import os
import re
from shapely.geometry import Polygon
from algam_pars_mif import UnpackingMif

class MifData:
    def __init__(self,directory,name,cad_number,additional_info,square):
        self.directory:str=directory
        self.name:str=name
        self.cad_number:str=cad_number
        self.additional_info:str=additional_info
        self.square:float=square

class ReadMifDirectory():
    def __init__(self):
        pass
    
    def find_mif_files(self,directory:str):
        mif_files = []
        for root, dirs, files in os.walk(directory):
            for file in files:
                if file.lower().endswith('.mif'):
                    # Ищем последовательность чисел, разделённых чем угодно, в начале имени файла
                    cad_match = re.search(r'(\d{2,}([_\W]+)?)+\d{2,}', file)
                    if cad_match:
                        # Извлекаем только цифры из найденной последовательности
                        numbers = re.findall(r'\d+', cad_match.group())
                        cad_number = ':'.join(numbers)
                        # Всё, что после кадастрового номера — это аббревиатуры и прочее
                        additional_info = file[cad_match.end():]
                        # Удаляем расширение
                        additional_info = re.sub(r'\.[^.]+$', '', additional_info)
                        # Чистим пробелы и спецсимволы
                        additional_info = re.sub(r'\s+', '_', additional_info)
                        additional_info = re.sub(r'^[^\wа-яА-ЯёЁ]*', '', additional_info)
                        additional_info = re.sub(r'[^\wа-яА-ЯёЁ_]', '', additional_info)
                        additional_info = re.sub(r'_+', '_', additional_info).strip('_')
                    else:
                        cad_number = ""
                        additional_info = re.sub(r'\.[^.]+$', '', file)
                        additional_info = re.sub(r'\s+', '_', additional_info)
                        additional_info = re.sub(r'^[^\wа-яА-ЯёЁ]*', '', additional_info)
                        additional_info = re.sub(r'[^\wа-яА-ЯёЁ_]', '', additional_info)
                        additional_info = re.sub(r'_+', '_', additional_info).strip('_')

                    mif_files.append(MifData(
                        directory=os.path.join(root, file),
                        name=os.path.basename(file),
                        cad_number=cad_number,
                        additional_info=additional_info,
                        square=UnpackingMif(file_directory=os.path.join(root, file))
                    ))
        if mif_files:
            print(f"{mif_files[0].directory=}/{mif_files[0].name=}/{mif_files[0].cad_number=}/{mif_files[0].additional_info=}/{mif_files[0].square=}")
        return mif_files