import os, sys, re

sys.stdout.reconfigure(encoding='utf-8')

ch9_path = "resources/knowledge/original/원가계산/건설공사_표준품셈/02_토목부문/제9장_측량.md"
with open(ch9_path, 'r', encoding='utf-8') as f:
    lines = f.readlines()

new_lines = []
fixed_cnt = 0

tech_words = [
    "특급기술자", "고급기술자", "중급기술자", "초급기술자", "초급기능사(측량)", "초급기능사", "인부", "측량사", "측량사보"
]

for idx, line in enumerate(lines):
    line_str = line.strip()
    if not (line_str.startswith('|') and line_str.endswith('|')):
        new_lines.append(line)
        continue
        
    cells = [c.strip() for c in line_str.split('|')[1:-1]]
    if len(cells) < 3 or cells[0] in ["계", "합계"]:
        new_lines.append(line)
        continue
        
    # Check if first cell contains multiple technician grades
    c0 = cells[0]
    c0_clean = c0.replace(" ", "")
    matched_grades = []
    temp_s = c0_clean
    for tw in sorted(tech_words, key=len, reverse=True):
        while tw in temp_s:
            matched_grades.append(tw)
            temp_s = temp_s.replace(tw, "", 1)
            
    if len(matched_grades) >= 2 and temp_s == "":
        # Reconstruct in order
        ordered_grades = []
        pos = 0
        while pos < len(c0_clean):
            for tw in sorted(tech_words, key=len, reverse=True):
                if c0_clean[pos:].startswith(tw):
                    ordered_grades.append(tw)
                    pos += len(tw)
                    break
                    
        N = len(ordered_grades)
        # Check column 1 (수량/산식)
        # Often contains N formulas ending with = number
        # e.g. "3×10/16×1.2×1.12= 2.52 24.5×10/16..."
        c1 = cells[1]
        # Match formulas like "...= <num>"
        f_matches = list(re.finditer(r'(.*?=\s*\d+(?:\.\d+)?)\s*(?=[^=]*=\s*\d+|$)', c1))
        c1_parts = [m.group(1).strip() for m in f_matches if m.group(1).strip()]
        
        # Check column 2 (단가)
        c2 = cells[2] if len(cells) > 2 else ""
        c2_toks = c2.split()
        
        # Check column 3 (금액)
        c3 = cells[3] if len(cells) > 3 else ""
        w_matches = list(re.finditer(r'(W\s*=\s*[^W]+?)(?=\s*W\s*=|(?:\s*1\s+)?$)', c3))
        c3_parts = [m.group(1).strip() for m in w_matches if m.group(1).strip()]
        
        if len(c1_parts) == N:
            # We can split!
            for r_idx in range(N):
                g_name = ordered_grades[r_idx]
                qty_val = c1_parts[r_idx]
                rate_val = c2_toks[r_idx] if len(c2_toks) == N else (c2_toks[0] if c2_toks else "")
                amt_val = c3_parts[r_idx] if len(c3_parts) == N else ""
                
                row_cells = [g_name, qty_val, rate_val, amt_val]
                # If there were more columns, append them
                for extra_c in range(4, len(cells)):
                    row_cells.append("" if r_idx > 0 else cells[extra_c])
                new_line_str = "| " + " | ".join(row_cells) + " |\n"
                new_lines.append(new_line_str)
            fixed_cnt += 1
            print(f"L{idx+1}: Fixed survey calc table with {N} grades: {c0[:40]}")
            continue

    new_lines.append(line)

print(f"\nTotal survey calc tables fixed: {fixed_cnt}")

if fixed_cnt > 0:
    with open(ch9_path, 'w', encoding='utf-8') as f:
        f.writelines(new_lines)
    print("Saved 제9장_측량.md successfully.")