ctypes-plus 0.1.0

A dataclass-like interface for ctypes data types.

ctypes-plus

A dataclass-like interface for ctypes data types.

[!WARNING] This package is under development. The API is subject to breaking changes.

Usage

Decorate an annotated class with @structure or @union. Each annotation becomes a _fields_ entry, and the result is a ctypes.Structure / ctypes.Union subclass usable anywhere ctypes is expected.

from ctypes_plus import structure, union
from ctypes_plus.types import CharP, Double, Int32, UInt64

@structure
class LogEvent:
    event_id: Int32
    timestamp: UInt64
    message: CharP

@union
class Number:
    as_int: Int32
    as_double: Double

event = LogEvent(event_id=2, timestamp=1234, message=b"foo")
print(event)  # LogEvent(event_id=2, timestamp=1234, message=b'foo')

Typed field aliases

ctypes_plus.types provides Annotated types that map a native Python type to its ctypes type, so construction and attribute access type-check cleanly.

Plain ctypes types (a: c_int), your own Annotated[..., ctype] aliases and type aliases of either are all accepted too.

Introspection

>>> from ctypes_plus import asdict, fields
>>> fields(LogEvent)
(Field(name='event_id', type=<class 'ctypes.c_int'>), ...)
>>> asdict(LogEvent(event_id=2, timestamp=1234, message=b"foo"))
{'event_id': 2, 'timestamp': 1234, 'message': b'foo'}

fields accepts a class or instance — including hand-written ctypes types — and returns a tuple of Field(name, type) records. asdict takes an instance and returns its fields as a dict.

Licence

MIT