Python values with units

Question

I need to keep track of units on float and int values in Python, but I don't want to use an external package like magnitude or others, because I don't need to perform operations on the values. Instead, all I want is to be able to define floats and ints that have a unit attribute (and I don't want to add a new dependency for something this simple). I tried doing:

class floatwithunit(float):

    __oldinit__ = float.__init__

    def __init__(self, *args, **kwargs):
        if 'unit' in kwargs:
            self.unit = kwargs.pop('unit')
        self.__oldinit__(*args, **kwargs)

But this doesn't work at all:

In [37]: a = floatwithunit(1.,unit=1.)
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)

/Users/tom/<ipython console> in <module>()

TypeError: float() takes at most 1 argument (2 given)

Any suggestions?

Answer 1

It looks like you need to check if kwargs is not None before you try to see if there is a label 'unit'.

Change your code to

 if kwargs and 'unit' in kwargs:

Updated answer:

don't pass kwargs to __oldinit__

Answer 2

I think you mean

class floatwithunit(float):

rather than

def floatwithunit(float):

Answer 3

You need to override __new__ (the "constructor proper", while __init__ is the "initializer"), otherwise float's __new__ gets called with extraneous arguments, which is the cause of the problem you're seeing. You don't need to call float's __init__ (it's a no-op). Here's how I'd code it:

class floatwithunit(float):

    def __new__(cls, value, *a, **k):
        return float.__new__(cls, value)

    def __init__(self, value, *args, **kwargs):
        self.unit = kwargs.pop('unit', None)

    def __str__(self):
        return '%f*%s' % (self, self.unit)

a = floatwithunit(1.,unit=1.)

print a

emitting 1.000000*1.0.

Answer 4

You might be looking for something like this:

class UnitFloat(float):

    def __new__(self, value, unit=None):
       return float.__new__(self, value)

    def __init__(self, value, unit=None):
        self.unit = unit


x = UnitFloat(35.5, "cm")
y = UnitFloat(42.5)

print x
print x.unit

print y
print y.unit

print x + y

Yields:

35.5
cm
42.5
None
78.0

Answer 5

Alex Martelli indeed pointed out the root of the problem. I always find __new__ quite confusing, however, so here's a working piece of example code:

class FloatWithUnit(float):
    def __new__(cls, *args, **kwargs):
        # avoid error in float.__new__
        # the original kwargs (with 'unit') will still be passed to __init__
        if 'unit' in kwargs:
            kwargs.pop('unit')
        return super(FloatWithUnit, cls).__new__(cls, *args, **kwargs)

    def __init__(self, *args, **kwargs):
        self.unit = kwargs.pop('unit') if 'unit' in kwargs else None
        super(FloatWithUnit, self).__init__(*args, **kwargs)