Dynamic Data

Any bstruct.Encoding[T] can only describe data with a fixed layout. To make working with dynamic data layouts more convenient, the library supports writing to and reading from bytes buffers.

from io import BytesIO
from dataclasses import dataclass

import bstruct


@dataclass
class Header:
    item_count: bstruct.u8


HeaderEncoding = bstruct.derive(Header)


@dataclass
class Item:
    value: bstruct.u8


ItemEncoding = bstruct.derive(Item)


def decode(buffer: BytesIO) -> list[Item]:
    header = HeaderEncoding.read(buffer)
    items = ItemEncoding.read_many(buffer, count=header.item_count)

    return list(items)


def encode(items: list[Item], buffer: BytesIO) -> None:
    HeaderEncoding.write(Header(len(items)), buffer)
    ItemEncoding.write_many(items, buffer)


items = [Item(i) for i in range(100)]

# This could also be a file or a stream.
buffer = BytesIO()
encode(items, buffer)

# Reset buffer to the start, so the decode function has something to read.
buffer.seek(0)
decoded = decode(buffer)

assert items == decoded