import os, sys

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

merged_path = "resources/knowledge/original/원가계산/건설공사_표준품셈/2026년_건설공사_표준품셈.md"
with open(merged_path, 'r', encoding='utf-8') as f:
    text_merged = f.read()

initial_len = len(text_merged)
applied_count = 0

# Import replacements from previous fix scripts
fix_scripts = [
    "scratch/fix_common_squashed_tables.py",
    "scratch/fix_civil_squashed_tables.py",
    "scratch/clean_civil_ch5_ch6.py",
    "scratch/fix_arch_squashed_tables.py",
    "scratch/fix_mech_squashed_tables.py",
    "scratch/fix_ch13_pipes.py",
    "scratch/fix_ch13_pipes_part2.py",
    "scratch/fix_maint_squashed_tables.py",
]

# Run regex/string replacements on text_merged
# Extract target and repl pairs from scripts
for script_path in fix_scripts:
    with open(script_path, 'r', encoding='utf-8') as f:
        code = f.read()
    
    # Execute code in a local dict to get targets and repls
    loc = {}
    exec(code, {}, loc)
    
    # Find all target_* and repl_* variables
    pairs = []
    for k in loc:
        if k.startswith('target_'):
            suffix = k[7:]
            repl_k = 'repl_' + suffix
            if repl_k in loc:
                pairs.append((loc[k], loc[repl_k]))
        elif k.startswith('old_l'):
            suffix = k[5:]
            repl_k = 'new_l' + suffix
            if repl_k in loc:
                pairs.append((loc[k], loc[repl_k]))
                
    for tgt, rpl in pairs:
        if tgt in text_merged:
            text_merged = text_merged.replace(tgt, rpl)
            applied_count += 1
            print(f"Applied replacement: {tgt[:40]}...")

print(f"\nTotal replacements applied to merged file: {applied_count}")

with open(merged_path, 'w', encoding='utf-8') as f:
    f.write(text_merged)

print(f"Merged file updated: len {initial_len} -> {len(text_merged)}")