はじめる
Statement Sorter — Categorized Monthly Table

Statement Sorter — Categorized Monthly Table

Send statement rows and category keywords. Get categorized transactions and monthly arithmetic you can inspect. Inputs: UTF-8 CSV. Outputs: CSV, Markdown, or JSON. Local processing; no bank connection.
#財務
評価
さらに評価が必要です
販売数
0
利用方法
ダウンロード

Statement Sorter — Bank Lines to a Categorized Monthly Table

Send bank or card statement rows (CSV) and the category names or keyword rules you want. Get a categorized row table, monthly category totals, and the addition formula for each total.

This local command-line tool reads UTF-8 CSV with date, description, and amount. It supports Python 3.9 and newer, does not connect to a bank, and makes no accounting or tax recommendations. Amount text is preserved in output; totals use decimal arithmetic. Amount inputs may use a period decimal, comma digit grouping, full-width digits, or a leading ¥/$; unmatched descriptions are marked Review unless a whole keyword matches.

Run

python3 scripts/statement_sorter.py examples/statement.csv --output /tmp/sorted.csv

Run package tests from this directory with python3 -m pytest -q tests (pytest required).

Options

Option Purpose
--format csv|markdown|json Choose output format
--rules categories.json Set category keywords
--month YYYY-MM Filter one month
--category NAME Filter category; repeatable
--min-amount N / --max-amount N Filter absolute amounts
--review-only Keep uncategorized rows
--list-categories Show built-in categories

Totals are grouped by month only for valid calendar dates written as YYYY-MM-DD. Dates that do not match that form, including impossible dates such as 2026-13-01 or 2026-02-30, remain visible under Unparsed date. A Review category signals a manual decision; it does not guess.

Input contract

UTF-8 CSV only. Required columns are date, description, and amount. Supported aliases include memo/details and transaction_date. Amounts accept a period decimal, optional comma digit grouping (for example, 1,200.00), full-width digits, or a leading ¥/$. The original amount text stays in the output; totals use its parsed numeric value. Other currency symbols, locale-specific decimal commas, and malformed grouping are unsupported. Unreadable or unsupported files return one error: line with exit code 2.