unithandler 1.6.0

numeric mimic classes with units

This is a package for managing and interacting with units. The focus of the package is to provide numeric type objects (e.g. float and int) which also have a unit attribute. The three core classes of this package are Unit, UnitFloat, and UnitInt.

While there are a variety of packages which support units, this package focused on enabling straightforward implementation of units without affecting Pythonic functionality of the numbers.

Additionally, the UnitFloat class was designed to represent the magnitude of the stored value as best as possible with an appropriate SI prefix. The value stored is scaled in its representation, but remains locked to a specified prefix when operated upon. For example, 0.005 with the unit 'L' would be represented as 5 mL if printed, but will behave as 0.005 for any numeric operations. This was implemented to automatically keep track of the scale of a unit while allowing the user to easily visualize the magnitude of the stored value.

Installation

unithandler is registered on PyPI.

pip install unithandler

Source files may be found in the GitLab repository.

The Unit class

This class handles all unit-associated attributes and methods. The unit attribute returns an appropriately formatted unit (complete with superscripts) as a string. Units are managed in dictionary format, associating a power to each unit denoted. The class (and its subclasses) are structured) so that the units may be easily modified using python built-in multiplication and division operators (exactly as one would track units with pen and paper).

The UnitFloat class

The UnitFloat class is a float-like numeric class that has an associated unit. The class has all defined Python magic methods, which enable all appropriate modifications and operations on or by UnitFloat objects. Typically, the return of each magic method is another UnitFloat instance, but see the UnitFloat documentation for further details.

Known limitations

Since this package is written entirely in Python, it does not provide direct access to values as the true float and int would. As such, some errors or unintended functionality may be encountered.

Currently numpy support is limited to basic ufunc implementation. Unexpected behaviour may be encountered when performing numpy operations on UnitFloat or UnitInt. We are working to implement full support for vectorized operations on the numeric classes in this package. To the best of our knowledge, the math package is fully supported, so any numeric operations should be performed with that package while we work to enable numpy support.