BNP_PDF_statement_parser/bnp_pdf_statement_parser.py

99 lines
3.0 KiB
Python
Executable File

#!/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
import operator
PATH = f'/home/benjamin/Desktop/bens_folder/bazaar/documents/bnp/bank_statements/'
os.chdir(PATH)
PRINT_TRANSACTIONS = False
MAIN_BANK_ACCOUNT = 'compte_de_cheques'
debits = []
credits_ = []
allTransactions = []
monthlyTransactions = {}
def appendTransactions(transactions, bankAccount):
global allTransactions
for transaction in transactions:
transaction['bank account'] = bankAccount
allTransactions += [transaction]
date = transaction['date'].replace(day = 1)
monthlyTransactions[date] = monthlyTransactions.get(date, []) + [transaction]
for folder in os.listdir():
if folder != MAIN_BANK_ACCOUNT:
for file in os.listdir(folder):
filePath = f'{folder}/{file}'
print(filePath)
print(readPdfBankStatement(filePath))
exit(0)
transactions = readPdfBankStatement(filePath)[3]
appendTransactions(transactions, folder)
#exit(1)
os.chdir(f'{MAIN_BANK_ACCOUNT}/')
for folder in sorted(os.listdir()):
for file in sorted(os.listdir(folder)):
filePath = f'{folder}/{file}'
print(filePath)
transactions = readPdfBankStatement(filePath)[3]
appendTransactions(transactions, MAIN_BANK_ACCOUNT)
if PRINT_TRANSACTIONS:
for transaction in transactions:
printTransaction(transaction)
#break
#break
from pprint import pprint
allTransactions.sort(key = operator.itemgetter('date'))
print(len(allTransactions))
pprint(allTransactions)
sortedMonths = sorted(monthlyTransactions.keys())
#pprint(sortedMonths)
# debit or credit?
totalMonthlyCredits = []
totalMonthlyDebits = []
totalMonthlyDifferences = []
#totalMonthlyDebits = [[ for transaction in monthlyTransactions[month]] for month in sortedMonths]
for month in sortedMonths:
#for transaction in monthlyTransactions[month]:
currentMonthlyTransactions = monthlyTransactions[month]
monthlyCredits = []
fig, ax = plt.subplots()
plt.title('Monthly debits and credits')
plt.xlabel('Date')
plt.ylabel('')
ALPHA = 0.5
xTicks = range(getMonthIndexSinceEpoch(sortedMonths[0]), getMonthIndexSinceEpoch(sortedMonths[-1]) + 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()