import json
import sys
import re
from pathlib import Path

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

sys.path.insert(0, '.')
from scratch.enrich_tables import 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

# We want to know: Did the trimming cut away REAL pricing coefficients/conditions?
# Keywords that indicate actual pricing rules:
rule_keywords = ['%', '가산', '할증', '공제', '손료', '손율', '감한다', '할인', '포함한다', '별도', '규격', '작업량']

cut_real_rules = []

for t in tables:
    tid = t['table_id']
    mt = master_tables.get(tid)
    if not mt:
        continue
        
    md_notes = t.get('notes', [])
    master_notes = mt.get('notes', [])
    
    # Check lines in md_notes that are NOT in master_notes
    for line in md_notes:
        line_clean = line.strip()
        if not line_clean:
            continue
        # check if line_clean is present in master_notes
        if not any(line_clean in mn for mn in master_notes):
            # Is this line a section title like "- 3. ..." or "- 라. ..."?
            # Check if it looks like a section heading
            is_heading = bool(re.match(r'^[-*•]?\s*(?:[0-9]+[\.\)]|[가-힣][\.\)]|[A-Z][\.\)])\s*[^:%~]+$', line_clean))
            is_example = '예시' in line_clean or line_clean.startswith('>')
            
            # Check if it has real pricing rules
            has_rule = any(kw in line_clean for kw in rule_keywords)
            
            if has_rule and not is_example and not is_heading:
                cut_real_rules.append({
                    'table_id': tid,
                    'chapter': t['chapter'],
                    'section': t['section'],
                    'line': t['start_line'],
                    'lost_rule': line_clean
                })

print(f"Total potential REAL RULES cut away: {len(cut_real_rules)}")
for cr in cut_real_rules:
    print(f"[{cr['table_id']} | line {cr['line']}] {cr['section']}")
    print(f"   CUT: {cr['lost_rule']}")