bankstatementparser-writer-xlsx: Excel writer for parsed bank statements

An Excel .xlsx writer for data parsed by bankstatementparser — turn parsed transactions (a pandas DataFrame, a list of Transaction objects, or a list of plain dicts) into a polished workbook that accountants, auditors, and reconciliation macros can open directly.

Latest release: v0.0.13 — single write_xlsx(data, path, ...) function, 100% line + branch coverage, 100% docstring coverage, mypy --strict clean.

Contents

• Overview
• Install
• Quick start
• Input shapes
• Value coercion
• The summary sheet
• Examples
• When not to use this package
• Development
• Security
• Documentation
• License
• Contributing
• Acknowledgements

Overview

bankstatementparser-writer-xlsx is a small, focused companion to the bankstatementparser library. It does one thing well: given already-parsed bank-statement records, write a clean Excel workbook with a bold header row, one row per transaction, auto-sized columns, and an optional second Summary sheet.

The package consumes parsed data — it does not read PDFs, CSVs, or XML itself. Parsing (and the security surface that comes with untrusted input) lives upstream in the bankstatementparser core.

Install

bankstatementparser-writer-xlsx runs on macOS, Linux, and Windows and requires Python 3.10+. It pulls in bankstatementparser, openpyxl, and pandas automatically.

pip install bankstatementparser-writer-xlsx

Quick start

from bankstatementparser import CsvStatementParser
from bankstatementparser_writer_xlsx import write_xlsx

parser = CsvStatementParser("statement.csv")
df = parser.parse()                      # a pandas DataFrame
write_xlsx(df, "statement.xlsx")         # one polished workbook

That's an Excel workbook ready for your accountant. Add a summary sheet in one extra argument:

from bankstatementparser import CsvStatementParser
from bankstatementparser_writer_xlsx import write_xlsx

parser = CsvStatementParser("statement.csv")
df = parser.parse()
write_xlsx(df, "statement.xlsx", summary=parser.get_summary())

Input shapes

write_xlsx(data, path, *, sheet_name="Transactions", summary=None) accepts three input shapes and normalises each to a flat table:

A header row (bold) is written to the sheet_name sheet, followed by one row per record. Columns are auto-sized to their widest cell, capped at 60 characters so wide descriptions don't run off-screen (the cell content itself is never truncated — only the displayed column width).

Empty input is accepted: an empty list writes an empty sheet (no header), while an empty DataFrame that still carries column labels writes a header-only sheet.

Value coercion

Spreadsheet cells can't hold arbitrary Python objects, so the writer coerces the rich types the parser emits:

bool is preserved as a true/false cell (it is not coerced to 0/1), and a missing list[dict] key writes a blank cell in the same way a None value does.

Errors

write_xlsx validates its data argument and fails fast with a precise exception rather than writing a malformed workbook:

from bankstatementparser_writer_xlsx import write_xlsx

try:
    write_xlsx("not-a-table", "out.xlsx")  # a bare str is not a table
except TypeError as exc:
    print(f"rejected: {exc}")

try:
    write_xlsx([42], "out.xlsx")           # 42 is not a dict/Transaction
except ValueError as exc:
    print(f"rejected: {exc}")

The summary sheet

If you pass summary= a mapping (for example a parser's get_summary() result), the writer adds a second sheet titled Summary with a bold Key / Value header and one row per item:

from decimal import Decimal

from bankstatementparser_writer_xlsx import write_xlsx

transactions = [
    {"date": "2026-06-01", "description": "Salary", "amount": Decimal("3000.00")},
    {"date": "2026-06-03", "description": "Coffee Shop", "amount": Decimal("-4.20")},
]

write_xlsx(
    transactions,
    "out.xlsx",
    summary={
        "account_id": "DE89370400440532013000",
        "transaction_count": 128,
        "total_amount": Decimal("12045.67"),
        "currency": "EUR",
    },
)

Examples

Six runnable examples live in examples/ and are exercised in CI on every commit. Together they cover every supported input shape, option, coercion rule, and error path of write_xlsx:

• 01_write_dataframe.py — write a pandas DataFrame to a single sheet.
• 02_write_transactions.py — write a list of Transaction objects in stable field order.
• 03_write_dicts.py — write a list of plain dict records (union of keys).
• 04_write_with_summary.py — write a DataFrame plus a second Summary sheet via summary=.
• 05_custom_sheet_name.py — rename the transactions sheet with sheet_name=.
• 06_value_coercion_and_errors.py — Decimal/date coercion, None/NaN/missing keys as blank cells, the 60-character column-width cap, and the TypeError/ValueError error paths.

When not to use this package

• You need a custom sheet layout. The single-sheet (+ optional Summary) structure is intentionally simple. Compose your own openpyxl workbook if you need pivot-ready, multi-sheet layouts.
• You need .xls (legacy binary). openpyxl writes .xlsx only; convert downstream if you must.
• You need encrypted output. Out of scope; encrypt the produced .xlsx downstream with a tool like msoffcrypto-tool.
• You want to read Excel. This package is a writer.

Development

git clone https://github.com/sebastienrousseau/bankstatementparser-writer-xlsx
cd bankstatementparser-writer-xlsx
poetry env use python3.12
poetry install
poetry run pytest        # 100% line + branch coverage gate
poetry run ruff check bankstatementparser_writer_xlsx tests
poetry run mypy bankstatementparser_writer_xlsx
poetry run interrogate -c pyproject.toml bankstatementparser_writer_xlsx

Security

bankstatementparser-writer-xlsx consumes already-parsed data, not raw statement files — the PDF/CSV/XML parsing surface lives upstream in the bankstatementparser core. Reporting practice, supported versions, and supply-chain posture are documented in SECURITY.md.

Documentation

• README.md — this file
• ARCHITECTURE.md — module map and design decisions
• CHANGELOG.md — release notes
• ROADMAP.md — what's next
• SECURITY.md — disclosure + supported versions
• examples/ — runnable scripts, exercised in CI

License

Licensed under the Apache License, Version 2.0. Any contribution submitted for inclusion shall be licensed as above, without additional terms.

Contributing

Contributions are welcome — open an issue or PR on the repository.

Acknowledgements

Built on the bankstatementparser library and openpyxl.