import json
import sys
import re
from pathlib import Path
sys.path.insert(0, str(Path(__file__).parent.parent))
from scratch.enrich_tables import tables, get_chapter

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

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

# Check chapter 5, 9, 12 specifically
for ch_num in ['제5장', '제9장', '제12장']:
    print(f"\n==================== {ch_num} ====================")
    ch_tables = [t for t in tables if ch_num in t['chapter']]
    for t in ch_tables:
        tid = t['table_id']
        mt = master_tables.get(tid)
        if not mt:
            continue
        m_qty = mt.get('basis_quantity')
        m_unit = mt.get('basis_unit')
        pform = mt.get('pum_form')
        
        # Look for unit patterns in pre_context and headers
        pre_text = " // ".join(t['pre_context'][-5:])
        notes_text = " // ".join(t['notes'][:3]) if t['notes'] else ""
        header_text = t['lines'][0].strip() if t['lines'] else ""
        
        # Check if basis missing or interesting
        if m_qty is None or m_unit is None or pform == 'reference':
            print(f"[{tid} | line {t['start_line']}] sec: {t['section']}")
            print(f"   pre: {pre_text}")
            print(f"   hdr: {header_text}")
            print(f"   master basis: {m_qty} {m_unit} | form: {pform}")
            if notes_text:
                print(f"   notes: {notes_text}")