import json
import sys
import re
from pathlib import Path

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

# Load files
md_path = Path("resources/knowledge/original/행정규칙/임도 품셈 적용기준 (현 산림사업 표준품셈)/첨부/(산림청고시 제2025-82호) 산림사업 표준품셈.md")
with open(md_path, "r", encoding="utf-8") as f:
    md_lines = f.readlines()

with open('resources/data_cost_input_value/pum_forest_2026.json', 'r', encoding='utf-8') as f:
    pum_data = json.load(f)

pum_tables = pum_data.get('variables', {}).get('pum', {}).get('tables', [])
pum_map = {t['table_id']: t for t in pum_tables}

with open('resources/data_work_item_master/work_item_master_2026-01-01.json', 'r', encoding='utf-8') as f:
    master = json.load(f)

master_tables = {}
for wi in master['work_items']:
    for t in wi.get('tables', []):
        tid = t.get('pum_table_id')
        if tid:
            master_tables[tid] = t
for t in master.get('orphan_tables', []):
    tid = t.get('pum_table_id')
    if tid:
        master_tables[tid] = t

# Extract all markdown tables from MD
md_tables = []
current_table = []
start_line = -1

for idx, line in enumerate(md_lines):
    line_str = line.strip()
    if line_str.startswith('|') and line_str.endswith('|'):
        if not current_table:
            start_line = idx + 1
        current_table.append((idx + 1, line))
    else:
        if current_table:
            has_sep = any(re.match(r'^\|(?:\s*:?-+:?\s*\|)+$', l.strip()) for _, l in current_table)
            if has_sep:
                md_tables.append({
                    'start_line': start_line,
                    'end_line': current_table[-1][0],
                    'lines': [l for _, l in current_table]
                })
            current_table = []
if current_table:
    has_sep = any(re.match(r'^\|(?:\s*:?-+:?\s*\|)+$', l.strip()) for _, l in current_table)
    if has_sep:
        md_tables.append({
            'start_line': start_line,
            'end_line': current_table[-1][0],
            'lines': [l for _, l in current_table]
        })

print(f"MD tables: {len(md_tables)}, pum_forest tables: {len(pum_tables)}")

# Function to parse table rows from md lines
def parse_md_cells(lines):
    rows = []
    for l in lines:
        l_str = l.strip()
        if re.match(r'^\|(?:\s*:?-+:?\s*\|)+$', l_str):
            continue
        cells = [c.strip() for c in l_str.split('|')[1:-1]]
        rows.append(cells)
    return rows

# Compare each table's cells between MD and pum_forest
cell_mismatches = []
row_count_mismatches = []

for i, mdt in enumerate(md_tables):
    tid = f"F{i+1:04d}"
    pt = pum_map.get(tid)
    if not pt:
        continue
    
    md_rows = parse_md_cells(mdt['lines'])
    # In pum_tables, headers is first row(s), rows is data rows
    pt_headers = pt.get('headers', [])
    pt_data_rows = pt.get('rows', [])
    pt_all_rows = []
    if pt_headers:
        pt_all_rows.append(pt_headers)
    pt_all_rows.extend(pt_data_rows)
    
    # Check row counts
    if len(md_rows) != len(pt_all_rows):
        row_count_mismatches.append((tid, mdt['start_line'], len(md_rows), len(pt_all_rows)))
        continue
        
    for r_idx, (mr, pr) in enumerate(zip(md_rows, pt_all_rows)):
        if len(mr) != len(pr):
            cell_mismatches.append((tid, mdt['start_line'] + r_idx, f"Col count diff: md={len(mr)} vs pt={len(pr)}"))
            continue
        for c_idx, (mc, pc) in enumerate(zip(mr, pr)):
            # Normalize strings for comparison
            # remove spaces, trailing .0, etc.
            mc_norm = mc.replace(" ", "").replace(",", "")
            pc_norm = str(pc).replace(" ", "").replace(",", "")
            if mc_norm != pc_norm:
                # Check numeric equivalence
                try:
                    if float(mc_norm) == float(pc_norm):
                        continue
                except:
                    pass
                cell_mismatches.append((tid, mdt['start_line'] + r_idx, f"Cell[{r_idx},{c_idx}] diff: md='{mc}' vs pt='{pc}'"))

print(f"Row count mismatches: {len(row_count_mismatches)}")
for r in row_count_mismatches[:10]:
    print(f"  {r[0]} line {r[1]}: md rows={r[2]} vs pt rows={r[3]}")

print(f"\nCell mismatches: {len(cell_mismatches)}")
for cm in cell_mismatches[:20]:
    print(f"  {cm[0]} line {cm[1]}: {cm[2]}")