Python Grundbuchauszug Parser – Grundbuchauszüge auslesen um diese in Excel, CSV oder andere Formate zu exportieren

				
					from pdfminer.pdfinterp import PDFResourceManager, PDFPageInterpreter
from pdfminer.converter import TextConverter
from pdfminer.layout import LAParams
from pdfminer.pdfpage import PDFPage
from io import StringIO
from pprint import pprint

def convert_pdf_to_txt(path):
    rsrcmgr = PDFResourceManager()
    retstr = StringIO()
    codec = 'utf-8'
    laparams = LAParams()
    device = TextConverter(rsrcmgr, retstr, codec=codec, laparams=laparams)
    fp = open(path, 'rb')
    interpreter = PDFPageInterpreter(rsrcmgr, device)
    password = ""
    maxpages = 0
    caching = True
    pagenos = set()

    for page in PDFPage.get_pages(fp, pagenos, maxpages=maxpages, password=password,caching=caching, check_extractable=True):
        interpreter.process_page(page)

    text = retstr.getvalue()

    fp.close()
    device.close()
    retstr.close()
    return text

def cleanup_pdf_text(text):

    output = []

    for line in text:

        if line != '':
            output.append(line)

    return output




pdf_text = cleanup_pdf_text(convert_pdf_to_txt('pdf/GB.pdf').split('\n'))

print(pdf_text)

output = {}
output['anteile'] = {}

temp_output = {}

current_stage = ''

current_substage = ''
current_substage_share = ''
current_substage_part = ''
current_substage_number = ''

for pdf_line in pdf_text:

    if 'Seite' in pdf_line:
        continue

    if '*** B ***' in pdf_line:

        current_stage = 'anteile'


    if current_stage == 'anteile':

        print(pdf_line)

        if 'ANTEIL' in pdf_line:

            print('------------------------------ Found new Anteil')

            current_substage_number = pdf_line.split('ANTEIL:')[0].strip().split(' ')[0].strip()

            print(current_substage_number)

            current_substage_share = pdf_line.split('ANTEIL:')[1].split('/')[0].strip()
            output['total_shares'] = pdf_line.split('ANTEIL:')[1].split('/')[1].strip()

            current_substage_part = 'anteilig'
            continue

        # next line is name of owner
        if current_substage_part == 'anteilig':

            current_substage_name = pdf_line.strip()

            temp_output = {}
            temp_output['share'] = current_substage_share

            temp_output['name'] = pdf_line.strip()
            temp_output['firstname'] = pdf_line.strip().split(' ')[-2]
            temp_output['lastname'] = pdf_line.strip().split(' ')[-1]

            if len(pdf_line.strip().split(' ')) > 2:
                temp_output['title'] = pdf_line.strip().split(' ')[0]
            else:
                temp_output['title'] = ''

            temp_output['family'] = '0'
            current_substage_part = 'anteilig-name'
            continue

        # next line is birthdate and address
        if current_substage_part == 'anteilig-name':

            print(pdf_line)

            address = pdf_line.split('ADR:')[1]

            address_with_country = address[:len(address) - 4].strip()

            if ',' in address_with_country:
                temp_output['address'] = address_with_country.split(',')[0].strip()
                temp_output['city'] = address_with_country.split(',')[1].strip()
            else:
                temp_output['address'] = address_with_country.strip()

            temp_output['postal_code'] = address[-4:].strip()

            current_substage_part = 'anteil_detail'
            continue

        if current_substage_part == 'anteil_detail':

            if 'Wohnungseigentum an' in pdf_line:

                temp_output['apartment'] = pdf_line.split('Wohnungseigentum an')[1].strip()

                # we need to find out if the current apartment is already stored in our output array, if not we append
                # if found we add a marker called family which indicates that more then one person owns given aparment

                add_to_output = True

                if len(output['anteile']) > 0:

                    for index, item in output['anteile'].items():

                        if item['apartment'] == temp_output['apartment']:

                            output['anteile'][index]['family'] = 1

                            if 'family_members' not in output['anteile'][index]:
                                output['anteile'][index]['family_members'] = []


                            output['anteile'][index]['family_members'].append(temp_output)

                            add_to_output = False

                if add_to_output is True:
                    output['anteile'][current_substage_number] = temp_output


                current_substage_part = 'finished'



pprint(output)