ieee754 – Handle IEEE754 floating-point format

Handle IEEE754 floating-point format.

The library.ieee754 module reads float numbers in a IEEE754 representation. This module do not implement the full specification of the standard and only read IEEE754 representation. The implementation supports the following subset:

• 32 bits - simple (binary32)

• 64 bits - double (binary64)

Using ieee754

The simple examples below shows how converting a IEEE754 representation of a float number expressed as a binary string in a float object.

We can read simple precision float number as shown below.

>>> from library.ieee754 import IEEE754
>>> IEEE754(b"\xab\xaa\xaa\x3e").value
0.3333333432674408
>>> IEEE754(b"\xab\xaa\xaa\x3e").value
0.3333333432674408
>>> IEEE754(b"\x00\x00\x00\x00").value
0.0
>>> IEEE754(b"\x00\x00\x80\x7f").value
inf

And the same with double precision.

>>> IEEE754( b"\x00\x00\x00\x00\x00\x00\xf0\x3f").value
1.0
>>> IEEE754(b"\x00\x00\x00\x00\x00\x00\xf0\xff").value
-inf
>>> IEEE754(b"\x55\x55\x55\x55\x55\x55\xd5\x3f").value
0.3333333333333333

The module supports only simple and double precision based on the lenght of the binary string; Any value other than 4 bytes (32 bits) or 8 bytes (64 bits) raise an KeyError Exception. The example below have a 56 bits lenght binary string.

>>> IEEE754(b"\x55\x55\x55\x55\x55\x55\xd5").value
KeyError: 'Unsupported IEEE754 Precision (56 bits)'

Reference

class darkbridge.library.ieee754.IEEE754(buffer: bytes)

Bases: object

IEEE754 float number.

The function supports single precision (32 bits), double precision (64 bits).

Parameters:

buffer – Bytes object of IEEE754 representation of the float number with the most significant byte is at the end of the byte array (byteorder = little).

Raises:

KeyError – Unsupported IEEE754 Precision.

value

Value of the float number.

Type:

float | None

_get_value(part: str) → int

Return the value of part.

This method extract the value from the IEEE representation and return it as an integer value.

Parameters:

part – name of the part (“sign”, “exponent”, “fraction”). Another name raise the KeyError exception.

Returns:

Value of the part as an integer