Full multi-currency support for python.
Quickstart Tutorial: https://grammacc.github.io/linearmoney/quickstart.html
Full Documentation: https://grammacc.github.io/linearmoney
License: MIT
This project uses semantic versioning. However, this is a pre-release piece of software, so until the first stable release, minor versions may contain breaking changes. Version 1.0.0 will mark the first stable release, at which point the regular rules of semver will apply and backwards-incompatible changes will only be introduced in major versions.
linearmoney was created to make calculations and formatting with multiple currencies in modern international applications easier and more reliable.
Key Features:
- Full support for arithmetic with monetary amounts in different currencies.
- Full support for non-destructive currency conversion.
- Full support for fractional currency rounding and fixed-point rounding.
- Full support for currency formatting and localization.
- No dependencies other than Python itself.
- Completely thread-safe.
- 100% Test and API Documentation coverage.
- Database/ORM integrations.
The linearmoney library takes a non-traditional approach to financial applications by using linear algebra internally to ensure the correctness of monetary calculations involving multiple currencies without passing this burden onto the programmer. Understanding of linear algebra is not needed to use and understand the linearmoney library, but an understanding of basic arithmetic with vectors is helpful for understanding how the library works under the hood.
For a technical explanation of the motivation and philosophy behind linearmoney as well as the complete pure-math model that defines the behaviors of the library, see the Linear Money Model article.
linearmoney uses the amazing Unicode CLDR-JSON data to provide data-driven interfaces for currency rounding, formatting, and localization.
linearmoney requires Python >= 3.10
From PyPi:
pip install linearmoneyFrom source:
git clone https://github.com/GrammAcc/linearmoney
cd linearmoney
python -m build .Then to install (virtual environment recommended):
pip install path/to/cloned/repo>>> import linearmoney as lm
>>> fo = lm.vector.forex({"base": "usd", "rates": {"jpy": 100}}) # 1 USD -> 100 JPY
>>> sp = lm.vector.space(fo)
>>> cart = []
>>> local_milk_price = lm.vector.asset(4.32, "usd", sp)
>>> cart.append(local_milk_price)
>>> foreign_eggs_price = lm.vector.asset(545, "jpy", sp)
>>> cart.append(foreign_eggs_price)
>>> sales_tax = 0.095
>>> subtotal = sum(cart)
>>> total = subtotal + (subtotal * sales_tax)
>>> total_usd = lm.vector.evaluate(total, "usd", fo)
>>> total_jpy = lm.vector.evaluate(total, "jpy", fo)
>>> usd = lm.data.currency("usd")
>>> jpy = lm.data.currency("jpy")
>>> rounded_total_usd = lm.round.as_currency(total_usd, usd)
>>> rounded_total_jpy = lm.round.as_currency(total_jpy, jpy)
>>> en_US = lm.data.locale("en", "us")
>>> localized_total_usd = lm.scalar.l10n(rounded_total_usd, usd, en_US)
>>> localized_total_jpy = lm.scalar.l10n(rounded_total_jpy, jpy, en_US)
>>> print(localized_total_usd)
$10.70
>>> print(localized_total_jpy)
¥1,070
linearymoney uses a functional/procedural style where all objects are immutable, so the code can become verbose compared to more idiomatic Python, but this also makes the code more explicit and easier to test.
See the Recipes section of the user guide for some examples of how to mitigate the verbosity of the library and other helpful patterns.
The linearmoney.ext sub-package provides optional integrations with other libraries/tools.
Most tools shouldn't need any kind of adapter layer for you to use linearmoney with them since you would normally evaluate a money vector whenever you need to do something with its value outside of linearmoney's functions or math between vectors. The exceptions to this are ORMs and similar tools that need some kind of serialization step to be performed.
SQLAlchemy integrations are implemented, and Django ORM is planned.
If there is a tool that you want better integration with or simply some kind of extra functionality that requires additional dependencies, please open an issue and we will evaluate if it is within the scope of the project to support as an extension.
Contributions are greatly appreciated!
See the contributing guidelines to get started.
Version 1.0.0:
- Redesign locale/formatting data structure
- Redesign caching system
- Higher-order serialization interface
- Serialization/deserialization of forex vectors
- Recipes to add
- Use-cases without vectors
- Refactor CLDR data processing script
- Add contributing guidelines and setup CI
- Contributing guidelines
- CI workflow