beancount-import example

import.py

#!/usr/bin/env python3

import glob
import os

import beancount_import.webserver

## I keep my scripts in a subdirectory relative to the main beancount
## file, so the main directory is this script's directory's parent.
JOURNAL_DIR = os.path.dirname(os.path.dirname(__file__))
DATA_DIR = os.path.join(JOURNAL_DIR, 'data')
CACHE_DIR = os.path.join(JOURNAL_DIR, 'cache')

if __name__ == '__main__':
    data_sources = [
        ## OFX data is in directories as "data/<institution>/<account-type>/<month>.ofx"
        ## e.g. "data/Bank-A/Checking/2021-10.ofx"
        ## The exception is one brokerage that gives me combined files for
        ## all of my accounts with them, plus they only do quarterly
        ## statements, so it has stuff like "data/Broker-B/2021-Q3.ofx"
        dict(
            module='beancount_import.source.ofx',
            ofx_filenames=(
                glob.glob(os.path.join(DATA_DIR, 'Bank-A/Checking/*.ofx')) \
                + glob.glob(os.path.join(DATA_DIR, 'Broker-B/*.ofx')) \
                + glob.glob(os.path.join(DATA_DIR, 'Bank-C/Credit-Card/*.ofx')) \
                + glob.glob(os.path.join(DATA_DIR, 'Bank-D/Savings/*.ofx')) \
                + glob.glob(os.path.join(DATA_DIR, 'Bank-D/Checking/*.ofx')) \
                + glob.glob(os.path.join(DATA_DIR, 'Bank-D/Credit-Card/*.ofx'))
            ),
            cache_filename=os.path.join(CACHE_DIR, 'ofx_cache.pickle')
        ),
        ## Amazon data is in "data/Amazon/<order-number>.html".  I download
        ## the invoice files from Amazon and then have a script that
        ## renames the files according to the order number.  (The script
        ## also sets the file modification date to the order date.)
        dict(
            module='beancount_import.source.amazon',
            directory=os.path.join(DATA_DIR, 'Amazon'),
            amazon_account='[email protected]',
            posttax_adjustment_accounts={
                'Gift Card Amount': 'Assets:Cash:Amazon:Gift-Cards',
                'Rewards Points': 'Assets:Cash:Chase:Rewards',
            },
        ),
    ]
    beancount_import.webserver.main(
        [],
        ## My main beancount file is "main.beancount", which basically only
        ## has include directives to pull in all the other files I use.
        journal_input=os.path.join(JOURNAL_DIR, 'main.beancount'),
        ## The ignored.beancount file is not included by main.beancount, so
        ## I can drop transactions here if I don't want them in my records
        ## for some reason.
        ignored_journal=os.path.join(JOURNAL_DIR, 'ignored.beancount'),
        ## While I have fairly granular files for my hand-maintained data,
        ## I just have a single "journal.beancount" file for most imported
        ## data.  I basically trat that file as write-only; I look at the
        ## transactions once, in the importer, and then never again.
        default_output=os.path.join(JOURNAL_DIR, 'journal.beancount'),
        ## Account definitions are in the "chart-of-accounts.beancount"
        ## file.  The importer adds new entries at the bottom and I
        ## periodically go in and sort the account definitions into
        ## appropriate sections earlier in the file.
        open_account_output_map=[
            ('.*', os.path.join(JOURNAL_DIR, 'chart-of-accounts.beancount')),
        ],
        ## I don't actually use imported balance assertions; I enter them
        ## by hand from the bank statements.  This setting is leftover
        ## from when I was initially setting up the importer.
        balance_account_output_map=[
            ('Bank-A', os.path.join(JOURNAL_DIR, 'bank-a.beancount')),
            ('.*', os.path.join(JOURNAL_DIR, 'journal.beancount')),
        ],
        data_sources=data_sources,
        verbose=True,
        dev=True,
    )