focused on document-specific heuristics.