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_tables = 0

# Common tasks lists in survey
known_task_patterns = [
    ["계획준비", "답사선점", "조표(매설)", "관 측", "계 산", "정리점검"],
    ["계획준비", "답사선점", "조표(매설)", "관측", "계산", "정리점검"],
    ["계획준비", "답사선점", "관 측", "계 산", "정리점검"],
    ["계획준비", "답사선점", "관측", "계산", "정리점검"],
    ["계획준비", "답사", "선점", "매설", "관측", "계산", "정리점검"],
    ["계획준비", "선점", "매설", "관측", "계산", "정리점검"],
]

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]]
    
    # Check if first cell contains tasks
    matched_tasks = None
    c0 = cells[0]
    
    # Try exact task sequence match
    for pat in known_task_patterns:
        pat_str = " ".join(pat)
        if c0 == pat_str or c0.replace(" ", "") == "".join(pat).replace(" ", ""):
            matched_tasks = pat
            break
            
    if not matched_tasks:
        # Try generic tasks split if tokens are known tasks
        toks = c0.split()
        if len(toks) >= 4 and all(any(k in t for k in ["계획", "준비", "답사", "선점", "조표", "매설", "관측", "계산", "정리", "점검", "수준", "측량", "항공", "도화", "편집"]) for t in toks):
            matched_tasks = toks

    if matched_tasks:
        N = len(matched_tasks)
        # Check if following data columns can be split into N tokens
        # Some columns might have single value (like 비고) or N tokens
        can_split = True
        col_splits = {}
        for c_idx in range(1, len(cells)):
            val = cells[c_idx]
            if not val or val == "-":
                col_splits[c_idx] = [val] * N
                continue
            v_toks = val.split()
            if len(v_toks) == N:
                col_splits[c_idx] = v_toks
            elif c_idx == len(cells) - 1: # Last column is often 비고
                col_splits[c_idx] = [val] + [""] * (N - 1)
            else:
                # Cannot cleanly split this column
                can_split = False
                break
                
        if can_split:
            # Generate N rows
            for r_idx in range(N):
                row_cells = [matched_tasks[r_idx]]
                for c_idx in range(1, len(cells)):
                    row_cells.append(col_splits[c_idx][r_idx])
                new_line_str = "| " + " | ".join(row_cells) + " |\n"
                new_lines.append(new_line_str)
            fixed_tables += 1
            print(f"L{idx+1}: Fixed survey table with {N} tasks: {cells[0][:30]}...")
            continue

    new_lines.append(line)

print(f"\nTotal survey task tables fixed in 제9장_측량.md: {fixed_tables}")

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