import re
import difflib

try:
    import Levenshtein
except ImportError:
    import subprocess
    subprocess.check_call(["pip", "install", "python-Levenshtein", "--quiet"])
    import Levenshtein

def extract_myanmar(text):
    return ''.join(c for c in text if '\u1000' <= c <= '\u109f')

with open('./An Duc Tam Bao/Tam_Bao_051_060_standard.txt', 'r', encoding='utf-8') as f:
    standard = extract_myanmar(f.read())

with open('./An Duc Tam Bao/tam_bao_051_060.md', 'r', encoding='utf-8') as f:
    extracted = extract_myanmar(f.read())

distance = Levenshtein.distance(standard, extracted)
max_len = max(len(standard), len(extracted))
accuracy = (1 - distance / max_len) * 100

print(f"Độ chính xác (Levenshtein): {accuracy:.2f}%")
print(f"Khoảng cách Levenshtein: {distance}")
