bitint 0.1.4

Pure Python implementation of extended integer with named bits.

bit-int

Pure Python implementation of extended integer with named bits.

I found that some Python engineers find working with bit math as sets unnatural. So I created this simple int extension to make bit manipulation more Pythonic and to demonstrate the beauty of Python magic.

What is BitInt good for in real life? It's as efficient as Python's own set type but it can be easily stored in databases, JSON or shared with other languages without any modification (C, Go, Java, JavaScript).

BitInt is highly inspired by great namedtuple from python standard library.

Installation

pip install bit-int

or pipenv

pipenv install -e 'git+https://github.com/czervenka/bit-set.git#egg=bitint'

Usage

>>> from bitint import bitint

BitInt is an integer type where bits can be named and manipulated as a set. Let's create a class Animals with first 8 bits named after animals kinds:

>>> Animals = bitint('Animals', 'cat, dog, mouse, bee, turtle, snake, frog, axolotl')

In this case, animals are represented by bits where cat is the first bit (2^0=1), dog is the second (2^1=2), mouse the third (2^2=4). You don't need to remember which animal is represented by which bit, because each species bit value is stored in property named after the species label:

>>> Animals.cat
1

Similarly dog label represents the second bit, which has value 2**1 = 2

>>> Animals.dog
2

We can now instantiate a subset of animals with set bits which corresponds to amphibians...

>>> amphibians = Animals('frog', 'axolotl')

... or animals which are commonly pets...

>>> pets = Animals('cat', 'dog')

Bit representation of pets is

0 0 0 0 0 0 1 1
            ^ ^
            | L cat (2^0 == 1 == Animals.cat)
            L dog (2^1 == 2 == Animals.dog)

Having this two sets of animals, we can use bit magic to test whether an animal bit is set to 1 or 0:

>>> "Dog is a pet" if Animals.dog & pets else "Dog is not a pet"
'Dog is a pet'

... or use in operator:

>>> Animals.dog in pets
True

What about if we try to check existence of an animal which has not been defined yet?

>>> Animals.elefant in pets
Traceback (most recent call last):
  ...
AttributeError: type object 'Animals' has no attribute 'elefant'

Beside checking for existense, it's possible to make basic set operations. For example, we can create a new Animals bitint which includes all pets and amphibians:

>>> pets | amphibians
Animals('cat', 'dog', 'frog', 'axolotl')

We can also find an intersection:

>>> hairy_animals = Animals('cat', 'dog', 'mouse')
>>> hairy_animals & pets
Animals('cat', 'dog')

... and even complement

>>> ~pets
Animals('mouse', 'bee', 'turtle', 'snake', 'frog', 'axolotl')

A set of all animals can be created using bit inversion of no animal

>>> all_animals = ~Animals()
>>> all_animals
Animals('cat', 'dog', 'mouse', 'bee', 'turtle', 'snake', 'frog', 'axolotl')

Animals can be added or removed from all animals by set and unset

>>> all_animals.unset(Animals.dog, Animals.bee, Animals.snake)
Animals('cat', 'mouse', 'turtle', 'frog', 'axolotl')

If you are curious which bits are set for named flags, print a bitint value.

>>> print(amphibians)
11000000
>>> print(pets | amphibians)
11000011

Furthermore you can also create list of or iterate over names of set bits

>>> list(pets)
['cat', 'dog']
>>> for pet in pets:
...     print(pet)
cat
dog

or for instance create regular Python set

>>> set(pets) == {'cat', 'dog'}
True

Beside everything else BitInt is just an enhanced integer :)

>>> pets + 1
4
>>> import json
>>> json.dumps({"pets_bitint": pets})
'{"pets_bitint": 3}'

Tip: If you change the named bits of a BitInt class (e.g. replace 'dog' with 'elephant') and you have stored bitints you need to make a migration otherwise all dogs become elephants. If you want to avoid migrations, never change existing named bit and only add new at the end of definition to utilize unused bits.

Update: My daughter found a bug in this readme:

>>> pets |= Animals.axolotl
>>> list(pets)
['cat', 'dog', 'axolotl']

Testing

Tests are documentation and vice-versa:

python -m doctest bitint.py README.md