BNP_PDF_statement_parser/bnp_pdf_statement_parser.py

#!/usr/bin/env python

# Depends on `pdftotext`.

import os
import subprocess
import re
import matplotlib.pyplot as plt
import matplotlib.ticker as ticker
from datetime import datetime

PATH = f'/home/benjamin/Desktop/bens_folder/bazaar/documents/bnp/bank_statements/compte_de_cheques/'

os.chdir(PATH)

'''
Assuming file hierarchy like:

2022
├── 20221121.pdf
└── 20221221.pdf
2023
├── 20230123.pdf
└── 20230221.pdf
'''

def execute(command):
    return subprocess.check_output(command).decode('utf-8')

def getTextFromPdf(pdfPath):
    return execute(['pdftotext', '-raw', pdfPath, '-'])

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\\ ([0-9 ]+,\\d{2})\\ ([0-9 ]+,\\d{2})')

PRINT_TRANSACTIONS = False

totalMonthlyDebits = []
totalMonthlyCredits = []
totalMonthlyDifferences = []
totals = []
firstDatetime = None
lastDatetime = None
for folder in sorted(os.listdir()):
    for file in sorted(os.listdir(folder)):
        filePath = f'{folder}/{file}'
        print(filePath)
        currentDatetime = getDatetime(file)
        if firstDatetime is None:
            firstDatetime = currentDatetime
        content = getTextFromPdf(filePath)
        lines = content.splitlines()
        started = False
        firstPage = True
        initialAmount = None
        currentAmount = None
        date = None
        comment = []
        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
                        print('Initial amount', initialAmount)
                        print()
                        totals += [initialAmount]
                    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)
                    totalMonthlyDebit, totalMonthlyCredit = [float(group.replace(',', '.').replace(' ', '')) for group in totalDesOperationsRegexMatch.groups()]
                    print(f'Total monthly debit: {totalMonthlyDebit}')
                    print(f'Total monthly credit: {totalMonthlyCredit}')
                    totalMonthlyDebits += [totalMonthlyDebit]
                    totalMonthlyCredits += [totalMonthlyCredit]
                    totalMonthlyDifference = totalMonthlyCredit - totalMonthlyDebit
                    totalMonthlyDifferences += [totalMonthlyDifference]
                    break
                if FIRST_LINE_OF_PAYMENT_REGEX.match(line) is not None:
                    if date is not None and PRINT_TRANSACTIONS:
                        print(date, valeur, amount, currentAmount)
                        print('\n'.join(comment))
                        print()
                    date, valeur, amount = line.split()
                    amount = float(amount.replace(',', '.'))
                    currentAmount -= amount
                    comment = []
                else:
                    comment += [line]
        #break
    #break
lastDatetime = getDatetime(file)

fig, ax = plt.subplots()
plt.title('Monthly debits and credits')
plt.xlabel('Date')
plt.ylabel('€')
ALPHA = 0.5

def getDatetime(aDatetimeStr):
    return datetime.strptime(aDatetimeStr, '%Y%m%d.pdf')

def getMonthIndex(aDatetime):
    return aDatetime.year * 12 + aDatetime.month

xTicks = range(getMonthIndex(firstDatetime), getMonthIndex(lastDatetime) + 1)
# sign does not seem respected for `totalMonthlyDifferences`.
totalMonthlyAmountAndLabel = (
    #(totalMonthlyDebits, 'Debit'),
    #(totalMonthlyCredits, 'Credit'),
    (totalMonthlyDifferences, 'Difference'),
    (totals, 'Total'),
)
for totalMonthlyAmount, totalMonthlyLabel in totalMonthlyAmountAndLabel:
    plt.bar(xTicks, totalMonthlyAmount, alpha = ALPHA, label = totalMonthlyLabel)
plt.legend()

#plt.yscale('symlog')
ax.yaxis.set_major_formatter(ticker.StrMethodFormatter('{x:,}'))

def getMonthName(monthIndex):
    return datetime((monthIndex - 1) // 12, 1 + (monthIndex - 1) % 12, 1).strftime('%b %Y')

ticksLabels = [getMonthName(monthIndex) for monthIndex in xTicks]
plt.xticks(xTicks, ticksLabels, rotation = 90)
#plt.tight_layout()
# How to show the horizontal lines for subticks?
plt.grid(axis = 'y')

plt.show()