Simple precision control

[timlnx]

From @kapsh in #61

What do you think about adding precision argument to best_prefix method as shortcut for most commonly required operation?

This is a good idea. I also find the method for controlling precision rather verbose

As a bitmath user I want a simple way to control precision printing so that I don't have to use extra verbose syntax

It's a fairly open definition at the moment, one which needs refinement. I think the idea is a very solid one though and I bet there are a lot of people who would find value in this.

[kapsh]

@tbielawa thanks for feedback.
I already took a look into the code and found that best_prefix actually returns instance of suitable subclass of Bitmath. In this case it would be awkward to pass precision argument to the constructor.

Next suggestion then. With custom __format__ magic method bitmath objects could partially understand format mini-language (like datetime objects). So I assume the usage example as:

>>> size = bitmath.Byte(314159265)
>>> size.best_prefix()
MiB(299.60562229156494)
>>> 'Size is {0:.2}'.format(size.best_prefix())
Size is 299.61 MiB

It looks quite pythonic IMO.

[timlnx]

I see what you're going for. I didn't realize before that there is a magic __format__ method. I'm going to look at that closer.

In the mean time, consider the following work-around as well: adjusting the default format string. As noted in the docs, each instance can already be formatted directory using the format method:

In [15]: ex = bitmath.MiB(10.0/3.0)

In [16]: print(ex)
3.33333333333 MiB

In [17]: print(ex.format('{value:0.2f} {unit}'))
3.33 MiB

However, if you want 2-digit precision by default, you can simply change the default formatting string:

In [1]: import bitmath

In [2]: bitmath.format_string = '{value:0.2f} {unit}'

In [3]: ex = bitmath.MiB(10.0/3.0)

In [4]: print(ex)
3.33 MiB

In [5]: print(ex.format('{value:0.2f} {unit}'))
3.33 MiB