Function walk_table

scripts/_debug_eurlex_table.py:28–59

(tbl, depth=0)

# Better approach: only process direct rows, handle nested tables recursively
print("\n=== PROPOSED: recursive table walker ===")
def walk_table(tbl, depth=0):
    prefix = "  " * depth
    # Get direct rows only
    tbody = tbl.find("tbody", recursive=False)
    container = tbody if tbody else tbl
    rows = [c for c in container.children if isinstance(c, Tag) and c.name == "tr"]
    
    for row in rows:
        cells = row.find_all(["td", "th"], recursive=False)
        if not cells:
            continue
        
        # For each cell, separate direct text from nested tables
        for cell in cells:
            inner_tables = cell.find_all("table", recursive=False)
            
            # Get text excluding nested tables
            direct_text = ""
            for child in cell.children:
                if isinstance(child, Tag) and child.name == "table":
                    continue
                if isinstance(child, Tag):
                    direct_text += child.get_text(" ", strip=False)
                else:
                    direct_text += str(child)
            direct_text = re.sub(r"\s+", " ", direct_text).strip()
            
            if direct_text:
                print(f"{prefix}TEXT: '{direct_text[:100]}'")
            
            for it in inner_tables:
                walk_table(it, depth + 1)

walk_table(table)