pintrs 0.2.2

Fast Rust-based Python extension for physical unit manipulation

pintrs

Fast physical units for Python, with a pint-compatible API and Rust performance.

pintrs is for people who like pint's ergonomics but not its runtime cost. It keeps the familiar UnitRegistry and Quantity workflow, with the hot path implemented in Rust.

• Drop into common pint workflows with minimal code changes
• Usually 8-150x faster on core operations
• Works with NumPy, pandas, Babel, measurements, contexts, groups, and systems

Why pintrs

If your code spends real time creating quantities, parsing unit strings, or converting units, pintrs removes a lot of overhead without asking you to relearn the API.

• Quantity creation: 9x faster
• Parsing unit strings: 152x faster
• Same-unit addition: 14x faster

Benchmarks below were measured with Python 3.12. Lower is better.

Operation	pintrs	pint	Speedup
Quantity creation	0.37 us	3.33 us	9x
Parse string ("9.81 m/s**2")	0.55 us	50.21 us	91x
Conversion (km -> m)	0.94 us	7.43 us	8x
Conversion (km/h -> m/s)	1.71 us	13.30 us	8x
Addition (same units)	0.34 us	4.64 us	14x
Addition (compatible units)	1.03 us	11.56 us	11x
Multiply by scalar	0.25 us	5.31 us	21x
Multiply quantities	0.38 us	5.01 us	13x
Comparison (>)	0.12 us	1.16 us	10x
To base units	0.37 us	6.43 us	17x
Parse units ("kg * m / s ** 2")	0.26 us	38.85 us	152x
String formatting	0.44 us	7.20 us	16x

Run python examples/benchmark.py to reproduce the numbers. Install pint alongside pintrs for the comparison run.

Migrating from pint

If you already use pint, the change is intentionally small: replace pint with pintrs in your dependencies and swap your imports.

- pint
+ pintrs

- from pint import UnitRegistry
+ from pintrs import UnitRegistry

ureg = UnitRegistry()

Your existing quantity code should continue to look like pint code:

distance = 5 * ureg.kilometer
time = 2 * ureg.hour
speed = distance / time

print(speed)           # 2.5 kilometer / hour
print(speed.to("m/s")) # 0.6944... meter / second

Compatibility with pint

pintrs targets full API compatibility with pint.

That includes the core registry and quantity model, conversions and formatting, decorators, measurements, contexts, groups, systems, and integrations with NumPy, pandas, and Babel.

If you already have working pint code and performance is the problem, pintrs is designed to be the least disruptive upgrade path.

Installation

pip install pintrs

NumPy, pandas, and Babel integrations are available when those packages are installed.

What you get

• The familiar pint API, with Rust underneath
• Substantial speedups on quantity creation, parsing, conversion, arithmetic, and formatting
• Support for NumPy, pandas, Babel, measurements, contexts, groups, systems, logarithmic units, and decorators
• Type information for mypy and pyright