Opinionated code formatter, just like Python's black code formatter but for Beancount
Try it out online here
- MIT licensed - based on beancount-parser, a Lark based LALR(1) Beancount syntax parser
- Extremely fast - 5K+ lines file generated by
bean-example
can be formatted in around 1 second - Section awareness - entries separated by Emac org symbol mark
*
will be formatted in groups without changing the overall structure - Comment preserving - comments are preserved and will be formatted as well
- Auto column width - calculate maximum column width and adjust accordingly
- Valid beancount file assumed - please notice that the formatter assumes the given beacnount file is valid, it doesn't not perform any kind of validation
The original project beancount-black was meant to be an internal tool built by Launch Platform LLC for
A modern accounting book service based on the most popular open source version control system Git and text-based double entry accounting book software Beancount. We realized adding new entries with BeanHub automatically over time makes beancount file a mess. So obviously, a strong code formatter is needed. While SaaS businesses won't be required to open source an internal tool like this, we still love that the service is only possible because of the open-source tool we are using. We think it would be greatly beneficial for the community to access a tool like this, so we've decided to open source it under MIT license, hope you find this tool useful 😄
To install the formatter, simply run
pip install beancount-black
Run
bean-black /path/to/file.bean
Then the file will be formatted.
Since this tool is still in its early stage, a backup file at <filepath>.backup
will be created automatically by default just in case.
The creation of backup files can be disabled by passing -n
or --no-backup
like this
bean-black -n /path/to/file.bean
It's highly recommended to use BeanHub, Git or other version control system to track your Beancount book files before running the formatter against them without a backup.
If you want to run the formatter programmatically, you can do this
import io
from beancount_parser.parser import make_parser
from beancount_black.formatter import Formatter
parser = make_parser()
formatter = Formatter()
tree = parser.parse(beancount_content)
output_file = io.StringIO()
formatter.format(tree, output_file)
You can run the formatter in STDIN mode by passing -s
or --stdin-mode
argument like this:
bean-black -s
Sometimes you might encounter problems when formatting some beancount files.
To debug and better understand which line is causing the problem, you can change the log level by passing -l
argument with debug
or verbose
.
For example:
bean-black -n /path/to/file.bean -l verbose
With verbose
, details of the parsed object will be printed.
With debug
, only the line number and type of entry will be printed.
You can also use LOG_LEVEL
environment variable as well.
The available log level options are:
- verbose
- debug
- info (default)
- error
- warning
- fatal
- Add argument for renaming account and commodity
- Add argument for following other files from
include
statements and also format those files
Feedbacks, bugs reporting or feature requests are welcome 🙌, just please open an issue. No guarantee we have time to deal with them, but will see what we can do.