Benjamin_Loison/BNP_PDF_statement_parser
1
0
Fork 0
Code Issues 2 Pull Requests Packages Projects Releases Wiki Activity

WIP adding other bank accounts

Browse Source
This commit is contained in:
Benjamin Loison 2024-10-03 19:33:54 +02:00
parent bbe93d0939
commit b99838696e
Signed by: Benjamin_Loison
SSH Key Fingerprint: SHA256:BtnEgYTlHdOg1u+RmYcDE0mnfz1rhv5dSbQ2gyxW8B8
2 changed files with 36 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

36
bnp_pdf_statement_parser.py
View File

@ -4,14 +4,31 @@ import os
import matplotlib.pyplot as plt import matplotlib.pyplot as plt
import matplotlib.ticker as ticker import matplotlib.ticker as ticker
from datetime import datetime from datetime import datetime
from utils import getTextFromPdf, getDatetimeFromFileName, getMonthIndexSinceEpoch, getMonthNameFromMonthIndex, FIRST_LINE_OF_PAYMENT_REGEX, END_PAGE_AFTER_THE_FIRST_ONE_REGEX, SOLDE_CREDITEUR_AU_REGEX, TOTAL_DES_OPERATIONS_REGEX from utils import getTextFromPdf, getDatetimeFromFileName, getMonthIndexSinceEpoch, getMonthNameFromMonthIndex, FIRST_LINE_OF_PAYMENT_REGEX, END_PAGE_AFTER_THE_FIRST_ONE_REGEX, SOLDE_CREDITEUR_AU_REGEX, TOTAL_DES_OPERATIONS_REGEX, TOTAL_DES_OPERATIONS_CREDIT_ONLY_REGEX
PATH = f'/home/benjamin/Desktop/bens_folder/bazaar/documents/bnp/bank_statements/compte_de_cheques/' PATH = f'/home/benjamin/Desktop/bens_folder/bazaar/documents/bnp/bank_statements/'
os.chdir(PATH) os.chdir(PATH)
PRINT_TRANSACTIONS = False PRINT_TRANSACTIONS = False
MAIN_BANK_ACCOUNT = 'compte_de_cheques'
# As far as I know there was no debit yet.
otherBankAccountsCredits = {}
for folder in os.listdir():
if folder != MAIN_BANK_ACCOUNT:
print(folder)
for file in os.listdir(folder):
print(file)
fileDatetime = getDatetimeFromFileName(file)
print(fileDatetime)
#otherBankAccountsCredits[fileDatetime] = otherBankAccountsCredits.get(fileDatetime, 0) +
os.chdir(f'{MAIN_BANK_ACCOUNT}/')
totalMonthlyDebits = [] totalMonthlyDebits = []
totalMonthlyCredits = [] totalMonthlyCredits = []
totalMonthlyDifferences = [] totalMonthlyDifferences = []
@ -22,9 +39,9 @@ for folder in sorted(os.listdir()):
for file in sorted(os.listdir(folder)): for file in sorted(os.listdir(folder)):
filePath = f'{folder}/{file}' filePath = f'{folder}/{file}'
print(filePath) print(filePath)
currentDatetime = getDatetimeFromFileName(file) fileDatetime = getDatetimeFromFileName(file)
if firstDatetime is None: if firstDatetime is None:
firstDatetime = currentDatetime firstDatetime = fileDatetime
content = getTextFromPdf(filePath) content = getTextFromPdf(filePath)
lines = content.splitlines() lines = content.splitlines()
started = False started = False
@ -54,7 +71,16 @@ for folder in sorted(os.listdir()):
# We aren't interested in the content after this line # We aren't interested in the content after this line
elif line.startswith('TOTAL DES OPERATIONS'): elif line.startswith('TOTAL DES OPERATIONS'):
totalDesOperationsRegexMatch = TOTAL_DES_OPERATIONS_REGEX.match(line) totalDesOperationsRegexMatch = TOTAL_DES_OPERATIONS_REGEX.match(line)
totalMonthlyDebit, totalMonthlyCredit = [float(group.replace(',', '.').replace(' ', '')) for group in totalDesOperationsRegexMatch.groups()] # 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.
toFloat = lambda group: float(group.replace(',', '.').replace(' ', ''))
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
if os.getcwd().endswith(f'/{MAIN_BANK_ACCOUNT}'):
totalMonthlyCredit += otherBankAccountsCredits[fileDatetime]
print(f'Total monthly debit: {totalMonthlyDebit}') print(f'Total monthly debit: {totalMonthlyDebit}')
print(f'Total monthly credit: {totalMonthlyCredit}') print(f'Total monthly credit: {totalMonthlyCredit}')
totalMonthlyDebits += [totalMonthlyDebit] totalMonthlyDebits += [totalMonthlyDebit]

6
utils.py
View File

@ -6,6 +6,7 @@ FIRST_LINE_OF_PAYMENT_REGEX = re.compile('\\d{2}\\.\\d{2} \\d{2}\\.\\d{2} \\d+,\
END_PAGE_AFTER_THE_FIRST_ONE_REGEX = re.compile('P\\. \\d+/\\d+') 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}') SOLDE_CREDITEUR_AU_REGEX = re.compile('SOLDE CREDITEUR AU \\d{2}\\.\\d{2}\\.\\d{4}')
TOTAL_DES_OPERATIONS_REGEX = re.compile('TOTAL\\ DES\\ OPERATIONS\\ ([0-9 ]+,\\d{2})\\ ([0-9 ]+,\\d{2})') TOTAL_DES_OPERATIONS_REGEX = re.compile('TOTAL\\ DES\\ OPERATIONS\\ ([0-9 ]+,\\d{2})\\ ([0-9 ]+,\\d{2})')
TOTAL_DES_OPERATIONS_CREDIT_ONLY_REGEX = re.compile('TOTAL\\ DES\\ OPERATIONS\\ ([0-9 ]+,\\d{2})')
def execute(command): def execute(command):
return subprocess.check_output(command).decode('utf-8') return subprocess.check_output(command).decode('utf-8')
@ -14,7 +15,10 @@ def getTextFromPdf(pdfPath):
return execute(['pdftotext', '-raw', pdfPath, '-']) return execute(['pdftotext', '-raw', pdfPath, '-'])
def getDatetimeFromFileName(aDatetimeStr): def getDatetimeFromFileName(aDatetimeStr):
return datetime.strptime(aDatetimeStr, '%Y%m%d.pdf') #aDatetime = datetime.strptime(aDatetimeStr, '%Y%m%d.pdf')
#aDatetime = datetime.strptime(aDatetimeStr[:6], '%Y%m')
aDatetime = datetime(int(aDatetimeStr[:4]), int(aDatetimeStr[4:6]), 1)
return aDatetime
def getMonthIndexSinceEpoch(aDatetime): def getMonthIndexSinceEpoch(aDatetime):
return aDatetime.year * 12 + aDatetime.month return aDatetime.year * 12 + aDatetime.month