如何正确序列化python bitarray？

Question

Simply using the tobytes and frombytes methods will not always give an equivalent object as shown below.

This seems to stem from the fact that the source bitarray does not have a length divisible by eight.

from bitarray import bitarray  # pip install bitarray


foo = bitarray('0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000010000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000010000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000010000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '000000000000000000000000000000000000000000000001000000',
               endian="little")


bar = bitarray(endian="little")
bar.frombytes(foo.tobytes())
assert(foo.tobytes() == bar.tobytes())  # passes
assert(foo == bar)  # fails

Answer 1

The solution is to save the length and trim the deserialized bitarray.

from bitarray import bitarray  # pip install bitarray
import json
from base64 import b64encode, b64decode


def serialize(ba):
    return json.dumps({
        "endian": ba.endian(),
        "bytes": b64encode(ba.tobytes()),
        "len": len(ba)
    })


def deserialize(data):
    data = json.loads(data)
    ba = bitarray(endian=data["endian"])
    ba.frombytes(b64decode(data["bytes"]))
    return ba[:data["len"]]


foo = bitarray('0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000010000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000010000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000010000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '0000000000000000000000000000000000000000000000000000000000'
               '000000000000000000000000000000000000000000000001000000',
               endian="little")


bar = deserialize(serialize(foo))
assert(foo.tobytes() == bar.tobytes())
assert(foo == bar)

Answer 2

You can save and load a bitarray directly with bitarray python module

# Save bitarray
ba = bitarray('11100000')
with open('bitarray.bin', 'wb') as f:
        ba.tofile(f)

# Load bitarray
ba = bitarray()
with open('bitarray.bin', 'rb') as f:
        ba.fromfile(f)