Skip to content
/ packer Public

fast and light packer python dataclasses to binary data

Notifications You must be signed in to change notification settings

moff4/packer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

BINARY PACKER

Light and fast packer python dataclasses to bytes

Build Status CodeFactor

Install

Using PIP:

pip install binary-packer

How it works

It's a wrap! A wrap over stdlib struct
So it converts values into bytes using struct.pack().
But it does not save keys and values' types in bytes, so be sure about how you configure packer for packing and unpacking.

Why and when should it be used?

We all like to use json/xml/pickle and other popular libraries for dumping data, but sometimes they are too slow or result's size ot too large.
This library works much faster than other packers like json and makes result a lot more compact.

Example of Usage

import uuid
from typing import Optional
from dataclasses import dataclass

from binary_packer import PackerFactory, FieldStruct


@dataclass
class Person:
    id: uuid.UUID
    name: Optional[str] = None  # max length is 20 bytes
    age: Optional[int] = None

person = Person(id=uuid.uuid4(), name='vaschevsky', age=33)


factory = PackerFactory(
    Person,
    id=FieldStruct[uuid.UUID, bytes](
        '16s',  # '16s' is for bytearray length of 16, uuid as bytearray also 16 bytes
        encoder=lambda uid: uid.bytes,
        decoder=lambda uid_as_bytes: uuid.UUID(bytes=uid_as_bytes),
    ),
    name=FieldStruct[str, bytes](
        '20s',  # '20s' is for bytearray length of 20, each byte is ascii char
        encoder=lambda name: name.encode(),
        decoder=lambda name_as_bytes: name_as_bytes.decode().strip('\x00'),
    ),
    age=FieldStruct[int, int](
        'B',  #  'B' is for unsigned tiny int (1 byte) => age can be any value from 0 to 255,
        # no need for custom encoder/decoder
    )
)

packer_1 = factory.make_packer('id', 'name', 'age')
packer_2 = factory.make_packer('id', 'name')
packer_3 = factory.make_packer('id')

for packer in (packer_1, packer_2, packer_3):
    data_as_bytes = packer.pack(person)
    print(f'{len(data_as_bytes)=}')
    unpacked_person = packer.unpack(data_as_bytes)
    print(f'{unpacked_person=}')

# will be printed:
# len(data_as_bytes)=37
# unpacked_person=Person(id=UUID('a72decb7-7f9e-497b-ac91-692e316a7580'), name='vaschevsky', age=33)
# len(data_as_bytes)=36
# unpacked_person=Person(id=UUID('a72decb7-7f9e-497b-ac91-692e316a7580'), name='vaschevsky', age=None)
# len(data_as_bytes)=16
# unpacked_person=Person(id=UUID('a72decb7-7f9e-497b-ac91-692e316a7580'), name=None, age=None)

Buy Me A Coffee

About

fast and light packer python dataclasses to binary data

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published