Users, please report broken documentation links here.

Copy the test suite implemented in pMuTT for constants.

Most of the package was derived from pMuTT's constants module. That being said, there are some remnants of pMuTT that still need to be removed.

Currently, VUnits stores the unit database in its code, which is imported whenever the vunits.db module is accessed. However, prefixes and suffixes are added during the import process, which can slow down the initialization of the database.

One option is to save the database as a JSON file, which will probably load faster and can be updated whenever necessary.

Currently, units are handled in a very simple way where inputted strings are looked up in a dictionary to produce the output (usually a float). There are several limitations, including:

• prefix unit conversions are not robust
• it becomes cumbersome to list all the possible unit combinations.

A (lightweight) class could be developed to handle units. A unit parser that can convert units to its fundamental SI equivalent would be useful.

The following script:

from vunits.quantity import Quantity

length = Quantity.from_units(units='m')
time = Quantity.from_units(units='s')
speed = length/time

produces the following error:

Traceback (most recent call last):
File "c:/Users/jonat/Desktop/test_qty.py", line 5, in
speed = length/time
TypeError: unsupported operand type(s) for /: 'Quantity' and 'Quantity'

This should be handled by __div__ magic method.

Currently, VUnits only attaches an 's' to the end of long units to create the plural form. This simple approach leads to incorrect words, like "inchs".

Instead of the boolean attribute plural_allowed, units should have a string attribute indicating the suffix to use.

Currently, Quantity objects can hold NumPy arrays and perform unit conversions with them. For example:

>>> import numpy as np
>>> from vunits.quantity import Quantity
>>> time = Quantity.from_units(np.array([0., 1., 2.]), 'hr') 
>>> time('min') 
array([  0.,  60., 120.])

However, more complex operations (for example, taking a sum of the array) is not possible.

>>> np.sum(time)
<vunits.quantity.Quantity object at 0x000001C1EB142190>

Perhaps implementing the magic methods suggested on NumPy's Dispatch page can increase Quantity object's utility.