BNP_PDF_statement_parser/bnp_pdf_statement_parser.py

#!/usr/bin/env python

import os
import matplotlib.pyplot as plt
import matplotlib.ticker as ticker
from datetime import datetime
from utils import getDatetimeFromFileName, getMonthIndexSinceEpoch, getMonthNameFromMonthIndex, readPdfBankStatement

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

os.chdir(PATH)

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():#[-2:]:
    if folder != MAIN_BANK_ACCOUNT:
        print(folder)
        for file in os.listdir(folder):#[-1:]:
            print(file)
            filePath = f'{folder}/{file}'
            initialAmount, totalMonthlyDebit, totalMonthlyCredit, transactions, fileDatetime = readPdfBankStatement(filePath)
            #print(totalMonthlyCredit)
            #import json
            #print(json.dumps(transactions, indent = 4))
            otherBankAccountsCredits[fileDatetime] = otherBankAccountsCredits.get(fileDatetime, 0) + totalMonthlyCredit

#exit(1)
os.chdir(f'{MAIN_BANK_ACCOUNT}/')

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)
        initialAmount, totalMonthlyDebit, totalMonthlyCredit, transactions, fileDatetime = readPdfBankStatement(filePath)
        print('Initial amount', initialAmount)
        print()
        totals += [initialAmount]
        print(f'Total monthly debit: {totalMonthlyDebit}')
        print(f'Total monthly credit: {totalMonthlyCredit}')
        totalMonthlyDebits += [totalMonthlyDebit]
        otherBankAccountsCredit = otherBankAccountsCredits.get(fileDatetime, 0)
        print(f'{otherBankAccountsCredit=}')
        totalMonthlyCredit += otherBankAccountsCredit
        totalMonthlyCredits += [totalMonthlyCredit]
        totalMonthlyDifference = totalMonthlyCredit - totalMonthlyDebit
        totalMonthlyDifferences += [totalMonthlyDifference]
        if PRINT_TRANSACTIONS:
            for transaction in transactions:
                # , transaction['currentAmount']
                print(transaction['date'], transaction['valeur'], transaction['amount'])
                print(transaction['comment'])
                print()
        if firstDatetime is None:
            firstDatetime = fileDatetime
        #break
    #break
lastDatetime = getDatetimeFromFileName(file)

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

xTicks = range(getMonthIndexSinceEpoch(firstDatetime), getMonthIndexSinceEpoch(lastDatetime) + 1)
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:,}'))

ticksLabels = [getMonthNameFromMonthIndex(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()