To pull and start the persistent postgres database container, run

POSTGRES_DB=postgres POSTGRES_USER=postgres POSTGRES_PASSWORD=<password> \
docker run -d -p 5433:5432 \
-v <absolute_path_to_desired_data_location_on_host>:/var/lib/postgresql/data \ 
--name postgres-finances-automation postgres:11

where <password> is your choice of password for the database, and <absolute_path_to_desired_data_location_on_host> is the location in which to store the database cluster on your machine. The database password can be stored locally by setting it as an environment variable, or declared with each command run.

After the database has been pulled and started for the first time, it can be stopped by

docker stop postgres-finances-automation

and started again by

docker start postgres-finances-automation

The database container must be running for any of the below commands to work properly.

pip3 install -e .[development]

or

pip3 install setup.py

The configuration for the database, tables, categories and more is included in finances_automation/configuration.py. To initialise the database with the tables from the configuration file, run

python3 finances_automation/initialise_database.py

from the repository root.

From the repository root and with the POSTGRES_PASSWORD set, run

finances-automation --help

The output will look like this:

usage: finances-automation [-h] {parse,categorise,analyse,view_latest} ...

Automate your finances analysis.

optional arguments:
  -h, --help            show this help message and exit

Subcommands:
  {parse,categorise,analyse,view_latest}
    parse               Parse a UTF-8 .csv financial statement.
    categorise          Categorise transactions from a transactions table.
    analyse             Analyse transactions from a transactions table.
    view_latest         View the latest entries to a table.

From the repository root, run:

pytest

The aim of this project is to automate the manual review of my finances I carry out each month in Excel. The project will involve:

• Automating the ingestion of financial statements in .csv or .xlsx format
• Creating a database for financial data
• Automating the categorisation of each transaction into e.g. travel, food etc. This will require some machine learning, although will initially be carried out manually by asking the user to label each transaction. This manual method will still be quicker and easier than the current method, and has the secondary use of creating a labelled dataset for a model to be trained on
• Calculating the expenditure in each category and producing summary statistics related to a pre-defined budget. This will include e.g. average expenditures (monthly, yearly etc.) and a forecast based on previous months/years.
• Creating a dashboard displaying these statistics in an elegant and insightful fashion (I'll probably use seaborn here). The interface will allow the user to see statistics for different date ranges and bank accounts.
• Providing a command line interface to the above