wojtekms / banker

App that parses HTML output of PKO bank transactions.

Banker

This app helps analyze PKO banking transactions by providing a grouping and a summary of transactions.

Manual

In order to use this application:

• download transactions from your bank website for a specific month
• run app:
  • select downloaded transactions
  • select "analysis" and pick output folder
• app produces files:
  • autogen_budget.xlsx which contains summary of your expenses categorized into groups
  • unmatched_transactions.html which contains transactions that couldn't be categorized automatically

App Personalization

Please keep in mind that application provides best value only when it is configured according to your own needs. This can be done by modification of categories.json file.

This file is located in this repo, path: src/resources/categories.json.

This file contains a list of categories with a:

• category name
• payment type (can be only one of: ["household", "recurring", "occasional", "optional"])
• list of matching regexes (all transactions that match at least one of these regexes will be applied to this category)

When adding your own categories please pay attention to correctly specify matching regexes, as they are inside JSON file (for example you have to use double escape slash to escape a special character in regex), eg: "EURO\\-NET"

Personalization recommendations

When adding new categories this simple heuristic for selecting payment type might help you:

• use ocassional if this transaction had to be done by certain deadline (eg: motor liability insurance; present for someone's birthday;)
• use optional if this transaction was not required in a specific month and could be postponed to another one (eg: going out to a theater; buying new furniture;)
• use recurring if this transaction is part of your monthly routine (eg: flat rental)
• use household if this transaction is part of your daily routine (eg: groceries; fuel; drugs from pharmacy; gym;)

Development setup

python3 -m venv env
source env/bin/activate
pip install -e .[dev]

banker's Issues

Consider using open banking API

Documentation:

• https://developers.pkobp.pl/documentation
• https://developers.pkobp.pl/
• https://enablebanking.com/open-banking-apis/pl00102/

read only data about history of transactions would be helpful and ease use of the app

Add output formatting

• matched categories should be formatted to CSV
• unmatched transactions should be formatted to HTML table

Make sure you use interfaces (one for transactions and one for categories) for output formatting.

Add loading categories file

Add support to select custom categories file via GUI

Consider refunds

Sometimes you buy something and then you return it back - acquiring a refund. Currently this is not taken into account when doing expense summary.

Eg: eobuwie

Analyze only transactions with negative value

In case any transaction has a positive value (that means money was payed into the account, not that the money was payed out of the account) it should be logged and ignored. We only want to analyze negative payments.

Add basic GUI support

Add 2 buttons:

• pick an input file with HTML transactions
• analyze transactions

Perhaps analyze transactions should also require user to pick output directory?

Add Windows support

App should be executed on windows

Add parsing categories from file

Categories in file will be treated as a config file for app. Bear in mind that file should be versioned using semantic versioning.

Add logging to file

When user will be using app via GUI logging will have to be redirected to a file.

NOTE: find correct log folder for both Linux and Windows. On windows logs cannot be saved to the same directory where the app .exe is located

Remove common/naming.py

This is a bad idea to keep these things shared in a common module.

Parser should have his own data and formatter should have its own data to help modularize application.

Add Windows CI

Includes test & deploying ready .exe for windows

Pyinstaller can be used to create an .exe file. On windows run:

• git clone banker
• cd banker
• pip install pyinstaller
• python -m venv env
• .\env\Scripts\activate
• pip install .
• pyinstaller --noconfirm --clean --onefile --windowed --add-data src\banker\resources:banker\resources -p env\Lib\site-packages src\banker\__main__.py