BNP_PDF_statement_parser/utils.py

95 lines
4.3 KiB
Python
Raw Normal View History

import subprocess
from datetime import datetime
import re
FIRST_LINE_OF_PAYMENT_REGEX = re.compile('(\\d{2}\\.\\d{2}) (\\d{2}\\.\\d{2}) ([\\d ]+,\\d{2})')
END_PAGE_AFTER_THE_FIRST_ONE_REGEX = re.compile('P\\. \\d+/\\d+')
SOLDE_CREDITEUR_AU_REGEX = re.compile('SOLDE CREDITEUR AU \\d{2}\\.\\d{2}\\.\\d{4}')
TOTAL_DES_OPERATIONS_REGEX = re.compile('TOTAL\\ DES\\ OPERATIONS\\ ([\\d ]+,\\d{2})\\ ([\\d ]+,\\d{2})')
TOTAL_DES_OPERATIONS_CREDIT_ONLY_REGEX = re.compile('TOTAL\\ DES\\ OPERATIONS\\ ([\\d ]+,\\d{2})')
def execute(command):
return subprocess.check_output(command).decode('utf-8')
def getTextFromPdf(pdfPath):
return execute(['pdftotext', '-raw', pdfPath, '-'])
def getDatetimeFromFileName(aDatetimeStr):
2024-10-03 19:33:54 +02:00
#aDatetime = datetime.strptime(aDatetimeStr, '%Y%m%d.pdf')
aDatetime = datetime(int(aDatetimeStr[:4]), int(aDatetimeStr[4:6]), 1)
return aDatetime
def getMonthIndexSinceEpoch(aDatetime):
return aDatetime.year * 12 + aDatetime.month
def getMonthNameFromMonthIndex(monthIndex):
2024-10-03 19:51:54 +02:00
return datetime((monthIndex - 1) // 12, 1 + (monthIndex - 1) % 12, 1).strftime('%b %Y')
def toFloat(group):
return float(group.replace(',', '.').replace(' ', ''))
2024-10-03 19:51:54 +02:00
def readPdfBankStatement(filePath):
file = filePath.split('/')[-1]
fileDatetime = getDatetimeFromFileName(file)
content = getTextFromPdf(filePath)
lines = content.splitlines()
#print('\n'.join(lines))
2024-10-03 19:51:54 +02:00
started = False
firstPage = True
initialAmount = None
#currentAmount = None
2024-10-03 19:51:54 +02:00
date = None
comment = []
transactions = []
for line in lines:
if not started:
# We are interested in the content after this line:
if SOLDE_CREDITEUR_AU_REGEX.match(line) is not None or (line.startswith('Date Nature des opérations Valeur Débit Crédit') and not firstPage):
if SOLDE_CREDITEUR_AU_REGEX.match(line):
initialAmount = float(SOLDE_CREDITEUR_AU_REGEX.sub('', line).replace(',', '.').replace(' ', ''))
#currentAmount = initialAmount
2024-10-03 19:51:54 +02:00
started = True
continue
else:
# We aren't interested in the content after this line:
if line.startswith('BNP PARIBAS SA au capital de') or END_PAGE_AFTER_THE_FIRST_ONE_REGEX.match(line) is not None:
firstPage = False
started = False
continue
# We aren't interested in the content after this line
elif line.startswith('TOTAL DES OPERATIONS'):
totalDesOperationsRegexMatch = TOTAL_DES_OPERATIONS_REGEX.match(line)
# Note that transfer between accounts will be noted in both debits and credits, as trying to cancel would make benefits show as negative debit which does not make sense.
# Cannot just consider January as benefits only as `20240122.pdf` also contains an additional transfer between my accounts.
if totalDesOperationsRegexMatch is not None:
totalMonthlyDebit, totalMonthlyCredit = [toFloat(group) for group in totalDesOperationsRegexMatch.groups()]
else:
totalMonthlyCredit = toFloat(TOTAL_DES_OPERATIONS_CREDIT_ONLY_REGEX.match(line).group(1))
totalMonthlyDebit = 0
break
firstLineOfPaymentRegexMatch = FIRST_LINE_OF_PAYMENT_REGEX.match(line)
if firstLineOfPaymentRegexMatch is not None:
2024-10-03 19:51:54 +02:00
if date is not None:
transactions += [{
'date': date,
'valeur': valeur,
'amount': amount,
#'currentAmount': currentAmount,
2024-10-03 19:51:54 +02:00
'comment': '\n'.join(comment)
}]
2024-10-03 20:10:06 +02:00
date = None
date, valeur, amount = firstLineOfPaymentRegexMatch.groups()
amount = toFloat(amount)
#currentAmount -= amount
2024-10-03 19:51:54 +02:00
comment = []
else:
comment += [line]
2024-10-03 20:10:06 +02:00
if date is not None:
transactions += [{
'date': date,
'valeur': valeur,
'amount': amount,
#'currentAmount': currentAmount,
2024-10-03 20:10:06 +02:00
'comment': '\n'.join(comment)
}]
2024-10-03 19:51:54 +02:00
return initialAmount, totalMonthlyDebit, totalMonthlyCredit, transactions, fileDatetime